Skip to content

PTQ-Level Efficiency, QAT-Level Accuracy | 训练后量化的效率,量化感知训练的精度

support to export deepseek v3 gguf format (#416)

Heng Guo6a49304ee904
diff --git a/auto_round/export/export_to_gguf/convert.py b/auto_round/export/export_to_gguf/convert.py
index 0e92d57f4..41f00aad7 100644
--- a/auto_round/export/export_to_gguf/convert.py
+++ b/auto_round/export/export_to_gguf/convert.py
@@ -32,7 +32,6 @@
# OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
# SOFTWARE.
-
from __future__ import annotations
import ast
@@ -58,11 +57,12 @@ if TYPE_CHECKING:
from auto_round.utils import logger, LazyImport
from .quant import ggml_quant
-gguf = LazyImport("gguf")
+gguf = LazyImport("gguf")
###### MODEL DEFINITIONS ######
+
class SentencePieceTokenTypes(IntEnum):
NORMAL = 1
UNKNOWN = 2
@@ -75,14 +75,14 @@ class SentencePieceTokenTypes(IntEnum):
AnyModel = TypeVar("AnyModel", bound="type[Model]")
-class Model:
+class OriModel:
_model_classes: dict[str, type[Model]] = {}
dir_model: Path
ftype: gguf.LlamaFileType
fname_out: Path
is_big_endian: bool
- endianness: gguf.GGUFEndian
+ endianess: gguf.GGUFEndian
use_temp_file: bool
lazy: bool
part_names: list[str]
@@ -99,30 +99,43 @@ class Model:
# subclasses should define this!
model_arch: gguf.MODEL_ARCH
- def __init__(self, model, dir_model: Path, ftype: gguf.LlamaFileType, fname_out: Path, is_big_endian: bool = False,
- layer_config = None, use_temp_file: bool = False, eager: bool = False,
- metadata_override: Path | None = None, model_name: str | None = None,
- split_max_tensors: int = 0, split_max_size: int = 0, dry_run: bool = False,
- small_first_shard: bool = False, hparams: dict[str, Any] | None = None):
+ def __init__(self,
+ dir_model: Path,
+ ftype: gguf.LlamaFileType,
+ fname_out: Path,
+ is_big_endian: bool = False,
+ use_temp_file: bool = False,
+ eager: bool = False,
+ metadata_override: Path | None = None,
+ model_name: str | None = None,
+ split_max_tensors: int = 0,
+ split_max_size: int = 0,
+ dry_run: bool = False,
+ small_first_shard: bool = False,
+ hparams: dict[str, Any] | None = None):
if type(self) is Model:
- raise TypeError(f"{type(self).__name__!r} should not be directly instantiated")
+ raise TypeError(
+ f"{type(self).__name__!r} should not be directly instantiated")
- self.model = model
self.dir_model = dir_model
self.ftype = ftype
self.fname_out = fname_out
- self.layer_config = layer_config
self.is_big_endian = is_big_endian
self.endianess = gguf.GGUFEndian.BIG if is_big_endian else gguf.GGUFEndian.LITTLE
self.use_temp_file = use_temp_file
self.lazy = not eager
- self.part_names = Model.get_model_part_names(self.dir_model, "model", ".safetensors")
+ self.part_names = Model.get_model_part_names(self.dir_model, "model",
+ ".safetensors")
self.is_safetensors = len(self.part_names) > 0
if not self.is_safetensors:
- self.part_names = Model.get_model_part_names(self.dir_model, "pytorch_model", ".bin")
- self.hparams = Model.load_hparams(self.dir_model) if hparams is None else hparams
- self.block_count = self.find_hparam(["n_layers", "num_hidden_layers", "n_layer", "num_layers"])
- self.tensor_map = gguf.get_tensor_name_map(self.model_arch, self.block_count)
+ self.part_names = Model.get_model_part_names(
+ self.dir_model, "pytorch_model", ".bin")
+ self.hparams = Model.load_hparams(
+ self.dir_model) if hparams is None else hparams
+ self.block_count = self.find_hparam(
+ ["n_layers", "num_hidden_layers", "n_layer", "num_layers"])
+ self.tensor_map = gguf.get_tensor_name_map(self.model_arch,
+ self.block_count)
self.tensor_names = None
self.metadata_override = metadata_override
self.model_name = model_name
@@ -133,25 +146,26 @@ class Model:
# NOTE: can't use field "torch_dtype" in config.json, because some finetunes lie.
_, first_tensor = next(self.get_tensors())
if first_tensor.dtype == torch.float16:
- logger.info(f"choosing --outtype f16 from first tensor type ({first_tensor.dtype})")
+ logger.info(
+ f"choosing --outtype f16 from first tensor type ({first_tensor.dtype})"
+ )
self.ftype = gguf.LlamaFileType.MOSTLY_F16
else:
- logger.info(f"choosing --outtype bf16 from first tensor type ({first_tensor.dtype})")
+ logger.info(
+ f"choosing --outtype bf16 from first tensor type ({first_tensor.dtype})"
+ )
self.ftype = gguf.LlamaFileType.MOSTLY_BF16
# Configure GGUF Writer
self.gguf_writer = gguf.GGUFWriter(
- path=None, arch=gguf.MODEL_ARCH_NAMES[self.model_arch],
- endianess=self.endianess, use_temp_file=self.use_temp_file,
- split_max_tensors=split_max_tensors, split_max_size=split_max_size,
- dry_run=dry_run, small_first_shard=small_first_shard)
-
- @classmethod
- def __init_subclass__(cls):
- # can't use an abstract property, because overriding it without type errors
- # would require using decorated functions instead of simply defining the property
- if "model_arch" not in cls.__dict__:
- raise TypeError(f"Missing property 'model_arch' for {cls.__name__!r}")
+ path=None,
+ arch=gguf.MODEL_ARCH_NAMES[self.model_arch],
+ endianess=self.endianess,
+ use_temp_file=self.use_temp_file,
+ split_max_tensors=split_max_tensors,
+ split_max_size=split_max_size,
+ dry_run=dry_run,
+ small_first_shard=small_first_shard)
def find_hparam(self, keys: Iterable[str], optional: bool = False) -> Any:
key = next((k for k in keys if k in self.hparams), None)
@@ -165,22 +179,95 @@ class Model:
self._set_vocab_gpt2()
def get_tensors(self) -> Iterator[tuple[str, Tensor]]:
- for name, tensor in self.model.named_parameters():
- yield name, tensor
-
-
- def format_tensor_name(
- self, key: gguf.MODEL_TENSOR, bid: int | None = None, suffix: str = ".weight") -> str:
+ tensor_names_from_parts: set[str] = set()
+
+ index_name = "model.safetensors" if self.is_safetensors else "pytorch_model.bin"
+ index_name += ".index.json"
+ index_file = self.dir_model / index_name
+
+ if index_file.is_file():
+ self.tensor_names = set()
+ logger.info(f"gguf: loading model weight map from '{index_name}'")
+ with open(index_file, "r", encoding="utf-8") as f:
+ index: dict[str, Any] = json.load(f)
+ weight_map = index.get("weight_map")
+ if weight_map is None or not isinstance(weight_map, dict):
+ raise ValueError(
+ f"Can't load 'weight_map' from {index_name!r}")
+ self.tensor_names.update(weight_map.keys())
+ else:
+ self.tensor_names = tensor_names_from_parts
+ weight_map = {}
+
+ for part_name in self.part_names:
+ logger.info(f"gguf: loading model part '{part_name}'")
+ ctx: ContextManager[Any]
+ if self.is_safetensors:
+ from safetensors import safe_open
+ ctx = cast(
+ ContextManager[Any],
+ safe_open(self.dir_model / part_name,
+ framework="pt",
+ device="cpu"))
+ else:
+ ctx = contextlib.nullcontext(
+ torch.load(str(self.dir_model / part_name),
+ map_location="cpu",
+ mmap=True,
+ weights_only=True))
+
+ with ctx as model_part:
+ tensor_names_from_parts.update(model_part.keys())
+
+ for name in model_part.keys():
+ if self.is_safetensors:
+ if self.lazy:
+ data = model_part.get_slice(name)
+ data = LazyTorchTensor.from_safetensors_slice(data)
+ else:
+ data = model_part.get_tensor(name)
+ else:
+ data = model_part[name]
+ if self.lazy:
+ data = LazyTorchTensor.from_eager(data)
+ yield name, data
+
+ # verify tensor name presence and identify potentially missing files
+ if len(tensor_names_from_parts.symmetric_difference(
+ self.tensor_names)) > 0:
+ missing = sorted(
+ self.tensor_names.difference(tensor_names_from_parts))
+ extra = sorted(
+ tensor_names_from_parts.difference(self.tensor_names))
+ missing_files = sorted(
+ set(weight_map[n] for n in missing if n in weight_map))
+ if len(extra) == 0 and len(missing_files) > 0:
+ raise ValueError(
+ f"Missing or incomplete model files: {missing_files}")
+ else:
+ raise ValueError(
+ "Mismatch between weight map and model parts for tensor names:\n"
+ f"Missing tensors: {missing}\n"
+ f"Extra tensors: {extra}")
+
+ def format_tensor_name(self,
+ key: gguf.MODEL_TENSOR,
+ bid: int | None = None,
+ suffix: str = ".weight") -> str:
if key not in gguf.MODEL_TENSORS[self.model_arch]:
- raise ValueError(f"Missing {key!r} for MODEL_TENSORS of {self.model_arch!r}")
+ raise ValueError(
+ f"Missing {key!r} for MODEL_TENSORS of {self.model_arch!r}")
name: str = gguf.TENSOR_NAMES[key]
if "{bid}" in name:
assert bid is not None
name = name.format(bid=bid)
return name + suffix
- def match_model_tensor_name(
- self, name: str, key: gguf.MODEL_TENSOR, bid: int | None, suffix: str = ".weight") -> bool:
+ def match_model_tensor_name(self,
+ name: str,
+ key: gguf.MODEL_TENSOR,
+ bid: int | None,
+ suffix: str = ".weight") -> bool:
if key not in gguf.MODEL_TENSORS[self.model_arch]:
return False
key_name: str = gguf.TENSOR_NAMES[key]
@@ -193,8 +280,11 @@ class Model:
return False
return name == (key_name + suffix)
- def map_tensor_name(self, name: str, try_suffixes: Sequence[str] = (".weight", ".bias")) -> str:
- new_name = self.tensor_map.get_name(key=name, try_suffixes=try_suffixes)
+ def map_tensor_name(
+ self, name: str,
+ try_suffixes: Sequence[str] = (".weight", ".bias")) -> str:
+ new_name = self.tensor_map.get_name(key=name,
+ try_suffixes=try_suffixes)
if new_name is None:
raise ValueError(f"Can not map tensor {name!r}")
return new_name
@@ -202,21 +292,25 @@ class Model:
def set_gguf_parameters(self):
self.gguf_writer.add_block_count(self.block_count)
- if (n_ctx := self.find_hparam(["max_position_embeddings", "n_ctx"], optional=True)) is not None:
+ if (n_ctx := self.find_hparam(["max_position_embeddings", "n_ctx"],
+ optional=True)) is not None:
self.gguf_writer.add_context_length(n_ctx)
logger.info(f"gguf: context length = {n_ctx}")
- n_embd = self.find_hparam(["hidden_size", "n_embd"])
- self.gguf_writer.add_embedding_length(n_embd)
- logger.info(f"gguf: embedding length = {n_embd}")
+ if (n_embd := self.find_hparam(["hidden_size", "n_embd"],
+ optional=True)) is not None:
+ self.gguf_writer.add_embedding_length(n_embd)
+ logger.info(f"gguf: embedding length = {n_embd}")
- if (n_ff := self.find_hparam(["intermediate_size", "n_inner"], optional=True)) is not None:
+ if (n_ff := self.find_hparam(["intermediate_size", "n_inner"],
+ optional=True)) is not None:
self.gguf_writer.add_feed_forward_length(n_ff)
logger.info(f"gguf: feed forward length = {n_ff}")
- n_head = self.find_hparam(["num_attention_heads", "n_head"])
- self.gguf_writer.add_head_count(n_head)
- logger.info(f"gguf: head count = {n_head}")
+ if (n_head := self.find_hparam(["num_attention_heads", "n_head"],
+ optional=True)) is not None:
+ self.gguf_writer.add_head_count(n_head)
+ logger.info(f"gguf: head count = {n_head}")
if (n_head_kv := self.hparams.get("num_key_value_heads")) is not None:
self.gguf_writer.add_head_count_kv(n_head_kv)
@@ -229,13 +323,15 @@ class Model:
self.gguf_writer.add_layer_norm_rms_eps(f_rms_eps)
logger.info(f"gguf: rms norm epsilon = {f_rms_eps}")
if (f_norm_eps := self.find_hparam(
- ["layer_norm_eps", "layer_norm_epsilon", "norm_epsilon"], optional=True)) is not None:
+ ["layer_norm_eps", "layer_norm_epsilon", "norm_epsilon"],
+ optional=True)) is not None:
self.gguf_writer.add_layer_norm_eps(f_norm_eps)
logger.info(f"gguf: layer norm epsilon = {f_norm_eps}")
if (n_experts := self.hparams.get("num_local_experts")) is not None:
self.gguf_writer.add_expert_count(n_experts)
logger.info(f"gguf: expert count = {n_experts}")
- if (n_experts_used := self.hparams.get("num_experts_per_tok")) is not None:
+ if (n_experts_used :=
+ self.hparams.get("num_experts_per_tok")) is not None:
self.gguf_writer.add_expert_used_count(n_experts_used)
logger.info(f"gguf: experts used count = {n_experts_used}")
@@ -246,13 +342,14 @@ class Model:
self.gguf_writer.add_file_type(self.ftype)
logger.info(f"gguf: file type = {self.ftype}")
- def modify_tensors(self, data_torch: Tensor, name: str, bid: int | None) -> Iterable[tuple[str, Tensor]]:
+ def modify_tensors(self, data_torch: Tensor, name: str,
+ bid: int | None) -> Iterable[tuple[str, Tensor]]:
del bid # unused
return [(self.map_tensor_name(name), data_torch)]
- def tensor_force_quant(
- self, name: str, new_name: str, bid: int | None, n_dims: int) -> gguf.GGMLQuantizationType | bool:
+ def tensor_force_quant(self, name: str, new_name: str, bid: int | None,
+ n_dims: int) -> gguf.GGMLQuantizationType | bool:
del name, new_name, bid, n_dims # unused
return False
@@ -262,11 +359,15 @@ class Model:
return ()
def prepare_tensors(self):
- max_name_len = max(len(s) for _, s in self.tensor_map.mapping.values()) + len(".weight,")
+ max_name_len = max(
+ len(s)
+ for _, s in self.tensor_map.mapping.values()) + len(".weight,")
- for name, data_torch in chain(self.generate_extra_tensors(), self.get_tensors()):
+ for name, data_torch in chain(self.generate_extra_tensors(),
+ self.get_tensors()):
# we don't need these
- if name.endswith((".attention.masked_bias", ".attention.bias", ".rotary_emb.inv_freq")):
+ if name.endswith((".attention.masked_bias", ".attention.bias",
+ ".rotary_emb.inv_freq")):
continue
old_dtype = data_torch.dtype
@@ -282,15 +383,19 @@ class Model:
bid = int(part)
break
- for new_name, data_torch in (self.modify_tensors(data_torch, name, bid)):
- data = data_torch.squeeze().numpy()
+ for new_name, data_torch in (self.modify_tensors(
+ data_torch, name, bid)):
+ # TODO: why do we squeeze here?
+ # data = data_torch.squeeze().numpy()
+ data = data_torch.numpy()
# if data ends up empty, it means data_torch was a scalar tensor -> restore
if len(data.shape) == 0:
data = data_torch.numpy()
n_dims = len(data.shape)
- data_qtype: gguf.GGMLQuantizationType | bool = self.tensor_force_quant(name, new_name, bid, n_dims)
+ data_qtype: gguf.GGMLQuantizationType | bool = self.tensor_force_quant(
+ name, new_name, bid, n_dims)
# Most of the codebase that takes in 1D tensors or norms only handles F32 tensors
if n_dims <= 1 or new_name.endswith("_norm.weight"):
@@ -298,8 +403,7 @@ class Model:
# Conditions should closely match those in llama_model_quantize_internal in llama.cpp
# Some tensor types are always in float32
- if data_qtype is False and (
- any(
+ if data_qtype is False and (any(
self.match_model_tensor_name(new_name, key, bid)
for key in (
gguf.MODEL_TENSOR.FFN_GATE_INP,
@@ -311,28 +415,24 @@ class Model:
gguf.MODEL_TENSOR.TIME_MIX_W2,
gguf.MODEL_TENSOR.TIME_MIX_DECAY_W1,
gguf.MODEL_TENSOR.TIME_MIX_DECAY_W2,
- )
- )
- or not new_name.endswith(".weight")
- ):
+ gguf.MODEL_TENSOR.TIME_MIX_LERP_FUSED,
+ gguf.MODEL_TENSOR.POSNET_NORM1,
+ gguf.MODEL_TENSOR.POSNET_NORM2,
+ )) or not new_name.endswith(".weight")):
data_qtype = gguf.GGMLQuantizationType.F32
if data_qtype is False and any(
- self.match_model_tensor_name(new_name, key, bid)
- for key in (
- gguf.MODEL_TENSOR.TOKEN_EMBD,
- gguf.MODEL_TENSOR.OUTPUT,
- )
- ):
+ self.match_model_tensor_name(new_name, key, bid)
+ for key in (
+ gguf.MODEL_TENSOR.TOKEN_EMBD,
+ gguf.MODEL_TENSOR.OUTPUT,
+ )):
if self.ftype in (
- gguf.LlamaFileType.MOSTLY_TQ1_0,
- gguf.LlamaFileType.MOSTLY_TQ2_0,
+ gguf.LlamaFileType.MOSTLY_TQ1_0,
+ gguf.LlamaFileType.MOSTLY_TQ2_0,
):
# TODO: use Q4_K and Q6_K
data_qtype = gguf.GGMLQuantizationType.F16
- # data_qtype = gguf.GGMLQuantizationType.Q8_0 # llama.cpp:llama_tensor_get_type
-
-
# No override (data_qtype is False), or wants to be quantized (data_qtype is True)
if isinstance(data_qtype, bool):
@@ -342,61 +442,50 @@ class Model:
data_qtype = gguf.GGMLQuantizationType.F16
elif self.ftype == gguf.LlamaFileType.MOSTLY_BF16:
data_qtype = gguf.GGMLQuantizationType.BF16
- # elif self.ftype == gguf.LlamaFileType.MOSTLY_Q8_0:
- # data_qtype = gguf.GGMLQuantizationType.Q8_0
- # elif self.ftype == gguf.LlamaFileType.MOSTLY_TQ1_0:
- # data_qtype = gguf.GGMLQuantizationType.TQ1_0
- # elif self.ftype == gguf.LlamaFileType.MOSTLY_TQ2_0:
- # data_qtype = gguf.GGMLQuantizationType.TQ2_0
- elif self.ftype == gguf.LlamaFileType.MOSTLY_Q4_0:
- data_qtype = gguf.GGMLQuantizationType.Q4_0
- elif self.ftype == gguf.LlamaFileType.MOSTLY_Q4_1:
- data_qtype = gguf.GGMLQuantizationType.Q4_1
+ elif self.ftype == gguf.LlamaFileType.MOSTLY_Q8_0:
+ data_qtype = gguf.GGMLQuantizationType.Q8_0
+ elif self.ftype == gguf.LlamaFileType.MOSTLY_TQ1_0:
+ data_qtype = gguf.GGMLQuantizationType.TQ1_0
+ elif self.ftype == gguf.LlamaFileType.MOSTLY_TQ2_0:
+ data_qtype = gguf.GGMLQuantizationType.TQ2_0
else:
- raise ValueError(f"Unknown file type: {self.ftype.name}")
-
- # try:
- # data = gguf.quants.quantize(data, data_qtype)
- # except gguf.QuantError as e:
- # logger.warning("%s, %s", e, "falling back to F16")
- # data_qtype = gguf.GGMLQuantizationType.F16
- # data = gguf.quants.quantize(data, data_qtype)
+ raise ValueError(
+ f"Unknown file type: {self.ftype.name}")
- # TODO: Quant here
- suffix = '.weight'
- if suffix in name and name[:-len(suffix)] in self.layer_config:
- layer_name = name[:-len(suffix)]
- if "scale" in self.layer_config[layer_name]:
- scale = self.layer_config[layer_name]['scale'].numpy()
- zp = self.layer_config[layer_name]['zp'].numpy()
- data = ggml_quant(data, data_qtype.name.lower(), scale, zp)
- else:
- data_qtype = gguf.GGMLQuantizationType.F32
- else:
- data_qtype = gguf.GGMLQuantizationType.F32
+ try:
+ data = gguf.quants.quantize(data, data_qtype)
+ except gguf.QuantError as e:
+ logger.warning("%s, %s", e, "falling back to F16")
+ data_qtype = gguf.GGMLQuantizationType.F16
+ data = gguf.quants.quantize(data, data_qtype)
shape = gguf.quant_shape_from_byte_shape(
- data.shape, data_qtype) if data.dtype == np.uint8 else data.shape
+ data.shape,
+ data_qtype) if data.dtype == np.uint8 else data.shape
# reverse shape to make it similar to the internal ggml dimension order
shape_str = f"{{{', '.join(str(n) for n in reversed(shape))}}}"
-
# n_dims is implicit in the shape
logger.info(
- f"{f'%-{max_name_len}s' % f'{new_name},'} {old_dtype}"
- f" --> {data_qtype.name}, shape = {shape_str}")
+ f"{f'%-{max_name_len}s' % f'{new_name},'} {old_dtype} --> {data_qtype.name}, shape = {shape_str}"
+ )
- self.gguf_writer.add_tensor(new_name, data, raw_dtype=data_qtype)
+ self.gguf_writer.add_tensor(new_name,
+ data,
+ raw_dtype=data_qtype)
def set_type(self):
self.gguf_writer.add_type(gguf.GGUFType.MODEL)
def prepare_metadata(self, vocab_only: bool):
- total_params, shared_params, expert_params, expert_count = self.gguf_writer.get_total_parameter_count()
+ total_params, shared_params, expert_params, expert_count = self.gguf_writer.get_total_parameter_count(
+ )
- self.metadata = gguf.Metadata.load(self.metadata_override, self.dir_model_card, self.model_name, total_params)
+ self.metadata = gguf.Metadata.load(self.metadata_override,
+ self.dir_model_card,
+ self.model_name, total_params)
# Fallback to model directory name if metadata name is still missing
if self.metadata.name is None:
@@ -404,7 +493,10 @@ class Model:
# Generate parameter weight class (useful for leader boards) if not yet determined
if self.metadata.size_label is None and total_params > 0:
- self.metadata.size_label = gguf.size_label(total_params, shared_params, expert_params, expert_count)
+ self.metadata.size_label = gguf.size_label(total_params,
+ shared_params,
+ expert_params,
+ expert_count)
# Extract the encoding scheme from the file type name. e.g. 'gguf.LlamaFileType.MOSTLY_Q8_0' --> 'Q8_0'
output_type: str = self.ftype.name.partition("_")[2]
@@ -414,13 +506,22 @@ class Model:
# Generate default filename based on model specification and available metadata
if not vocab_only:
fname_default: str = gguf.naming_convention(
- self.metadata.name, self.metadata.basename, self.metadata.finetune,
- self.metadata.version, self.metadata.size_label, output_type,
+ self.metadata.name,
+ self.metadata.basename,
+ self.metadata.finetune,
+ self.metadata.version,
+ self.metadata.size_label,
+ output_type,
model_type="LoRA" if total_params < 0 else None)
else:
fname_default: str = gguf.naming_convention(
- self.metadata.name, self.metadata.basename, self.metadata.finetune,
- self.metadata.version, size_label=None, output_type=None, model_type="vocab")
+ self.metadata.name,
+ self.metadata.basename,
+ self.metadata.finetune,
+ self.metadata.version,
+ size_label=None,
+ output_type=None,
+ model_type="vocab")
# Use the default filename
self.fname_out = self.fname_out / f"{fname_default}.gguf"
@@ -430,7 +531,8 @@ class Model:
# file template strings as it doesn't actually exist as a file
# Process templated file name with the output ftype, useful with the "auto" ftype
- self.fname_out = self.fname_out.parent / gguf.fill_templated_filename(self.fname_out.name, output_type)
+ self.fname_out = self.fname_out.parent / gguf.fill_templated_filename(
+ self.fname_out.name, output_type)
self.set_type()
@@ -464,7 +566,8 @@ class Model:
self.gguf_writer.close()
@staticmethod
- def get_model_part_names(dir_model: Path, prefix: str, suffix: str) -> list[str]:
+ def get_model_part_names(dir_model: Path, prefix: str,
+ suffix: str) -> list[str]:
part_names: list[str] = []
for filename in os.listdir(dir_model):
if filename.startswith(prefix) and filename.endswith(suffix):
@@ -487,14 +590,21 @@ class Model:
for name in names:
cls._model_classes[name] = modelcls
return modelcls
+
return func
+ @classmethod
+ def print_registered_models(cls):
+ for name in sorted(cls._model_classes.keys()):
+ logger.error(f"- {name}")
+
@classmethod
def from_model_architecture(cls, arch: str) -> type[Model]:
try:
return cls._model_classes[arch]
except KeyError:
- raise NotImplementedError(f'Architecture {arch!r} not supported!') from None
+ raise NotImplementedError(
+ f'Architecture {arch!r} not supported!') from None
def does_token_look_special(self, token: str | bytes) -> bool:
if isinstance(token, (bytes, bytearray)):
@@ -508,14 +618,21 @@ class Model:
# (e.g. command-r, command-r-plus, deepseek-coder, gemma{,-2})
seems_special = token_text in (
"<pad>", # deepseek-coder
- "<mask>", "<2mass>", "[@BOS@]", # gemma{,-2}
+ "<mask>",
+ "<2mass>",
+ "[@BOS@]", # gemma{,-2}
)
- seems_special = seems_special or (token_text.startswith("<|") and token_text.endswith("|>"))
- seems_special = seems_special or (token_text.startswith("<|") and token_text.endswith("|>")) # deepseek-coder
+ seems_special = seems_special or (token_text.startswith("<|")
+ and token_text.endswith("|>"))
+ seems_special = seems_special or (token_text.startswith("<|")
+ and token_text.endswith("|>")
+ ) # deepseek-coder
# TODO: should these be marked as UNUSED instead? (maybe not)
- seems_special = seems_special or (token_text.startswith("<unused") and token_text.endswith(">")) # gemma{,-2}
+ seems_special = seems_special or (token_text.startswith("<unused")
+ and token_text.endswith(">")
+ ) # gemma{,-2}
return seems_special
@@ -531,7 +648,10 @@ class Model:
tokpre = self.get_vocab_base_pre(tokenizer)
- reverse_vocab = {id_: encoded_tok for encoded_tok, id_ in tokenizer.vocab.items()}
+ reverse_vocab = {
+ id_: encoded_tok
+ for encoded_tok, id_ in tokenizer.vocab.items()
+ }
added_vocab = tokenizer.get_added_vocab()
for i in range(vocab_size):
@@ -541,10 +661,27 @@ class Model:
else:
token: str = reverse_vocab[i]
if token in added_vocab:
- if tokenizer.added_tokens_decoder[i].special or self.does_token_look_special(token):
+ # The tokenizer in llama.cpp assumes the CONTROL and USER_DEFINED tokens are pre-normalized.
+ # To avoid unexpected issues - we make sure to normalize non-normalized tokens
+ if not tokenizer.added_tokens_decoder[i].normalized:
+ previous_token = token
+ token = tokenizer.decode(
+ tokenizer.encode(token, add_special_tokens=False))
+ if previous_token != token:
+ logger.info(
+ f"{repr(previous_token)} is encoded and decoded"
+ f" back to {repr(token)} using AutoTokenizer"
+ )
+
+ if tokenizer.added_tokens_decoder[
+ i].special or self.does_token_look_special(token):
toktypes.append(gguf.TokenType.CONTROL)
else:
- token = token.replace(b"\xe2\x96\x81".decode("utf-8"), " ") # pre-normalize user-defined spaces
+ # NOTE: this was added for Gemma.
+ # Encoding and decoding the tokens above isn't sufficient for this case.
+ token = token.replace(
+ b"\xe2\x96\x81".decode("utf-8"),
+ " ") # pre-normalize user-defined spaces
toktypes.append(gguf.TokenType.USER_DEFINED)
else:
toktypes.append(gguf.TokenType.NORMAL)
@@ -562,11 +699,7 @@ class Model:
# we will use this unique identifier to write a "tokenizer.ggml.pre" entry in the GGUF file which we can
# use in llama.cpp to implement the same pre-tokenizer
- chktxt = '\n \n\n \n\n\n \t \t\t \t\n \n \n \n \n🚀 (normal)' \
- +' 😶\u200d🌫️ (multiple emojis concatenated) ✅ 🦙🦙 3 33 333 3333 33333' \
- +' 333333 3333333 33333333 3.3 3..3 3...3 កាន់តែពិសេសអាច😁 ?我想在apple工作1314151天~' \
- +' ------======= нещо на Български \'\'\'\'\'\'```````""""......!!!!!!?????? I\'ve been' \
- +' \'told he\'s there, \'RE you sure? \'M not sure I\'ll make it, \'D you like some tea? We\'Ve a\'lL'
+ chktxt = ('\n \n\n \n\n\n \t \t\t \t\n \n \n \n \n🚀 (normal) 😶\u200d🌫️ (multiple emojis concatenated) ✅ 🦙🦙 3 33 333 3333 33333 333333 3333333 33333333 3.3 3..3 3...3 កាន់តែពិសេសអាច😁 ?我想在apple工作1314151天~ ------======= нещо на Български \'\'\'\'\'\'```````""""......!!!!!!?????? I\'ve been \'told he\'s there, \'RE you sure? \'M not sure I\'ll make it, \'D you like some tea? We\'Ve a\'lL') # pylint: disable=C0301
chktok = tokenizer.encode(chktxt)
chkhsh = sha256(str(chktok).encode()).hexdigest()
@@ -591,6 +724,9 @@ class Model:
if chkhsh == "8aeee3860c56296a157a1fe2fad249ec40aa59b1bb5709f4ade11c4e6fe652ed":
# ref: https://huggingface.co/tiiuae/falcon-7b
res = "falcon"
+ if chkhsh == "9d032fcbd5501f4a38150912590928bfb36091efb5df11b8e2124b0390e3fb1e":
+ # ref: https://huggingface.co/tiiuae/Falcon3-7B-Base
+ res = "falcon3"
if chkhsh == "0876d13b50744004aa9aeae05e7b0647eac9d801b5ba4668afc01e709c15e19f":
# ref: https://huggingface.co/BAAI/bge-small-en-v1.5
res = "bert-bge"
@@ -678,22 +814,51 @@ class Model:
if chkhsh == "60824e3c0d9401f89943cbb2fff727f0e2d4c545ba4df2d6e4f09a6db0f5b450":
# ref: https://huggingface.co/facebook/chameleon-7b
res = "chameleon"
+ if chkhsh == "1431a23e583c97432bc230bff598d103ddb5a1f89960c8f1d1051aaa944d0b35":
+ # ref: https://huggingface.co/sapienzanlp/Minerva-7B-base-v1.0
+ res = "minerva-7b"
+ if chkhsh == "8b5a93ed704057481f240da0be7e7dca721d7f8f4755263b6807227a2cbeae65":
+ # ref: https://huggingface.co/sentence-transformers/stsb-roberta-base
+ res = "roberta-bpe"
+ if chkhsh == "ad851be1dba641f2e3711822f816db2c265f788b37c63b4e1aeacb9ee92de8eb":
+ # ref: https://huggingface.co/ai-sage/GigaChat-20B-A3B-instruct
+ res = "gigachat"
+ if chkhsh == "d4c8f286ea6b520b3d495c4455483cfa2302c0cfcd4be05d781b6a8a0a7cdaf1":
+ # ref: https://huggingface.co/Infinigence/Megrez-3B-Instruct
+ res = "megrez"
+ if chkhsh == "877081d19cf6996e2c4ff0e1236341e9b7bde288f5311a56a937f0afbbb3aeb5":
+ # ref: https://huggingface.co/deepseek-ai/DeepSeek-V3
+ res = "deepseek-v3"
if res is None:
logger.warning("\n")
- logger.warning("**************************************************************************************")
- logger.warning("** WARNING: The BPE pre-tokenizer was not recognized!")
- logger.warning("** There are 2 possible reasons for this:")
- logger.warning("** - the model has not been added to convert_hf_to_gguf_update.py yet")
- logger.warning("** - the pre-tokenization config has changed upstream")
logger.warning(
- "** Check your model files and convert_hf_to_gguf_update.py and update them accordingly.")
- logger.warning("** ref: https://github.com/ggerganov/llama.cpp/pull/6920")
+ "**************************************************************************************"
+ )
+ logger.warning(
+ "** WARNING: The BPE pre-tokenizer was not recognized!")
+ logger.warning(
+ "** There are 2 possible reasons for this:")
+ logger.warning(
+ "** - the model has not been added to convert_hf_to_gguf_update.py yet"
+ )
+ logger.warning(
+ "** - the pre-tokenization config has changed upstream"
+ )
+ logger.warning(
+ "** Check your model files and convert_hf_to_gguf_update.py and update them accordingly."
+ )
+ logger.warning(
+ "** ref: https://github.com/ggerganov/llama.cpp/pull/6920")
logger.warning("**")
logger.warning(f"** chkhsh: {chkhsh}")
- logger.warning("**************************************************************************************")
+ logger.warning(
+ "**************************************************************************************"
+ )
logger.warning("\n")
- raise NotImplementedError("BPE pre-tokenizer was not recognized - update get_vocab_base_pre()")
+ raise NotImplementedError(
+ "BPE pre-tokenizer was not recognized - update get_vocab_base_pre()"
+ )
logger.debug(f"tokenizer.ggml.pre: {repr(res)}")
logger.debug(f"chkhsh: {chkhsh}")
@@ -701,6 +866,9 @@ class Model:
return res
# Marker: End get_vocab_base_pre
+ def _set_vocab_none(self) -> None:
+ self.gguf_writer.add_tokenizer_model("none")
+
def _set_vocab_gpt2(self) -> None:
tokens, toktypes, tokpre = self.get_vocab_base()
self.gguf_writer.add_tokenizer_model("gpt2")
@@ -718,7 +886,8 @@ class Model:
toktypes: list[int] = []
from transformers import AutoTokenizer
- tokenizer = AutoTokenizer.from_pretrained(dir_model, trust_remote_code=True)
+ tokenizer = AutoTokenizer.from_pretrained(dir_model,
+ trust_remote_code=True)
vocab_size = hparams["vocab_size"]
assert max(tokenizer.get_vocab().values()) < vocab_size
@@ -733,11 +902,18 @@ class Model:
continue
merged = QwenModel.bpe(mergeable_ranks, token, max_rank=rank)
assert len(merged) == 2
- merges.append(' '.join(map(QwenModel.token_bytes_to_string, merged)))
+ merges.append(' '.join(map(QwenModel.token_bytes_to_string,
+ merged)))
# for this kind of tokenizer, added_vocab is not a subset of vocab, so they need to be combined
added_vocab = tokenizer.special_tokens
- reverse_vocab = {id_ : encoded_tok for encoded_tok, id_ in {**vocab, **added_vocab}.items()}
+ reverse_vocab = {
+ id_: encoded_tok
+ for encoded_tok, id_ in {
+ **vocab,
+ **added_vocab
+ }.items()
+ }
for i in range(vocab_size):
if i not in reverse_vocab:
@@ -759,10 +935,13 @@ class Model:
special_vocab.merges = merges
# only add special tokens when they were not already loaded from config.json
if len(special_vocab.special_token_ids) == 0:
- special_vocab._set_special_token("bos", tokenizer.special_tokens["<|endoftext|>"])
- special_vocab._set_special_token("eos", tokenizer.special_tokens["<|endoftext|>"])
+ special_vocab._set_special_token(
+ "bos", tokenizer.special_tokens["<|endoftext|>"])
+ special_vocab._set_special_token(
+ "eos", tokenizer.special_tokens["<|endoftext|>"])
# this one is usually not in config.json anyway
- special_vocab._set_special_token("unk", tokenizer.special_tokens["<|endoftext|>"])
+ special_vocab._set_special_token(
+ "unk", tokenizer.special_tokens["<|endoftext|>"])
special_vocab.add_to_gguf(self.gguf_writer)
def _set_vocab_sentencepiece(self, add_to_gguf=True):
@@ -790,7 +969,9 @@ class Model:
vocab_size = self.hparams.get('vocab_size', tokenizer.vocab_size())
- tokens: list[bytes] = [f"[PAD{i}]".encode("utf-8") for i in range(vocab_size)]
+ tokens: list[bytes] = [
+ f"[PAD{i}]".encode("utf-8") for i in range(vocab_size)
+ ]
scores: list[float] = [-10000.0] * vocab_size
toktypes: list[int] = [SentencePieceTokenTypes.UNUSED] * vocab_size
@@ -820,7 +1001,9 @@ class Model:
for key in added_tokens_json:
token_id = added_tokens_json[key]
if token_id >= vocab_size:
- logger.warning(f'ignore token {token_id}: id is out of range, max={vocab_size - 1}')
+ logger.warning(
+ f'ignore token {token_id}: id is out of range, max={vocab_size - 1}'
+ )
continue
tokens[token_id] = key.encode("utf-8")
@@ -831,26 +1014,34 @@ class Model:
if tokenizer_config_file.is_file():
with open(tokenizer_config_file, "r", encoding="utf-8") as f:
tokenizer_config_json = json.load(f)
- added_tokens_decoder = tokenizer_config_json.get("added_tokens_decoder", {})
+ added_tokens_decoder = tokenizer_config_json.get(
+ "added_tokens_decoder", {})
for token_id, token_data in added_tokens_decoder.items():
token_id = int(token_id)
token: str = token_data["content"]
if toktypes[token_id] != SentencePieceTokenTypes.UNUSED:
if tokens[token_id] != token.encode("utf-8"):
logger.warning(
- f'replacing token {token_id}: {tokens[token_id].decode("utf-8")!r} -> {token!r}')
- if token_data.get("special") or self.does_token_look_special(token):
+ f'replacing token {token_id}: {tokens[token_id].decode("utf-8")!r} -> {token!r}'
+ )
+ if token_data.get(
+ "special") or self.does_token_look_special(token):
toktypes[token_id] = SentencePieceTokenTypes.CONTROL
else:
- token = token.replace(b"\xe2\x96\x81".decode("utf-8"), " ") # pre-normalize user-defined spaces
- toktypes[token_id] = SentencePieceTokenTypes.USER_DEFINED
+ token = token.replace(
+ b"\xe2\x96\x81".decode("utf-8"),
+ " ") # pre-normalize user-defined spaces
+ toktypes[
+ token_id] = SentencePieceTokenTypes.USER_DEFINED
scores[token_id] = -1000.0
tokens[token_id] = token.encode("utf-8")
if vocab_size > len(tokens):
pad_count = vocab_size - len(tokens)
- logger.debug(f"Padding vocab with {pad_count} token(s) - [PAD1] through [PAD{pad_count}]")
+ logger.debug(
+ f"Padding vocab with {pad_count} token(s) - [PAD1] through [PAD{pad_count}]"
+ )
for i in range(1, pad_count + 1):
tokens.append(bytes(f"[PAD{i}]", encoding="utf-8"))
scores.append(-1000.0)
@@ -880,52 +1071,262 @@ class Model:
special_vocab = gguf.SpecialVocab(self.dir_model, n_vocab=len(tokens))
special_vocab.add_to_gguf(self.gguf_writer)
- def _set_vocab_builtin(self, model_name: Literal["gpt-neox", "llama-spm"], vocab_size: int):
- tokenizer_path = Path(sys.path[0]) / "models" / f"ggml-vocab-{model_name}.gguf"
- logger.warning(f"Using tokenizer from '{os.path.relpath(tokenizer_path, os.getcwd())}'")
+ def _set_vocab_builtin(self, model_name: Literal["gpt-neox", "llama-spm"],
+ vocab_size: int):
+ tokenizer_path = Path(
+ sys.path[0]) / "models" / f"ggml-vocab-{model_name}.gguf"
+ logger.warning(
+ f"Using tokenizer from '{os.path.relpath(tokenizer_path, os.getcwd())}'"
+ )
vocab_reader = gguf.GGUFReader(tokenizer_path, "r")
default_pre = "mpt" if model_name == "gpt-neox" else "default"
field = vocab_reader.get_field(gguf.Keys.Tokenizer.MODEL)
assert field # tokenizer model
- self.gguf_writer.add_tokenizer_model(bytes(field.parts[-1]).decode("utf-8"))
+ self.gguf_writer.add_tokenizer_model(
+ bytes(field.parts[-1]).decode("utf-8"))
field = vocab_reader.get_field(gguf.Keys.Tokenizer.PRE)
- self.gguf_writer.add_tokenizer_pre(bytes(field.parts[-1]).decode("utf-8") if field else default_pre)
+ self.gguf_writer.add_tokenizer_pre(
+ bytes(field.parts[-1]).decode("utf-8") if field else default_pre)
field = vocab_reader.get_field(gguf.Keys.Tokenizer.LIST)
assert field # token list
- self.gguf_writer.add_token_list([bytes(field.parts[i]) for i in field.data][:vocab_size])
+ self.gguf_writer.add_token_list(
+ [bytes(field.parts[i]) for i in field.data][:vocab_size])
if model_name == "llama-spm":
field = vocab_reader.get_field(gguf.Keys.Tokenizer.SCORES)
assert field # token scores
- self.gguf_writer.add_token_scores([field.parts[i].tolist()[0] for i in field.data][:vocab_size])
+ self.gguf_writer.add_token_scores(
+ [field.parts[i].tolist()[0] for i in field.data][:vocab_size])
field = vocab_reader.get_field(gguf.Keys.Tokenizer.TOKEN_TYPE)
assert field # token types
- self.gguf_writer.add_token_types([field.parts[i].tolist()[0] for i in field.data][:vocab_size])
+ self.gguf_writer.add_token_types(
+ [field.parts[i].tolist()[0] for i in field.data][:vocab_size])
if model_name != "llama-spm":
field = vocab_reader.get_field(gguf.Keys.Tokenizer.MERGES)
assert field # token merges
- self.gguf_writer.add_token_merges([bytes(field.parts[i]) for i in field.data])
+ self.gguf_writer.add_token_merges(
+ [bytes(field.parts[i]) for i in field.data])
- if (field := vocab_reader.get_field(gguf.Keys.Tokenizer.BOS_ID)) is not None:
+ if (field := vocab_reader.get_field(
+ gguf.Keys.Tokenizer.BOS_ID)) is not None:
self.gguf_writer.add_bos_token_id(field.parts[-1].tolist()[0])
- if (field := vocab_reader.get_field(gguf.Keys.Tokenizer.EOS_ID)) is not None:
+ if (field := vocab_reader.get_field(
+ gguf.Keys.Tokenizer.EOS_ID)) is not None:
self.gguf_writer.add_eos_token_id(field.parts[-1].tolist()[0])
- if (field := vocab_reader.get_field(gguf.Keys.Tokenizer.UNK_ID)) is not None:
+ if (field := vocab_reader.get_field(
+ gguf.Keys.Tokenizer.UNK_ID)) is not None:
self.gguf_writer.add_unk_token_id(field.parts[-1].tolist()[0])
- if (field := vocab_reader.get_field(gguf.Keys.Tokenizer.PAD_ID)) is not None:
+ if (field := vocab_reader.get_field(
+ gguf.Keys.Tokenizer.PAD_ID)) is not None:
self.gguf_writer.add_pad_token_id(field.parts[-1].tolist()[0])
- if (field := vocab_reader.get_field(gguf.Keys.Tokenizer.ADD_BOS)) is not None:
+ if (field := vocab_reader.get_field(
+ gguf.Keys.Tokenizer.ADD_BOS)) is not None:
self.gguf_writer.add_add_bos_token(field.parts[-1].tolist()[0])
- if (field := vocab_reader.get_field(gguf.Keys.Tokenizer.ADD_EOS)) is not None:
+ if (field := vocab_reader.get_field(
+ gguf.Keys.Tokenizer.ADD_EOS)) is not None:
self.gguf_writer.add_add_eos_token(field.parts[-1].tolist()[0])
+class Model(OriModel):
+
+ def __init__(self,
+ model,
+ dir_model: Path,
+ ftype: gguf.LlamaFileType,
+ fname_out: Path,
+ is_big_endian: bool = False,
+ layer_config: dict = {},
+ use_temp_file: bool = False,
+ eager: bool = False,
+ metadata_override: Path | None = None,
+ model_name: str | None = None,
+ split_max_tensors: int = 0,
+ split_max_size: int = 0,
+ dry_run: bool = False,
+ small_first_shard: bool = False,
+ hparams: dict[str, Any] | None = None):
+ self.model = model
+ self.layer_config = layer_config
+ super().__init__(dir_model=dir_model,
+ ftype=ftype,
+ fname_out=fname_out,
+ is_big_endian=is_big_endian,
+ use_temp_file=use_temp_file,
+ eager=eager,
+ metadata_override=metadata_override,
+ model_name=model_name,
+ split_max_tensors=split_max_tensors,
+ split_max_size=split_max_size,
+ dry_run=dry_run,
+ small_first_shard=small_first_shard,
+ hparams=hparams)
+
+ @classmethod
+ def __init_subclass__(cls):
+ # can't use an abstract property, because overriding it without type errors
+ # would require using decorated functions instead of simply defining the property
+ if "model_arch" not in cls.__dict__:
+ raise TypeError(
+ f"Missing property 'model_arch' for {cls.__name__!r}")
+
+ def get_tensors(self) -> Iterator[tuple[str, Tensor]]:
+ for name, tensor in self.model.named_parameters():
+ yield name, tensor
+
+ def prepare_tensors(self):
+ max_name_len = max(len(s) for _, s in self.tensor_map.mapping.values()) + len(".weight,")
+
+ for name, data_torch in chain(self.generate_extra_tensors(), self.get_tensors()):
+ # we don't need these
+ if name.endswith((".attention.masked_bias", ".attention.bias", ".rotary_emb.inv_freq")):
+ continue
+
+ old_dtype = data_torch.dtype
+
+ # convert any unsupported data types to float32
+ if data_torch.dtype not in (torch.float16, torch.float32):
+ data_torch = data_torch.to(torch.float32)
+
+ # use the first number-like part of the tensor name as the block id
+ bid = None
+ for part in name.split("."):
+ if part.isdecimal():
+ bid = int(part)
+ break
+
+ for new_name, data_torch in (self.modify_tensors(data_torch, name, bid)):
+ data = data_torch.squeeze().cpu().numpy()
+
+ # if data ends up empty, it means data_torch was a scalar tensor -> restore
+ if len(data.shape) == 0:
+ data = data_torch.numpy()
+
+ n_dims = len(data.shape)
+ data_qtype: gguf.GGMLQuantizationType | bool = self.tensor_force_quant(name, new_name, bid, n_dims)
+
+ # Most of the codebase that takes in 1D tensors or norms only handles F32 tensors
+ if n_dims <= 1 or new_name.endswith("_norm.weight"):
+ data_qtype = gguf.GGMLQuantizationType.F32
+
+ # Conditions should closely match those in llama_model_quantize_internal in llama.cpp
+ # Some tensor types are always in float32
+ if data_qtype is False and (any(
+ self.match_model_tensor_name(new_name, key, bid)
+ for key in (
+ gguf.MODEL_TENSOR.FFN_GATE_INP,
+ gguf.MODEL_TENSOR.POS_EMBD,
+ gguf.MODEL_TENSOR.TOKEN_TYPES,
+ gguf.MODEL_TENSOR.SSM_CONV1D,
+ gguf.MODEL_TENSOR.TIME_MIX_FIRST,
+ gguf.MODEL_TENSOR.TIME_MIX_W1,
+ gguf.MODEL_TENSOR.TIME_MIX_W2,
+ gguf.MODEL_TENSOR.TIME_MIX_DECAY_W1,
+ gguf.MODEL_TENSOR.TIME_MIX_DECAY_W2,
+ )) or not new_name.endswith(".weight")):
+ data_qtype = gguf.GGMLQuantizationType.F32
+
+ if data_qtype is False and any(
+ self.match_model_tensor_name(new_name, key, bid)
+ for key in (
+ gguf.MODEL_TENSOR.TOKEN_EMBD,
+ gguf.MODEL_TENSOR.OUTPUT,
+ )):
+ if self.ftype in (
+ gguf.LlamaFileType.MOSTLY_TQ1_0,
+ gguf.LlamaFileType.MOSTLY_TQ2_0,
+ ):
+ # TODO: use Q4_K and Q6_K
+ data_qtype = gguf.GGMLQuantizationType.F16
+ # data_qtype = gguf.GGMLQuantizationType.Q8_0 # llama.cpp:llama_tensor_get_type
+
+ # No override (data_qtype is False), or wants to be quantized (data_qtype is True)
+ if isinstance(data_qtype, bool):
+ if self.ftype == gguf.LlamaFileType.ALL_F32:
+ data_qtype = gguf.GGMLQuantizationType.F32
+ elif self.ftype == gguf.LlamaFileType.MOSTLY_F16:
+ data_qtype = gguf.GGMLQuantizationType.F16
+ elif self.ftype == gguf.LlamaFileType.MOSTLY_BF16:
+ data_qtype = gguf.GGMLQuantizationType.BF16
+ # elif self.ftype == gguf.LlamaFileType.MOSTLY_Q8_0:
+ # data_qtype = gguf.GGMLQuantizationType.Q8_0
+ # elif self.ftype == gguf.LlamaFileType.MOSTLY_TQ1_0:
+ # data_qtype = gguf.GGMLQuantizationType.TQ1_0
+ # elif self.ftype == gguf.LlamaFileType.MOSTLY_TQ2_0:
+ # data_qtype = gguf.GGMLQuantizationType.TQ2_0
+ elif self.ftype == gguf.LlamaFileType.MOSTLY_Q4_0:
+ data_qtype = gguf.GGMLQuantizationType.Q4_0
+ elif self.ftype == gguf.LlamaFileType.MOSTLY_Q4_1:
+ data_qtype = gguf.GGMLQuantizationType.Q4_1
+ else:
+ raise ValueError(
+ f"Unknown file type: {self.ftype.name}")
+
+ # try:
+ # data = gguf.quants.quantize(data, data_qtype)
+ # except gguf.QuantError as e:
+ # logger.warning("%s, %s", e, "falling back to F16")
+ # data_qtype = gguf.GGMLQuantizationType.F16
+ # data = gguf.quants.quantize(data, data_qtype)
+
+ # Quant here
+ def _quant_data(data, data_qtype):
+ suffix = '.weight'
+ if suffix in name and name[:-len(suffix)] in self.layer_config:
+ layer_name = name[:-len(suffix)]
+ if "scale" in self.layer_config[layer_name]:
+ scale = self.layer_config[layer_name]['scale']
+ if isinstance(scale, torch.Tensor):
+ scale = scale.numpy()
+ zp = self.layer_config[layer_name]['zp']
+ if isinstance(zp, torch.Tensor):
+ zp = zp.numpy()
+ data = ggml_quant(data, data_qtype.name.lower(), scale,
+ zp)
+ else:
+ data_qtype = gguf.GGMLQuantizationType.F32
+ else:
+ data_qtype = gguf.GGMLQuantizationType.F32
+ return data, data_qtype
+
+ # for MOE model
+ if len(data.shape) == 3:
+ new_data = []
+ for idx, arr in enumerate(data):
+ arr_name = name.split('.')
+ for i in range(len(arr_name)-1, -1, -1):
+ if arr_name[i].isdecimal() and int(arr_name[i]) == (data.shape[0] - 1):
+ arr_name[i] = str(idx)
+ arr_name = ".".join(arr_name)
+ arr, data_qtype = _quant_data(arr, data_qtype)
+ new_data.append(arr)
+ data = np.array(new_data)
+ del new_data
+ else:
+ data, data_qtype = _quant_data(data, data_qtype)
+
+ shape = gguf.quant_shape_from_byte_shape(
+ data.shape,
+ data_qtype) if data.dtype == np.uint8 else data.shape
+
+ # reverse shape to make it similar to the internal ggml dimension order
+ shape_str = f"{{{', '.join(str(n) for n in reversed(shape))}}}"
+
+ # n_dims is implicit in the shape
+ logger.info(
+ f"{f'%-{max_name_len}s' % f'{new_name},'} {old_dtype}"
+ f" --> {data_qtype.name}, shape = {shape_str}")
+
+ self.gguf_writer.add_tensor(new_name,
+ data,
+ raw_dtype=data_qtype)
+
+
@Model.register("GPTNeoXForCausalLM")
class GPTNeoXModel(Model):
model_arch = gguf.MODEL_ARCH.GPTNEOX
@@ -933,27 +1334,36 @@ class GPTNeoXModel(Model):
def set_gguf_parameters(self):
block_count = self.hparams["num_hidden_layers"]
- self.gguf_writer.add_context_length(self.hparams["max_position_embeddings"])
+ self.gguf_writer.add_context_length(
+ self.hparams["max_position_embeddings"])
self.gguf_writer.add_embedding_length(self.hparams["hidden_size"])
self.gguf_writer.add_block_count(block_count)
- self.gguf_writer.add_feed_forward_length(self.hparams["intermediate_size"])
+ self.gguf_writer.add_feed_forward_length(
+ self.hparams["intermediate_size"])
self.gguf_writer.add_rope_dimension_count(
- int(self.hparams["rotary_pct"] * (self.hparams["hidden_size"] // self.hparams["num_attention_heads"])),
- )
+ int(self.hparams["rotary_pct"] *
+ (self.hparams["hidden_size"] //
+ self.hparams["num_attention_heads"])), )
self.gguf_writer.add_head_count(self.hparams["num_attention_heads"])
- self.gguf_writer.add_parallel_residual(self.hparams.get("use_parallel_residual", True))
+ self.gguf_writer.add_parallel_residual(
+ self.hparams.get("use_parallel_residual", True))
self.gguf_writer.add_layer_norm_eps(self.hparams["layer_norm_eps"])
- def modify_tensors(self, data_torch: Tensor, name: str, bid: int | None) -> Iterable[tuple[str, Tensor]]:
+ def modify_tensors(self, data_torch: Tensor, name: str,
+ bid: int | None) -> Iterable[tuple[str, Tensor]]:
del bid # unused
- n_head = self.hparams.get("n_head", self.hparams.get("num_attention_heads"))
+ n_head = self.hparams.get("n_head",
+ self.hparams.get("num_attention_heads"))
n_embed = self.hparams.get("hidden_size", self.hparams.get("n_embed"))
tensors: list[tuple[str, Tensor]] = []
- if re.match(r"gpt_neox\.layers\.\d+\.attention\.query_key_value\.weight", name):
- qkv_weights = data_torch.reshape((n_head, 3, n_embed // n_head, n_embed))
+ if re.match(
+ r"gpt_neox\.layers\.\d+\.attention\.query_key_value\.weight",
+ name):
+ qkv_weights = data_torch.reshape(
+ (n_head, 3, n_embed // n_head, n_embed))
data_torch = torch.cat(
(
qkv_weights[:, 0, :, :].reshape((-1, n_embed)),
@@ -963,13 +1373,15 @@ class GPTNeoXModel(Model):
dim=0,
)
logger.info("re-format attention.linear_qkv.weight")
- elif re.match(r"gpt_neox\.layers\.\d+\.attention\.query_key_value\.bias", name):
+ elif re.match(
+ r"gpt_neox\.layers\.\d+\.attention\.query_key_value\.bias",
+ name):
qkv_bias = data_torch.reshape((n_head, 3, n_embed // n_head))
data_torch = torch.cat(
(
- qkv_bias[:, 0, :].reshape((n_embed,)),
- qkv_bias[:, 1, :].reshape((n_embed,)),
- qkv_bias[:, 2, :].reshape((n_embed,)),
+ qkv_bias[:, 0, :].reshape((n_embed, )),
+ qkv_bias[:, 1, :].reshape((n_embed, )),
+ qkv_bias[:, 2, :].reshape((n_embed, )),
),
dim=0,
)
@@ -986,8 +1398,10 @@ class BloomModel(Model):
def set_gguf_parameters(self):
n_embed = self.hparams.get("hidden_size", self.hparams.get("n_embed"))
- n_head = self.hparams.get("n_head", self.hparams.get("num_attention_heads"))
- self.gguf_writer.add_context_length(self.hparams.get("seq_length", n_embed))
+ n_head = self.hparams.get("n_head",
+ self.hparams.get("num_attention_heads"))
+ self.gguf_writer.add_context_length(
+ self.hparams.get("seq_length", n_embed))
self.gguf_writer.add_embedding_length(n_embed)
self.gguf_writer.add_feed_forward_length(4 * n_embed)
self.gguf_writer.add_block_count(self.hparams["n_layer"])
@@ -996,10 +1410,12 @@ class BloomModel(Model):
self.gguf_writer.add_layer_norm_eps(self.hparams["layer_norm_epsilon"])
self.gguf_writer.add_file_type(self.ftype)
- def modify_tensors(self, data_torch: Tensor, name: str, bid: int | None) -> Iterable[tuple[str, Tensor]]:
+ def modify_tensors(self, data_torch: Tensor, name: str,
+ bid: int | None) -> Iterable[tuple[str, Tensor]]:
del bid # unused
- n_head = self.hparams.get("n_head", self.hparams.get("num_attention_heads"))
+ n_head = self.hparams.get("n_head",
+ self.hparams.get("num_attention_heads"))
n_embed = self.hparams.get("hidden_size", self.hparams.get("n_embed"))
name = re.sub(r'transformer\.', '', name)
@@ -1007,7 +1423,8 @@ class BloomModel(Model):
tensors: list[tuple[str, Tensor]] = []
if re.match(r"h\.\d+\.self_attention\.query_key_value\.weight", name):
- qkv_weights = data_torch.reshape((n_head, 3, n_embed // n_head, n_embed))
+ qkv_weights = data_torch.reshape(
+ (n_head, 3, n_embed // n_head, n_embed))
data_torch = torch.cat(
(
qkv_weights[:, 0, :, :].reshape((-1, n_embed)),
@@ -1021,9 +1438,9 @@ class BloomModel(Model):
qkv_bias = data_torch.reshape((n_head, 3, n_embed // n_head))
data_torch = torch.cat(
(
- qkv_bias[:, 0, :].reshape((n_embed,)),
- qkv_bias[:, 1, :].reshape((n_embed,)),
- qkv_bias[:, 2, :].reshape((n_embed,)),
+ qkv_bias[:, 0, :].reshape((n_embed, )),
+ qkv_bias[:, 1, :].reshape((n_embed, )),
+ qkv_bias[:, 2, :].reshape((n_embed, )),
),
dim=0,
)
@@ -1035,8 +1452,11 @@ class BloomModel(Model):
assert self.tensor_names is not None
# TODO: tie them at runtime, don't duplicate in the model file
- if all(s not in self.tensor_names for s in ("lm_head.weight", "output.weight")): # pylint: disable=E1135
- tensors.append((self.format_tensor_name(gguf.MODEL_TENSOR.OUTPUT), data_torch))
+ if all(s not in self.tensor_names
+ for s in ("lm_head.weight", "output.weight")):
+ tensors.append(
+ (self.format_tensor_name(gguf.MODEL_TENSOR.OUTPUT),
+ data_torch))
return tensors
@@ -1067,20 +1487,26 @@ class MPTModel(Model):
self.gguf_writer.add_head_count_kv(kv_n_heads)
self.gguf_writer.add_layer_norm_eps(1e-5)
if self.hparams["attn_config"]["clip_qkv"] is not None:
- self.gguf_writer.add_clamp_kqv(self.hparams["attn_config"]["clip_qkv"])
+ self.gguf_writer.add_clamp_kqv(
+ self.hparams["attn_config"]["clip_qkv"])
if self.hparams["attn_config"]["alibi"]:
- self.gguf_writer.add_max_alibi_bias(self.hparams["attn_config"]["alibi_bias_max"])
+ self.gguf_writer.add_max_alibi_bias(
+ self.hparams["attn_config"]["alibi_bias_max"])
else:
self.gguf_writer.add_max_alibi_bias(0.0)
- def modify_tensors(self, data_torch: Tensor, name: str, bid: int | None) -> Iterable[tuple[str, Tensor]]:
+ def modify_tensors(self, data_torch: Tensor, name: str,
+ bid: int | None) -> Iterable[tuple[str, Tensor]]:
del bid # unused
if "scales" in name:
- new_name = self.map_tensor_name(name, try_suffixes=(".weight", ".bias", ".scales"))
+ new_name = self.map_tensor_name(name,
+ try_suffixes=(".weight", ".bias",
+ ".scales"))
new_name = new_name.replace("scales", "act.scales")
else:
- new_name = self.map_tensor_name(name, try_suffixes=(".weight", ".bias"))
+ new_name = self.map_tensor_name(name,
+ try_suffixes=(".weight", ".bias"))
return [(new_name, data_torch)]
@@ -1112,9 +1538,11 @@ class OrionModel(Model):
self.gguf_writer.add_context_length(ctx_length)
self.gguf_writer.add_embedding_length(self.hparams["hidden_size"])
self.gguf_writer.add_block_count(block_count)
- self.gguf_writer.add_feed_forward_length(self.hparams["intermediate_size"])
+ self.gguf_writer.add_feed_forward_length(
+ self.hparams["intermediate_size"])
self.gguf_writer.add_head_count(head_count)
self.gguf_writer.add_head_count_kv(head_count_kv)
+ # note: config provides rms norm but it is actually layer norm
self.gguf_writer.add_layer_norm_eps(self.hparams["rms_norm_eps"])
@@ -1144,19 +1572,26 @@ class BaichuanModel(Model):
self.gguf_writer.add_context_length(ctx_length)
self.gguf_writer.add_embedding_length(self.hparams["hidden_size"])
self.gguf_writer.add_block_count(block_count)
- self.gguf_writer.add_feed_forward_length(self.hparams["intermediate_size"])
- self.gguf_writer.add_rope_dimension_count(self.hparams["hidden_size"] // self.hparams["num_attention_heads"])
+ self.gguf_writer.add_feed_forward_length(
+ self.hparams["intermediate_size"])
+ self.gguf_writer.add_rope_dimension_count(
+ self.hparams["hidden_size"] // self.hparams["num_attention_heads"])
self.gguf_writer.add_head_count(head_count)
self.gguf_writer.add_head_count_kv(head_count_kv)
self.gguf_writer.add_layer_norm_rms_eps(self.hparams["rms_norm_eps"])
self.gguf_writer.add_file_type(self.ftype)
- if self.hparams.get("rope_scaling") is not None and "factor" in self.hparams["rope_scaling"]:
+ if self.hparams.get(
+ "rope_scaling"
+ ) is not None and "factor" in self.hparams["rope_scaling"]:
if self.hparams["rope_scaling"].get("type") == "linear":
- self.gguf_writer.add_rope_scaling_type(gguf.RopeScalingType.LINEAR)
- self.gguf_writer.add_rope_scaling_factor(self.hparams["rope_scaling"]["factor"])
+ self.gguf_writer.add_rope_scaling_type(
+ gguf.RopeScalingType.LINEAR)
+ self.gguf_writer.add_rope_scaling_factor(
+ self.hparams["rope_scaling"]["factor"])
- def modify_tensors(self, data_torch: Tensor, name: str, bid: int | None) -> Iterable[tuple[str, Tensor]]:
+ def modify_tensors(self, data_torch: Tensor, name: str,
+ bid: int | None) -> Iterable[tuple[str, Tensor]]:
head_count = self.hparams["num_attention_heads"]
head_count_kv = self.hparams.get("num_key_value_heads", head_count)
@@ -1166,32 +1601,40 @@ class BaichuanModel(Model):
logger.info(f"Unpacking and permuting layer {bid}")
tensors = [
(self.format_tensor_name(gguf.MODEL_TENSOR.ATTN_Q, bid),
- self._reverse_hf_permute_part(data_torch, 0, head_count, head_count)),
+ self._reverse_hf_permute_part(data_torch, 0, head_count,
+ head_count)),
(self.format_tensor_name(gguf.MODEL_TENSOR.ATTN_K, bid),
- self._reverse_hf_permute_part(data_torch, 1, head_count, head_count_kv)),
+ self._reverse_hf_permute_part(data_torch, 1, head_count,
+ head_count_kv)),
(self.format_tensor_name(gguf.MODEL_TENSOR.ATTN_V, bid),
- self._reverse_hf_part(data_torch, 2)),
+ self._reverse_hf_part(data_torch, 2)),
]
else:
tensors = [(self.map_tensor_name(name), data_torch)]
return tensors
- def _reverse_hf_permute(self, weights: Tensor, n_head: int, n_kv_head: int | None = None) -> Tensor:
+ def _reverse_hf_permute(self,
+ weights: Tensor,
+ n_head: int,
+ n_kv_head: int | None = None) -> Tensor:
if n_kv_head is not None and n_head != n_kv_head:
n_head //= n_kv_head
- return (
- weights.reshape(n_head, 2, weights.shape[0] // n_head // 2, *weights.shape[1:])
- .swapaxes(1, 2)
- .reshape(weights.shape)
- )
+ return (weights.reshape(n_head, 2, weights.shape[0] // n_head // 2,
+ *weights.shape[1:]).swapaxes(1, 2).reshape(
+ weights.shape))
def _reverse_hf_permute_part(
- self, weights: Tensor, n_part: int, n_head: int, n_head_kv: int | None = None,
+ self,
+ weights: Tensor,
+ n_part: int,
+ n_head: int,
+ n_head_kv: int | None = None,
) -> Tensor:
r = weights.shape[0] // 3
- return self._reverse_hf_permute(weights[r * n_part:r * n_part + r, ...], n_head, n_head_kv)
+ return self._reverse_hf_permute(
+ weights[r * n_part:r * n_part + r, ...], n_head, n_head_kv)
def _reverse_hf_part(self, weights: Tensor, n_part: int) -> Tensor:
r = weights.shape[0] // 3
@@ -1219,7 +1662,10 @@ class XverseModel(Model):
if max_vocab_index >= vocab_size:
raise ValueError("Vocabulary size exceeds expected maximum size.")
- reverse_vocab: dict[int, str] = {id_: encoded_tok for encoded_tok, id_ in tokenizer.vocab.items()}
+ reverse_vocab: dict[int, str] = {
+ id_: encoded_tok
+ for encoded_tok, id_ in tokenizer.vocab.items()
+ }
added_vocab = tokenizer.get_added_vocab()
for token_id in range(vocab_size):
@@ -1268,19 +1714,26 @@ class XverseModel(Model):
self.gguf_writer.add_context_length(ctx_length)
self.gguf_writer.add_embedding_length(self.hparams["hidden_size"])
self.gguf_writer.add_block_count(block_count)
- self.gguf_writer.add_feed_forward_length(self.hparams["intermediate_size"])
- self.gguf_writer.add_rope_dimension_count(self.hparams["hidden_size"] // self.hparams["num_attention_heads"])
+ self.gguf_writer.add_feed_forward_length(
+ self.hparams["intermediate_size"])
+ self.gguf_writer.add_rope_dimension_count(
+ self.hparams["hidden_size"] // self.hparams["num_attention_heads"])
self.gguf_writer.add_head_count(head_count)
self.gguf_writer.add_head_count_kv(head_count_kv)
self.gguf_writer.add_layer_norm_rms_eps(self.hparams["rms_norm_eps"])
self.gguf_writer.add_file_type(self.ftype)
- if self.hparams.get("rope_scaling") is not None and "factor" in self.hparams["rope_scaling"]:
+ if self.hparams.get(
+ "rope_scaling"
+ ) is not None and "factor" in self.hparams["rope_scaling"]:
if self.hparams["rope_scaling"].get("type") == "linear":
- self.gguf_writer.add_rope_scaling_type(gguf.RopeScalingType.LINEAR)
- self.gguf_writer.add_rope_scaling_factor(self.hparams["rope_scaling"]["factor"])
+ self.gguf_writer.add_rope_scaling_type(
+ gguf.RopeScalingType.LINEAR)
+ self.gguf_writer.add_rope_scaling_factor(
+ self.hparams["rope_scaling"]["factor"])
- def modify_tensors(self, data_torch: Tensor, name: str, bid: int | None) -> Iterable[tuple[str, Tensor]]:
+ def modify_tensors(self, data_torch: Tensor, name: str,
+ bid: int | None) -> Iterable[tuple[str, Tensor]]:
del bid # unused
head_count = self.hparams["num_attention_heads"]
@@ -1288,21 +1741,24 @@ class XverseModel(Model):
# HF models permute some of the tensors, so we need to undo that
if name.endswith("q_proj.weight"):
- data_torch = self._reverse_hf_permute(data_torch, head_count, head_count)
+ data_torch = self._reverse_hf_permute(data_torch, head_count,
+ head_count)
if name.endswith("k_proj.weight"):
- data_torch = self._reverse_hf_permute(data_torch, head_count, head_count_kv)
+ data_torch = self._reverse_hf_permute(data_torch, head_count,
+ head_count_kv)
return [(self.map_tensor_name(name), data_torch)]
- def _reverse_hf_permute(self, weights: Tensor, n_head: int, n_kv_head: int | None = None) -> Tensor:
+ def _reverse_hf_permute(self,
+ weights: Tensor,
+ n_head: int,
+ n_kv_head: int | None = None) -> Tensor:
if n_kv_head is not None and n_head != n_kv_head:
n_head //= n_kv_head
- return (
- weights.reshape(n_head, 2, weights.shape[0] // n_head // 2, *weights.shape[1:])
- .swapaxes(1, 2)
- .reshape(weights.shape)
- )
+ return (weights.reshape(n_head, 2, weights.shape[0] // n_head // 2,
+ *weights.shape[1:]).swapaxes(1, 2).reshape(
+ weights.shape))
@Model.register("FalconForCausalLM", "RWForCausalLM")
@@ -1323,16 +1779,19 @@ class FalconModel(Model):
n_head_kv = self.hparams.get("n_head_kv", 1) # old name
self.gguf_writer.add_context_length(2048) # not in config.json
- self.gguf_writer.add_tensor_data_layout("jploski") # qkv tensor transform
+ self.gguf_writer.add_tensor_data_layout(
+ "jploski") # qkv tensor transform
self.gguf_writer.add_embedding_length(self.hparams["hidden_size"])
- self.gguf_writer.add_feed_forward_length(4 * self.hparams["hidden_size"])
+ self.gguf_writer.add_feed_forward_length(4 *
+ self.hparams["hidden_size"])
self.gguf_writer.add_block_count(block_count)
self.gguf_writer.add_head_count(n_head)
self.gguf_writer.add_head_count_kv(n_head_kv)
self.gguf_writer.add_layer_norm_eps(self.hparams["layer_norm_epsilon"])
self.gguf_writer.add_file_type(self.ftype)
- def modify_tensors(self, data_torch: Tensor, name: str, bid: int | None) -> Iterable[tuple[str, Tensor]]:
+ def modify_tensors(self, data_torch: Tensor, name: str,
+ bid: int | None) -> Iterable[tuple[str, Tensor]]:
del bid # unused
# QKV tensor transform
@@ -1347,10 +1806,12 @@ class FalconModel(Model):
if "query_key_value" in name:
n_head = self.find_hparam(["num_attention_heads", "n_head"])
- n_head_kv = self.find_hparam(["num_kv_heads", "n_head_kv"], optional=True) or 1
+ n_head_kv = self.find_hparam(["num_kv_heads", "n_head_kv"],
+ optional=True) or 1
head_dim = self.hparams["hidden_size"] // n_head
- qkv = data_torch.view(n_head_kv, n_head // n_head_kv + 2, head_dim, head_dim * n_head)
+ qkv = data_torch.view(n_head_kv, n_head // n_head_kv + 2, head_dim,
+ head_dim * n_head)
q = qkv[:, :-2].reshape(n_head * head_dim, head_dim * n_head)
k = qkv[:, [-2]].reshape(n_head_kv * head_dim, head_dim * n_head)
v = qkv[:, [-1]].reshape(n_head_kv * head_dim, head_dim * n_head)
@@ -1384,8 +1845,10 @@ class RefactModel(Model):
super().set_vocab()
# TODO: how to determine special FIM tokens automatically?
- special_vocab = gguf.SpecialVocab(self.dir_model, load_merges=False,
- special_token_types = ['prefix', 'suffix', 'middle', 'eot'])
+ special_vocab = gguf.SpecialVocab(
+ self.dir_model,
+ load_merges=False,
+ special_token_types=['prefix', 'suffix', 'middle', 'eot'])
special_vocab._set_special_token("prefix", 1)
special_vocab._set_special_token("suffix", 3)
special_vocab._set_special_token("middle", 2)
@@ -1409,10 +1872,12 @@ class RefactModel(Model):
self.gguf_writer.add_block_count(block_count)
self.gguf_writer.add_head_count(self.hparams["n_head"])
self.gguf_writer.add_head_count_kv(1)
- self.gguf_writer.add_layer_norm_rms_eps(self.hparams["layer_norm_epsilon"])
+ self.gguf_writer.add_layer_norm_rms_eps(
+ self.hparams["layer_norm_epsilon"])
self.gguf_writer.add_file_type(self.ftype)
- def modify_tensors(self, data_torch: Tensor, name: str, bid: int | None) -> Iterable[tuple[str, Tensor]]:
+ def modify_tensors(self, data_torch: Tensor, name: str,
+ bid: int | None) -> Iterable[tuple[str, Tensor]]:
hidden_dim = self.hparams["n_embd"]
inner_dim = 4 * hidden_dim
hidden_dim = int(2 * inner_dim / 3)
@@ -1426,15 +1891,23 @@ class RefactModel(Model):
if bid is not None:
if name == f"transformer.h.{bid}.attn.kv.weight":
- tensors.append((self.format_tensor_name(gguf.MODEL_TENSOR.ATTN_K, bid),
- data_torch[:n_head_kv * head_dim]))
- tensors.append((self.format_tensor_name(gguf.MODEL_TENSOR.ATTN_V, bid),
- data_torch[n_head_kv * head_dim:]))
+ tensors.append(
+ (self.format_tensor_name(gguf.MODEL_TENSOR.ATTN_K, bid),
+ data_torch[:n_head_kv * head_dim]))
+ tensors.append(
+ (self.format_tensor_name(gguf.MODEL_TENSOR.ATTN_V, bid),
+ data_torch[n_head_kv * head_dim:]))
elif name == f"transformer.h.{bid}.attn.q.weight":
- tensors.append((self.format_tensor_name(gguf.MODEL_TENSOR.ATTN_Q, bid), data_torch))
+ tensors.append(
+ (self.format_tensor_name(gguf.MODEL_TENSOR.ATTN_Q,
+ bid), data_torch))
elif name == f"transformer.h.{bid}.mlp.gate_up_proj.weight":
- tensors.append((self.format_tensor_name(gguf.MODEL_TENSOR.FFN_GATE, bid), data_torch[:ff_dim]))
- tensors.append((self.format_tensor_name(gguf.MODEL_TENSOR.FFN_UP, bid), data_torch[ff_dim:]))
+ tensors.append(
+ (self.format_tensor_name(gguf.MODEL_TENSOR.FFN_GATE,
+ bid), data_torch[:ff_dim]))
+ tensors.append(
+ (self.format_tensor_name(gguf.MODEL_TENSOR.FFN_UP,
+ bid), data_torch[ff_dim:]))
if len(tensors) == 0:
tensors.append((self.map_tensor_name(name), data_torch))
@@ -1442,7 +1915,8 @@ class RefactModel(Model):
return tensors
[email protected]("StableLmForCausalLM", "StableLMEpochForCausalLM", "LlavaStableLMEpochForCausalLM")
[email protected]("StableLmForCausalLM", "StableLMEpochForCausalLM",
+ "LlavaStableLMEpochForCausalLM")
class StableLMModel(Model):
model_arch = gguf.MODEL_ARCH.STABLELM
@@ -1463,18 +1937,22 @@ class StableLMModel(Model):
self.gguf_writer.add_feed_forward_length(hparams["intermediate_size"])
rotary_factor = self.find_hparam(["partial_rotary_factor", "rope_pct"])
self.gguf_writer.add_rope_dimension_count(
- int(rotary_factor * (hparams["hidden_size"] // hparams["num_attention_heads"])))
+ int(rotary_factor *
+ (hparams["hidden_size"] // hparams["num_attention_heads"])))
self.gguf_writer.add_head_count(hparams["num_attention_heads"])
self.gguf_writer.add_head_count_kv(hparams["num_key_value_heads"])
self.gguf_writer.add_parallel_residual(
- hparams["use_parallel_residual"] if "use_parallel_residual" in hparams else True)
- self.gguf_writer.add_layer_norm_eps(self.find_hparam(["layer_norm_eps", "norm_eps"]))
+ hparams["use_parallel_residual"] if "use_parallel_residual" in
+ hparams else True)
+ self.gguf_writer.add_layer_norm_eps(
+ self.find_hparam(["layer_norm_eps", "norm_eps"]))
self.gguf_writer.add_file_type(self.ftype)
_q_norms: list[dict[str, Tensor]] | None = None
_k_norms: list[dict[str, Tensor]] | None = None
- def modify_tensors(self, data_torch: Tensor, name: str, bid: int | None) -> Iterable[tuple[str, Tensor]]:
+ def modify_tensors(self, data_torch: Tensor, name: str,
+ bid: int | None) -> Iterable[tuple[str, Tensor]]:
n_head = self.hparams["num_attention_heads"]
n_kv_head = self.hparams["num_key_value_heads"]
@@ -1487,7 +1965,8 @@ class StableLMModel(Model):
self._q_norms[bid][name] = data_torch
if len(self._q_norms[bid]) >= n_head:
- return self._stack_qk_norm(bid, n_head, self._q_norms[bid], "q_layernorm")
+ return self._stack_qk_norm(bid, n_head, self._q_norms[bid],
+ "q_layernorm")
else:
return []
@@ -1500,13 +1979,18 @@ class StableLMModel(Model):
self._k_norms[bid][name] = data_torch
if len(self._k_norms[bid]) >= n_kv_head:
- return self._stack_qk_norm(bid, n_kv_head, self._k_norms[bid], "k_layernorm")
+ return self._stack_qk_norm(bid, n_kv_head, self._k_norms[bid],
+ "k_layernorm")
else:
return []
return [(self.map_tensor_name(name), data_torch)]
- def _stack_qk_norm(self, bid: int, n_head: int, norms: dict[str, Tensor], layer_name: str = "q_layernorm"):
+ def _stack_qk_norm(self,
+ bid: int,
+ n_head: int,
+ norms: dict[str, Tensor],
+ layer_name: str = "q_layernorm"):
data: list[Tensor] = []
# extract the norms in order
for xid in range(n_head):
@@ -1525,16 +2009,16 @@ class StableLMModel(Model):
if self._q_norms is not None or self._k_norms is not None:
# flatten two `list[dict[str, Tensor]]` into a single `list[str]`
- norms = (
- [k for d in self._q_norms for k in d.keys()] if self._q_norms is not None else []
- ) + (
- [k for d in self._k_norms for k in d.keys()] if self._k_norms is not None else []
- )
+ norms = ([k for d in self._q_norms
+ for k in d.keys()] if self._q_norms is not None else
+ []) + ([k for d in self._k_norms for k in d.keys()]
+ if self._k_norms is not None else [])
if len(norms) > 0:
raise ValueError(f"Unprocessed norms: {norms}")
[email protected]("LLaMAForCausalLM", "LlamaForCausalLM", "MistralForCausalLM", "MixtralForCausalLM")
[email protected]("LLaMAForCausalLM", "LlamaForCausalLM", "MistralForCausalLM",
+ "MixtralForCausalLM")
class LlamaModel(Model):
model_arch = gguf.MODEL_ARCH.LLAMA
@@ -1551,13 +2035,13 @@ class LlamaModel(Model):
# Apply to CodeLlama only (and ignore for Llama 3 with a vocab size of 128256)
if self.hparams.get("vocab_size", 32000) == 32016:
special_vocab = gguf.SpecialVocab(
- self.dir_model, load_merges=False,
- special_token_types = ['prefix', 'suffix', 'middle', 'eot']
- )
+ self.dir_model,
+ load_merges=False,
+ special_token_types=['prefix', 'suffix', 'middle', 'eot'])
special_vocab._set_special_token("prefix", 32007)
special_vocab._set_special_token("suffix", 32008)
special_vocab._set_special_token("middle", 32009)
- special_vocab._set_special_token("eot", 32010)
+ special_vocab._set_special_token("eot", 32010)
special_vocab.add_to_gguf(self.gguf_writer)
tokenizer_config_file = self.dir_model / 'tokenizer_config.json'
@@ -1565,7 +2049,8 @@ class LlamaModel(Model):
with open(tokenizer_config_file, "r", encoding="utf-8") as f:
tokenizer_config_json = json.load(f)
if "add_prefix_space" in tokenizer_config_json:
- self.gguf_writer.add_add_space_prefix(tokenizer_config_json["add_prefix_space"])
+ self.gguf_writer.add_add_space_prefix(
+ tokenizer_config_json["add_prefix_space"])
# Apply to granite small models only
if self.hparams.get("vocab_size", 32000) == 49152:
@@ -1582,22 +2067,27 @@ class LlamaModel(Model):
rope_dim = hparams["hidden_size"] // hparams["num_attention_heads"]
self.gguf_writer.add_rope_dimension_count(rope_dim)
- if self.hparams.get("rope_scaling") is not None and "factor" in self.hparams["rope_scaling"]:
+ if self.hparams.get(
+ "rope_scaling"
+ ) is not None and "factor" in self.hparams["rope_scaling"]:
if self.hparams["rope_scaling"].get("type") == "linear":
- self.gguf_writer.add_rope_scaling_type(gguf.RopeScalingType.LINEAR)
- self.gguf_writer.add_rope_scaling_factor(self.hparams["rope_scaling"]["factor"])
+ self.gguf_writer.add_rope_scaling_type(
+ gguf.RopeScalingType.LINEAR)
+ self.gguf_writer.add_rope_scaling_factor(
+ self.hparams["rope_scaling"]["factor"])
@staticmethod
def permute(weights: Tensor, n_head: int, n_head_kv: int | None):
if n_head_kv is not None and n_head != n_head_kv:
n_head = n_head_kv
- return (weights.reshape(n_head, 2, weights.shape[0] // n_head // 2, *weights.shape[1:])
- .swapaxes(1, 2)
- .reshape(weights.shape))
+ return (weights.reshape(n_head, 2, weights.shape[0] // n_head // 2,
+ *weights.shape[1:]).swapaxes(1, 2).reshape(
+ weights.shape))
_experts: list[dict[str, Tensor]] | None = None
- def modify_tensors(self, data_torch: Tensor, name: str, bid: int | None) -> Iterable[tuple[str, Tensor]]:
+ def modify_tensors(self, data_torch: Tensor, name: str,
+ bid: int | None) -> Iterable[tuple[str, Tensor]]:
n_head = self.hparams["num_attention_heads"]
n_kv_head = self.hparams.get("num_key_value_heads")
@@ -1646,13 +2136,17 @@ class LlamaModel(Model):
if rope_scaling := self.find_hparam(["rope_scaling"], optional=True):
if rope_scaling.get("rope_type", '').lower() == "llama3":
base = self.hparams.get("rope_theta", 10000.0)
- dim = self.hparams.get("head_dim", self.hparams["hidden_size"] // self.hparams["num_attention_heads"])
- freqs = 1.0 / (base ** (torch.arange(0, dim, 2, dtype=torch.float32) / dim))
+ dim = self.hparams.get(
+ "head_dim", self.hparams["hidden_size"] //
+ self.hparams["num_attention_heads"])
+ freqs = 1.0 / (base**(
+ torch.arange(0, dim, 2, dtype=torch.float32) / dim))
factor = rope_scaling.get("factor", 8.0)
low_freq_factor = rope_scaling.get("low_freq_factor", 1.0)
high_freq_factor = rope_scaling.get("high_freq_factor", 4.0)
- old_context_len = self.hparams.get("original_max_position_embeddings", 8192)
+ old_context_len = self.hparams.get(
+ "original_max_position_embeddings", 8192)
low_freq_wavelen = old_context_len / low_freq_factor
high_freq_wavelen = old_context_len / high_freq_factor
@@ -1666,8 +2160,10 @@ class LlamaModel(Model):
elif wavelen > low_freq_wavelen:
rope_factors.append(factor)
else:
- smooth = (old_context_len / wavelen - low_freq_factor) / (high_freq_factor - low_freq_factor)
- rope_factors.append(1 / ((1 - smooth) / factor + smooth))
+ smooth = (old_context_len / wavelen - low_freq_factor
+ ) / (high_freq_factor - low_freq_factor)
+ rope_factors.append(1 /
+ ((1 - smooth) / factor + smooth))
yield (self.format_tensor_name(gguf.MODEL_TENSOR.ROPE_FREQS),
torch.tensor(rope_factors, dtype=torch.float32))
@@ -1682,6 +2178,206 @@ class LlamaModel(Model):
raise ValueError(f"Unprocessed experts: {experts}")
[email protected]("DeciLMForCausalLM")
+class DeciModel(Model):
+ model_arch = gguf.MODEL_ARCH.DECI
+
+ @staticmethod
+ def _ffn_mult_to_intermediate_size(ffn_mult: float, n_embd: int) -> int:
+ # DeciLM-specific code
+ intermediate_size = int(2 * ffn_mult * n_embd / 3)
+ return DeciModel._find_multiple(intermediate_size, 256)
+
+ @staticmethod
+ def _find_multiple(n: int, k: int) -> int:
+ # DeciLM-specific code
+ if n % k == 0:
+ return n
+ return n + k - (n % k)
+
+ def __init__(self, *args, **kwargs):
+ super().__init__(*args, **kwargs)
+
+ if "block_configs" in self.hparams: # Llama-3_1-Nemotron-51B
+ _block_configs: list[dict[str,
+ Any]] = self.hparams["block_configs"]
+ assert self.block_count == len(_block_configs)
+ self._num_kv_heads = list()
+ self._num_heads = list()
+ _ffn_multipliers = list()
+ # ***linear attention layer***
+ # if n_heads_in_group is None and replace_with_linear is True
+ # then _num_kv_heads[il] is 0 and _num_heads[il] is num_attention_heads
+ # ***attention-free layer***
+ # if n_heads_in_group is None and replace_with_linear is False
+ # then _num_kv_heads[il] is 0 and _num_heads[il] is 0
+ # ***normal attention-layer***
+ # if n_heads_in_group is not None, then
+ # _num_kv_heads[il] is num_attention_head // n_heads_in_group and
+ # _num_heads[il] is num_attention_head
+ for il in range(len(_block_configs)):
+ if _block_configs[il]["attention"]["n_heads_in_group"] is None:
+ if _block_configs[il]["attention"][
+ "replace_with_linear"] is True:
+ self._num_kv_heads.append(0)
+ self._num_heads.append(
+ self.hparams["num_attention_heads"])
+ else:
+ self._num_kv_heads.append(0)
+ self._num_heads.append(0)
+ else:
+ self._num_kv_heads.append(
+ self.hparams["num_attention_heads"] //
+ _block_configs[il]["attention"]["n_heads_in_group"])
+ self._num_heads.append(self.hparams["num_attention_heads"])
+ _ffn_multipliers.append(_block_configs[il]["ffn"]["ffn_mult"])
+ assert self.block_count == len(self._num_kv_heads)
+ assert self.block_count == len(self._num_heads)
+ assert self.block_count == len(_ffn_multipliers)
+ assert isinstance(self._num_kv_heads, list) and isinstance(
+ self._num_kv_heads[0], int)
+ assert isinstance(self._num_heads, list) and isinstance(
+ self._num_heads[0], int)
+ assert isinstance(_ffn_multipliers, list) and isinstance(
+ _ffn_multipliers[0], float)
+ self._ffn_dims: list[int] = [
+ DeciModel._ffn_mult_to_intermediate_size(
+ multiplier, self.hparams["hidden_size"])
+ for multiplier in _ffn_multipliers
+ ]
+
+ def set_vocab(self):
+ # Please change tokenizer_config.json of Llama-3_1-Nemotron-51B's
+ # eos_token from '|eot_id|' to '|end_of_text|'
+ if self.hparams.get("vocab_size", 128256) == 128256:
+ tokens, toktypes, tokpre = self.get_vocab_base()
+ self.gguf_writer.add_tokenizer_model("gpt2")
+ self.gguf_writer.add_tokenizer_pre(tokpre)
+ self.gguf_writer.add_token_list(tokens)
+ self.gguf_writer.add_token_types(toktypes)
+
+ special_vocab = gguf.SpecialVocab(self.dir_model, load_merges=True)
+ special_vocab.add_to_gguf(self.gguf_writer)
+ else:
+ # DeciLM-7B
+ self._set_vocab_llama_hf()
+
+ def set_gguf_parameters(self):
+ if "block_configs" in self.hparams: # Llama-3_1-Nemotron-51B
+ assert self.block_count == len(self._num_kv_heads)
+ assert self.block_count == len(self._num_heads)
+ assert self.block_count == len(self._ffn_dims)
+ if (rope_theta := self.hparams.get("rope_theta")) is not None:
+ self.gguf_writer.add_rope_freq_base(rope_theta)
+ self.gguf_writer.add_head_count_kv(self._num_kv_heads)
+ self.gguf_writer.add_head_count(self._num_heads)
+ self.gguf_writer.add_feed_forward_length(self._ffn_dims)
+ self.gguf_writer.add_block_count(self.block_count)
+ self.gguf_writer.add_context_length(
+ self.hparams["max_position_embeddings"])
+ self.gguf_writer.add_embedding_length(self.hparams["hidden_size"])
+ self.gguf_writer.add_layer_norm_rms_eps(
+ self.hparams["rms_norm_eps"])
+ self.gguf_writer.add_key_length(
+ self.hparams["hidden_size"] //
+ self.hparams["num_attention_heads"])
+ self.gguf_writer.add_value_length(
+ self.hparams["hidden_size"] //
+ self.hparams["num_attention_heads"])
+ self.gguf_writer.add_file_type(self.ftype)
+ else: # DeciLM-7B
+ super().set_gguf_parameters()
+ if "num_key_value_heads_per_layer" in self.hparams: # DeciLM-7B
+ self._num_kv_heads: list[int] = self.hparams[
+ "num_key_value_heads_per_layer"]
+ assert self.block_count == len(self._num_kv_heads)
+ self.gguf_writer.add_head_count_kv(self._num_kv_heads)
+ hparams = self.hparams
+ self.gguf_writer.add_vocab_size(hparams["vocab_size"])
+
+ if "head_dim" in hparams:
+ rope_dim = hparams["head_dim"]
+ else:
+ rope_dim = hparams["hidden_size"] // hparams["num_attention_heads"]
+ self.gguf_writer.add_rope_dimension_count(rope_dim)
+
+ if self.hparams.get(
+ "rope_scaling"
+ ) is not None and "factor" in self.hparams["rope_scaling"]:
+ if self.hparams["rope_scaling"].get("type") == "linear":
+ self.gguf_writer.add_rope_scaling_type(
+ gguf.RopeScalingType.LINEAR)
+ self.gguf_writer.add_rope_scaling_factor(
+ self.hparams["rope_scaling"]["factor"])
+
+ @staticmethod
+ def permute(weights: Tensor, n_head: int, n_head_kv: int | None):
+ if n_head_kv is not None and n_head != n_head_kv:
+ n_head = n_head_kv
+ return (weights.reshape(n_head, 2, weights.shape[0] // n_head // 2,
+ *weights.shape[1:]).swapaxes(1, 2).reshape(
+ weights.shape))
+
+ def modify_tensors(self, data_torch: Tensor, name: str,
+ bid: int | None) -> Iterable[tuple[str, Tensor]]:
+ n_head = self.hparams["num_attention_heads"]
+ if bid is not None:
+ if "num_key_value_heads_per_layer" in self.hparams:
+ n_kv_head = self.hparams["num_key_value_heads_per_layer"][bid]
+ elif "block_configs" in self.hparams:
+ n_kv_head = self._num_kv_heads[bid]
+ n_head = self._num_heads[bid]
+ else:
+ n_kv_head = self.hparams.get("num_key_value_heads")
+ else:
+ n_kv_head = self.hparams.get("num_key_value_heads")
+
+ if name.endswith(("q_proj.weight", "q_proj.bias")):
+ data_torch = DeciModel.permute(data_torch, n_head, n_head)
+ if name.endswith(("k_proj.weight", "k_proj.bias")):
+ data_torch = DeciModel.permute(data_torch, n_head, n_kv_head)
+ return [(self.map_tensor_name(name), data_torch)]
+
+ def generate_extra_tensors(self) -> Iterable[tuple[str, Tensor]]:
+ if rope_scaling := self.find_hparam(["rope_scaling"], optional=True):
+ if rope_scaling.get("rope_type", '').lower() == "llama3":
+ base = self.hparams.get("rope_theta", 10000.0)
+ dim = self.hparams.get(
+ "head_dim", self.hparams["hidden_size"] //
+ self.hparams["num_attention_heads"])
+ freqs = 1.0 / (base**(
+ torch.arange(0, dim, 2, dtype=torch.float32) / dim))
+
+ factor = rope_scaling.get("factor", 8.0)
+ low_freq_factor = rope_scaling.get("low_freq_factor", 1.0)
+ high_freq_factor = rope_scaling.get("high_freq_factor", 4.0)
+ old_context_len = self.hparams.get(
+ "original_max_position_embeddings", 8192)
+
+ low_freq_wavelen = old_context_len / low_freq_factor
+ high_freq_wavelen = old_context_len / high_freq_factor
+ assert low_freq_wavelen != high_freq_wavelen
+
+ rope_factors = []
+ for freq in freqs:
+ wavelen = 2 * math.pi / freq
+ if wavelen < high_freq_wavelen:
+ rope_factors.append(1)
+ elif wavelen > low_freq_wavelen:
+ rope_factors.append(factor)
+ else:
+ smooth = (old_context_len / wavelen - low_freq_factor
+ ) / (high_freq_factor - low_freq_factor)
+ rope_factors.append(1 /
+ ((1 - smooth) / factor + smooth))
+
+ yield (self.format_tensor_name(gguf.MODEL_TENSOR.ROPE_FREQS),
+ torch.tensor(rope_factors, dtype=torch.float32))
+
+ def prepare_tensors(self):
+ super().prepare_tensors()
+
+
@Model.register("BitnetForCausalLM")
class BitnetModel(Model):
model_arch = gguf.MODEL_ARCH.BITNET
@@ -1699,21 +2395,25 @@ class BitnetModel(Model):
weight = weight.float()
scale = weight.abs().mean().clamp(min=1e-5)
iscale = 1 / scale
+ # TODO: multiply by the scale directly instead of inverting it twice
+ # (this is also unnecessarily doubly inverted upstream)
result = (weight * iscale).round().clamp(-1, 1) / iscale
return result.type(dtype)
- def modify_tensors(self, data_torch: Tensor, name: str, bid: int | None) -> Iterable[tuple[str, Tensor]]:
+ def modify_tensors(self, data_torch: Tensor, name: str,
+ bid: int | None) -> Iterable[tuple[str, Tensor]]:
new_name = self.map_tensor_name(name)
- if any(self.match_model_tensor_name(new_name, key, bid) for key in [
- gguf.MODEL_TENSOR.ATTN_Q,
- gguf.MODEL_TENSOR.ATTN_K,
- gguf.MODEL_TENSOR.ATTN_V,
- gguf.MODEL_TENSOR.ATTN_OUT,
- gguf.MODEL_TENSOR.FFN_UP,
- gguf.MODEL_TENSOR.FFN_DOWN,
- gguf.MODEL_TENSOR.FFN_GATE,
- ]):
+ if any(
+ self.match_model_tensor_name(new_name, key, bid) for key in [
+ gguf.MODEL_TENSOR.ATTN_Q,
+ gguf.MODEL_TENSOR.ATTN_K,
+ gguf.MODEL_TENSOR.ATTN_V,
+ gguf.MODEL_TENSOR.ATTN_OUT,
+ gguf.MODEL_TENSOR.FFN_UP,
+ gguf.MODEL_TENSOR.FFN_DOWN,
+ gguf.MODEL_TENSOR.FFN_GATE,
+ ]):
# transform weight into 1/0/-1 (in fp32)
data_torch = self.weight_quant(data_torch)
@@ -1735,7 +2435,8 @@ class GrokModel(Model):
_experts: list[dict[str, Tensor]] | None = None
- def modify_tensors(self, data_torch: Tensor, name: str, bid: int | None) -> Iterable[tuple[str, Tensor]]:
+ def modify_tensors(self, data_torch: Tensor, name: str,
+ bid: int | None) -> Iterable[tuple[str, Tensor]]:
# process the experts separately
if name.find(".moe.") != -1:
n_experts = self.hparams["num_local_experts"]
@@ -1801,7 +2502,8 @@ class DbrxModel(Model):
self.gguf_writer.add_file_type(self.ftype)
logger.info(f"gguf: file type = {self.ftype}")
- def modify_tensors(self, data_torch: Tensor, name: str, bid: int | None) -> Iterable[tuple[str, Tensor]]:
+ def modify_tensors(self, data_torch: Tensor, name: str,
+ bid: int | None) -> Iterable[tuple[str, Tensor]]:
del bid # unused
n_expert = self.hparams["ffn_config"]["moe_num_experts"]
@@ -1814,16 +2516,21 @@ class DbrxModel(Model):
# AND the dimensions in ggml are typically in the reverse order of the pytorch dimensions
# so (n_expert, n_ff, n_embd) in pytorch is {n_embd, n_ff, n_expert} in ggml_tensor
exp_tensor_names = {
- "ffn.experts.mlp.w1": None, # LLM_TENSOR_FFN_GATE_EXPS ggml_tensor->ne{n_embd, n_ff, n_expert}
- "ffn.experts.mlp.w2": (0, 2, 1), # LLM_TENSOR_FFN_DOWN_EXPS ggml_tensor->ne{n_ff, n_embd, n_expert}
- "ffn.experts.mlp.v1": None} # LLM_TENSOR_FFN_UP_EXPS ggml_tensor->ne{n_embd, n_ff, n_expert}
+ "ffn.experts.mlp.w1":
+ None, # LLM_TENSOR_FFN_GATE_EXPS ggml_tensor->ne{n_embd, n_ff, n_expert}
+ "ffn.experts.mlp.w2": (
+ 0, 2, 1
+ ), # LLM_TENSOR_FFN_DOWN_EXPS ggml_tensor->ne{n_ff, n_embd, n_expert}
+ "ffn.experts.mlp.v1": None
+ } # LLM_TENSOR_FFN_UP_EXPS ggml_tensor->ne{n_embd, n_ff, n_expert}
experts = False
for exp_tensor_name in exp_tensor_names.keys():
if name.find(exp_tensor_name) != -1 and name.find(".weight") == -1:
experts = True
data_torch = data_torch.view(n_expert, n_ff, n_embd)
- if (permute_tensor := exp_tensor_names[exp_tensor_name]) is not None:
+ if (permute_tensor :=
+ exp_tensor_names[exp_tensor_name]) is not None:
data_torch = data_torch.permute(*permute_tensor)
break
@@ -1833,12 +2540,14 @@ class DbrxModel(Model):
# Every other model has the weight names ending in .weight,
# let's assume that is the convention which is not the case for dbrx:
# https://huggingface.co/databricks/dbrx-instruct/blob/main/model.safetensors.index.json#L15
- new_name = self.map_tensor_name(name if not experts else name + ".weight", try_suffixes=(".weight",))
+ new_name = self.map_tensor_name(name if not experts else name +
+ ".weight",
+ try_suffixes=(".weight", ))
return [(new_name, data_torch)]
- def tensor_force_quant(
- self, name: str, new_name: str, bid: int | None, n_dims: int) -> gguf.GGMLQuantizationType | bool:
+ def tensor_force_quant(self, name: str, new_name: str, bid: int | None,
+ n_dims: int) -> gguf.GGMLQuantizationType | bool:
del name, new_name, bid # unused
return n_dims > 1
@@ -1849,31 +2558,58 @@ class MiniCPMModel(Model):
model_arch = gguf.MODEL_ARCH.MINICPM
def set_gguf_parameters(self):
- block_count = self.hparams["num_hidden_layers"]
- self.gguf_writer.add_context_length(self.hparams["max_position_embeddings"])
- self.gguf_writer.add_embedding_length(self.hparams["hidden_size"])
- self.gguf_writer.add_block_count(block_count)
- self.gguf_writer.add_feed_forward_length(self.hparams["intermediate_size"])
- self.gguf_writer.add_rope_dimension_count(self.hparams["hidden_size"] // self.hparams["num_attention_heads"])
- self.gguf_writer.add_head_count(self.hparams["num_attention_heads"])
- self.gguf_writer.add_head_count_kv(self.hparams["num_key_value_heads"])
- self.gguf_writer.add_layer_norm_rms_eps(self.hparams["rms_norm_eps"])
- self.gguf_writer.add_file_type(self.ftype)
+ super().set_gguf_parameters()
+ embedding_scale = float(self.hparams["scale_emb"])
+ self.gguf_writer.add_embedding_scale(embedding_scale)
+ logger.info(f"gguf: (minicpm) embedding_scale = {embedding_scale}")
+ residual_scale = self.hparams["scale_depth"] / self.hparams[
+ "num_hidden_layers"]**0.5
+ self.gguf_writer.add_residual_scale(residual_scale)
+ logger.info(f"gguf: (minicpm) residual_scale = {residual_scale}")
+ logit_scale = self.hparams["hidden_size"] / self.hparams[
+ "dim_model_base"]
+ self.gguf_writer.add_logit_scale(logit_scale)
+ logger.info(f"gguf: (minicpm) logit_scale = {logit_scale}")
+ if self.hparams.get("rope_scaling") is not None:
+ if self.hparams["rope_scaling"].get("type") == "longrope":
+ self.gguf_writer.add_rope_scaling_type(
+ gguf.RopeScalingType.LONGROPE)
+ logger.info(
+ f"gguf: (minicpm) rope_scaling_type = {gguf.RopeScalingType.LONGROPE}"
+ )
- def set_vocab(self):
- self._set_vocab_llama_hf()
+ def generate_extra_tensors(self) -> Iterable[tuple[str, Tensor]]:
+ rope_dims = self.hparams["hidden_size"] // self.hparams[
+ "num_attention_heads"]
- def _reverse_hf_permute(self, weights: Tensor, n_head: int, n_kv_head: int | None = None) -> Tensor:
- if n_kv_head is not None and n_head != n_kv_head:
- n_head //= n_kv_head
+ rope_scaling = self.find_hparam(['rope_scaling'], True)
+ if rope_scaling is not None:
+ long_factors = rope_scaling.get('long_factor', None)
+ short_factors = rope_scaling.get('short_factor', None)
- return (
- weights.reshape(n_head, 2, weights.shape[0] // n_head // 2, *weights.shape[1:])
- .swapaxes(1, 2)
- .reshape(weights.shape)
- )
+ if long_factors is None or short_factors is None:
+ raise KeyError(
+ 'Missing the required key rope_scaling.long_factor or rope_scaling_short_factor'
+ )
+
+ if len(long_factors) != len(short_factors) or len(
+ long_factors) != rope_dims / 2:
+ raise ValueError(
+ f'The length of rope long and short factors must be {rope_dims / 2}'
+ )
+
+ yield (self.format_tensor_name(
+ gguf.MODEL_TENSOR.ROPE_FACTORS_LONG),
+ torch.tensor(long_factors, dtype=torch.float32))
+ yield (self.format_tensor_name(
+ gguf.MODEL_TENSOR.ROPE_FACTORS_SHORT),
+ torch.tensor(short_factors, dtype=torch.float32))
+
+ def set_vocab(self):
+ self._set_vocab_sentencepiece()
- def modify_tensors(self, data_torch: Tensor, name: str, bid: int | None) -> Iterable[tuple[str, Tensor]]:
+ def modify_tensors(self, data_torch: Tensor, name: str,
+ bid: int | None) -> Iterable[tuple[str, Tensor]]:
del bid # unused
n_head = self.hparams["num_attention_heads"]
@@ -1881,9 +2617,9 @@ class MiniCPMModel(Model):
# HF models permute some of the tensors, so we need to undo that
if name.endswith(("q_proj.weight")):
- data_torch = self._reverse_hf_permute(data_torch, n_head, n_head)
+ data_torch = LlamaModel.permute(data_torch, n_head, n_head)
if name.endswith(("k_proj.weight")):
- data_torch = self._reverse_hf_permute(data_torch, n_head, n_kv_head)
+ data_torch = LlamaModel.permute(data_torch, n_head, n_kv_head)
return [(self.map_tensor_name(name), data_torch)]
@@ -1907,7 +2643,8 @@ class MiniCPM3Model(Model):
if "q_lora_rank" in hparams and hparams["q_lora_rank"] is not None:
self.gguf_writer.add_q_lora_rank(hparams["q_lora_rank"])
self.gguf_writer.add_kv_lora_rank(hparams["kv_lora_rank"])
- self.gguf_writer.add_key_length(hparams["qk_nope_head_dim"] + hparams["qk_rope_head_dim"])
+ self.gguf_writer.add_key_length(hparams["qk_nope_head_dim"] +
+ hparams["qk_rope_head_dim"])
self.gguf_writer.add_rope_dimension_count(hparams["qk_rope_head_dim"])
def generate_extra_tensors(self) -> Iterable[tuple[str, Tensor]]:
@@ -1919,28 +2656,36 @@ class MiniCPM3Model(Model):
short_factors = rope_scaling.get('short_factor', None)
if long_factors is None or short_factors is None:
- raise KeyError('Missing the required key rope_scaling.long_factor or rope_scaling_short_factor')
+ raise KeyError(
+ 'Missing the required key rope_scaling.long_factor or rope_scaling_short_factor'
+ )
- if len(long_factors) != len(short_factors) or len(long_factors) != rope_dims / 2:
- raise ValueError(f'The length of rope long and short factors must be {rope_dims / 2}')
+ if len(long_factors) != len(short_factors) or len(
+ long_factors) != rope_dims / 2:
+ raise ValueError(
+ f'The length of rope long and short factors must be {rope_dims / 2}'
+ )
yield (self.format_tensor_name(
- gguf.MODEL_TENSOR.ROPE_FACTORS_LONG), torch.tensor(long_factors, dtype=torch.float32))
+ gguf.MODEL_TENSOR.ROPE_FACTORS_LONG),
+ torch.tensor(long_factors, dtype=torch.float32))
yield (self.format_tensor_name(
- gguf.MODEL_TENSOR.ROPE_FACTORS_SHORT), torch.tensor(short_factors, dtype=torch.float32))
+ gguf.MODEL_TENSOR.ROPE_FACTORS_SHORT),
+ torch.tensor(short_factors, dtype=torch.float32))
def set_vocab(self):
self._set_vocab_sentencepiece()
- def _reverse_hf_permute(self, weights: Tensor, n_head: int, n_kv_head: int | None = None) -> Tensor:
+ def _reverse_hf_permute(self,
+ weights: Tensor,
+ n_head: int,
+ n_kv_head: int | None = None) -> Tensor:
if n_kv_head is not None and n_head != n_kv_head:
n_head //= n_kv_head
- return (
- weights.reshape(n_head, 2, weights.shape[0] // n_head // 2, *weights.shape[1:])
- .swapaxes(1, 2)
- .reshape(weights.shape)
- )
+ return (weights.reshape(n_head, 2, weights.shape[0] // n_head // 2,
+ *weights.shape[1:]).swapaxes(1, 2).reshape(
+ weights.shape))
@Model.register("QWenLMHeadModel")
@@ -1951,10 +2696,13 @@ class QwenModel(Model):
def token_bytes_to_string(b):
from transformers.models.gpt2.tokenization_gpt2 import bytes_to_unicode
byte_encoder = bytes_to_unicode()
- return ''.join([byte_encoder[ord(char)] for char in b.decode('latin-1')])
+ return ''.join(
+ [byte_encoder[ord(char)] for char in b.decode('latin-1')])
@staticmethod
- def bpe(mergeable_ranks: dict[bytes, int], token: bytes, max_rank: int | None = None) -> list[bytes]:
+ def bpe(mergeable_ranks: dict[bytes, int],
+ token: bytes,
+ max_rank: int | None = None) -> list[bytes]:
parts = [bytes([b]) for b in token]
while True:
min_idx = None
@@ -1964,24 +2712,30 @@ class QwenModel(Model):
if rank is not None and (min_rank is None or rank < min_rank):
min_idx = i
min_rank = rank
- if min_rank is None or (max_rank is not None and min_rank >= max_rank):
+ if min_rank is None or (max_rank is not None
+ and min_rank >= max_rank):
break
assert min_idx is not None
- parts = parts[:min_idx] + [parts[min_idx] + parts[min_idx + 1]] + parts[min_idx + 2:]
+ parts = parts[:min_idx] + [parts[min_idx] + parts[min_idx + 1]
+ ] + parts[min_idx + 2:]
return parts
def set_vocab(self):
self._set_vocab_qwen()
def set_gguf_parameters(self):
- self.gguf_writer.add_context_length(self.hparams["max_position_embeddings"])
+ self.gguf_writer.add_context_length(
+ self.hparams["max_position_embeddings"])
self.gguf_writer.add_block_count(self.hparams["num_hidden_layers"])
self.gguf_writer.add_embedding_length(self.hparams["hidden_size"])
- self.gguf_writer.add_feed_forward_length(self.hparams["intermediate_size"])
+ self.gguf_writer.add_feed_forward_length(
+ self.hparams["intermediate_size"])
self.gguf_writer.add_rope_freq_base(self.hparams["rotary_emb_base"])
- self.gguf_writer.add_rope_dimension_count(self.hparams["hidden_size"] // self.hparams["num_attention_heads"])
+ self.gguf_writer.add_rope_dimension_count(
+ self.hparams["hidden_size"] // self.hparams["num_attention_heads"])
self.gguf_writer.add_head_count(self.hparams["num_attention_heads"])
- self.gguf_writer.add_layer_norm_rms_eps(self.hparams["layer_norm_epsilon"])
+ self.gguf_writer.add_layer_norm_rms_eps(
+ self.hparams["layer_norm_epsilon"])
self.gguf_writer.add_file_type(self.ftype)
@@ -1995,6 +2749,87 @@ class Qwen2Model(Model):
except FileNotFoundError:
self._set_vocab_gpt2()
+ def set_gguf_parameters(self):
+ super().set_gguf_parameters()
+ if self.hparams.get(
+ "rope_scaling"
+ ) is not None and "factor" in self.hparams["rope_scaling"]:
+ if self.hparams["rope_scaling"].get("type") == "yarn":
+ self.gguf_writer.add_rope_scaling_type(
+ gguf.RopeScalingType.YARN)
+ self.gguf_writer.add_rope_scaling_factor(
+ self.hparams["rope_scaling"]["factor"])
+ self.gguf_writer.add_rope_scaling_orig_ctx_len(
+ self.hparams["rope_scaling"]
+ ["original_max_position_embeddings"])
+
+
[email protected]("Qwen2VLForConditionalGeneration")
+class Qwen2VLModel(Model):
+ model_arch = gguf.MODEL_ARCH.QWEN2VL
+
+ def set_gguf_parameters(self):
+ super().set_gguf_parameters()
+ mrope_section = self.hparams["rope_scaling"]["mrope_section"]
+ mrope_section += [0] * max(0, 4 - len(mrope_section))
+ self.gguf_writer.add_rope_dimension_sections(mrope_section)
+
+ def set_vocab(self):
+ try:
+ self._set_vocab_sentencepiece()
+ except FileNotFoundError:
+ self._set_vocab_gpt2()
+
+ def get_tensors(self) -> Iterator[tuple[str, Tensor]]:
+ for name, data in super().get_tensors():
+ if name.startswith("visual."):
+ continue
+ yield name, data
+
+
[email protected]("WavTokenizerDec")
+class WavTokenizerDecModel(Model):
+ model_arch = gguf.MODEL_ARCH.WAVTOKENIZER_DEC
+
+ def modify_tensors(self, data_torch: Tensor, name: str,
+ bid: int | None) -> Iterable[tuple[str, Tensor]]:
+ del bid # unused
+
+ if \
+ name.endswith("codebook.cluster_size") or \
+ name.endswith("codebook.embed_avg") or \
+ name.endswith("codebook.inited"):
+ logger.debug(f"Skipping {name!r}")
+ return []
+
+ logger.info(f"{self.map_tensor_name(name)} -> {data_torch.shape}")
+
+ return [(self.map_tensor_name(name), data_torch)]
+
+ def set_vocab(self):
+ self._set_vocab_none()
+
+ def set_gguf_parameters(self):
+ super().set_gguf_parameters()
+ self.gguf_writer.add_vocab_size(self.hparams["vocab_size"])
+ self.gguf_writer.add_features_length(self.hparams["n_embd_features"])
+ self.gguf_writer.add_feed_forward_length(self.hparams["n_ff"])
+ self.gguf_writer.add_group_norm_eps(self.hparams["group_norm_epsilon"])
+ self.gguf_writer.add_group_norm_groups(
+ self.hparams["group_norm_groups"])
+
+ self.gguf_writer.add_posnet_embedding_length(
+ self.hparams["posnet"]["n_embd"])
+ self.gguf_writer.add_posnet_block_count(
+ self.hparams["posnet"]["n_layer"])
+
+ self.gguf_writer.add_convnext_embedding_length(
+ self.hparams["convnext"]["n_embd"])
+ self.gguf_writer.add_convnext_block_count(
+ self.hparams["convnext"]["n_layer"])
+
+ self.gguf_writer.add_causal_attention(False)
+
@Model.register("Qwen2MoeForCausalLM")
class Qwen2MoeModel(Model):
@@ -2004,16 +2839,25 @@ class Qwen2MoeModel(Model):
super().set_gguf_parameters()
if (n_experts := self.hparams.get("num_experts")) is not None:
self.gguf_writer.add_expert_count(n_experts)
- if (moe_intermediate_size := self.hparams.get("moe_intermediate_size")) is not None:
- self.gguf_writer.add_expert_feed_forward_length(moe_intermediate_size)
- logger.info(f"gguf: expert feed forward length = {moe_intermediate_size}")
- if (shared_expert_intermediate_size := self.hparams.get('shared_expert_intermediate_size')) is not None:
- self.gguf_writer.add_expert_shared_feed_forward_length(shared_expert_intermediate_size)
- logger.info(f"gguf: expert shared feed forward length = {shared_expert_intermediate_size}")
+ if (moe_intermediate_size :=
+ self.hparams.get("moe_intermediate_size")) is not None:
+ self.gguf_writer.add_expert_feed_forward_length(
+ moe_intermediate_size)
+ logger.info(
+ f"gguf: expert feed forward length = {moe_intermediate_size}")
+ if (shared_expert_intermediate_size :=
+ self.hparams.get('shared_expert_intermediate_size')
+ ) is not None:
+ self.gguf_writer.add_expert_shared_feed_forward_length(
+ shared_expert_intermediate_size)
+ logger.info(
+ f"gguf: expert shared feed forward length = {shared_expert_intermediate_size}"
+ )
_experts: list[dict[str, Tensor]] | None = None
- def modify_tensors(self, data_torch: Tensor, name: str, bid: int | None) -> Iterable[tuple[str, Tensor]]:
+ def modify_tensors(self, data_torch: Tensor, name: str,
+ bid: int | None) -> Iterable[tuple[str, Tensor]]:
# process the experts separately
if name.find("experts") != -1:
n_experts = self.hparams["num_experts"]
@@ -2072,7 +2916,8 @@ class GPT2Model(Model):
self.gguf_writer.add_layer_norm_eps(self.hparams["layer_norm_epsilon"])
self.gguf_writer.add_file_type(self.ftype)
- def modify_tensors(self, data_torch: Tensor, name: str, bid: int | None) -> Iterable[tuple[str, Tensor]]:
+ def modify_tensors(self, data_torch: Tensor, name: str,
+ bid: int | None) -> Iterable[tuple[str, Tensor]]:
del bid # unused
tensors: list[tuple[str, Tensor]] = []
@@ -2081,7 +2926,8 @@ class GPT2Model(Model):
if name.endswith((".attn.bias", ".attn.masked_bias")):
return tensors
- if name.endswith((".c_attn.weight", ".c_proj.weight", ".c_fc.weight", ".c_proj.weight")):
+ if name.endswith((".c_attn.weight", ".c_proj.weight", ".c_fc.weight",
+ ".c_proj.weight")):
data_torch = data_torch.transpose(1, 0)
new_name = self.map_tensor_name(name)
@@ -2090,7 +2936,8 @@ class GPT2Model(Model):
# note: GPT2 output is tied to (same as) wte in original model
if new_name == self.format_tensor_name(gguf.MODEL_TENSOR.TOKEN_EMBD):
- tensors.append((self.format_tensor_name(gguf.MODEL_TENSOR.OUTPUT), data_torch))
+ tensors.append((self.format_tensor_name(gguf.MODEL_TENSOR.OUTPUT),
+ data_torch))
return tensors
@@ -2106,15 +2953,18 @@ class Phi2Model(Model):
n_embd = self.find_hparam(["hidden_size", "n_embd"])
n_head = self.find_hparam(["num_attention_heads", "n_head"])
- self.gguf_writer.add_context_length(self.find_hparam(["n_positions", "max_position_embeddings"]))
+ self.gguf_writer.add_context_length(
+ self.find_hparam(["n_positions", "max_position_embeddings"]))
self.gguf_writer.add_embedding_length(n_embd)
self.gguf_writer.add_feed_forward_length(4 * n_embd)
self.gguf_writer.add_block_count(block_count)
self.gguf_writer.add_head_count(n_head)
self.gguf_writer.add_head_count_kv(n_head)
- self.gguf_writer.add_layer_norm_eps(self.find_hparam(["layer_norm_epsilon", "layer_norm_eps"]))
- self.gguf_writer.add_rope_dimension_count(int(rot_pct * n_embd) // n_head)
+ self.gguf_writer.add_layer_norm_eps(
+ self.find_hparam(["layer_norm_epsilon", "layer_norm_eps"]))
+ self.gguf_writer.add_rope_dimension_count(
+ int(rot_pct * n_embd) // n_head)
self.gguf_writer.add_file_type(self.ftype)
self.gguf_writer.add_add_bos_token(False)
@@ -2124,6 +2974,15 @@ class Phi3MiniModel(Model):
model_arch = gguf.MODEL_ARCH.PHI3
def set_vocab(self):
+ # Phi-4 model uses GPT2Tokenizer
+ tokenizer_config_file = self.dir_model / 'tokenizer_config.json'
+ if tokenizer_config_file.is_file():
+ with open(tokenizer_config_file, "r", encoding="utf-8") as f:
+ tokenizer_config_json = json.load(f)
+ tokenizer_class = tokenizer_config_json['tokenizer_class']
+ if tokenizer_class == 'GPT2Tokenizer':
+ return self._set_vocab_gpt2()
+
from sentencepiece import SentencePieceProcessor
tokenizer_path = self.dir_model / 'tokenizer.model'
@@ -2136,7 +2995,9 @@ class Phi3MiniModel(Model):
vocab_size = self.hparams.get('vocab_size', tokenizer.vocab_size())
- tokens: list[bytes] = [f"[PAD{i}]".encode("utf-8") for i in range(vocab_size)]
+ tokens: list[bytes] = [
+ f"[PAD{i}]".encode("utf-8") for i in range(vocab_size)
+ ]
scores: list[float] = [-10000.0] * vocab_size
toktypes: list[int] = [SentencePieceTokenTypes.UNUSED] * vocab_size
@@ -2168,7 +3029,9 @@ class Phi3MiniModel(Model):
for key in added_tokens_json:
token_id = added_tokens_json[key]
if token_id >= vocab_size:
- logger.debug(f'ignore token {token_id}: id is out of range, max={vocab_size - 1}')
+ logger.debug(
+ f'ignore token {token_id}: id is out of range, max={vocab_size - 1}'
+ )
continue
tokens[token_id] = key.encode("utf-8")
@@ -2179,15 +3042,17 @@ class Phi3MiniModel(Model):
if tokenizer_config_file.is_file():
with open(tokenizer_config_file, "r", encoding="utf-8") as f:
tokenizer_config_json = json.load(f)
- added_tokens_decoder = tokenizer_config_json.get("added_tokens_decoder", {})
+ added_tokens_decoder = tokenizer_config_json.get(
+ "added_tokens_decoder", {})
for token_id, foken_data in added_tokens_decoder.items():
token_id = int(token_id)
token = foken_data["content"].encode("utf-8")
if toktypes[token_id] != SentencePieceTokenTypes.UNUSED:
if tokens[token_id] != token:
logger.warning(
- f'replacing token {token_id}: {tokens[token_id].decode("utf-8")!r}' \
- + ' -> {token.decode("utf-8")!r}')
+ f"replacing token {token_id}: {tokens[token_id].decode('utf-8')!r} "
+ f"-> {token.decode('utf-8')!r}"
+ )
tokens[token_id] = token
scores[token_id] = -1000.0
toktypes[token_id] = SentencePieceTokenTypes.USER_DEFINED
@@ -2205,8 +3070,9 @@ class Phi3MiniModel(Model):
if toktypes[token_id] != SentencePieceTokenTypes.UNUSED:
if tokens[token_id] != token:
logger.warning(
- f'replacing token {token_id}: {tokens[token_id].decode("utf-8")!r}' \
- +' -> {token.decode("utf-8")!r}')
+ f"replacing token {token_id}: {tokens[token_id].decode('utf-8')!r} "
+ f"-> {token.decode('utf-8')!r}"
+ )
tokens[token_id] = token
scores[token_id] = -1000.0
toktypes[token_id] = SentencePieceTokenTypes.USER_DEFINED
@@ -2229,14 +3095,17 @@ class Phi3MiniModel(Model):
n_head = self.find_hparam(["num_attention_heads", "n_head"])
n_head_kv = self.find_hparam(["num_key_value_heads", "n_head_kv"])
rms_eps = self.find_hparam(["rms_norm_eps"])
- max_pos_embds = self.find_hparam(["n_positions", "max_position_embeddings"])
- orig_max_pos_embds = self.find_hparam(["original_max_position_embeddings"])
+ max_pos_embds = self.find_hparam(
+ ["n_positions", "max_position_embeddings"])
+ orig_max_pos_embds = self.find_hparam(
+ ["original_max_position_embeddings"])
rope_dims = n_embd // n_head
self.gguf_writer.add_context_length(max_pos_embds)
self.gguf_writer.add_rope_scaling_orig_ctx_len(orig_max_pos_embds)
self.gguf_writer.add_embedding_length(n_embd)
- self.gguf_writer.add_feed_forward_length(self.find_hparam(["intermediate_size"]))
+ self.gguf_writer.add_feed_forward_length(
+ self.find_hparam(["intermediate_size"]))
self.gguf_writer.add_block_count(block_count)
self.gguf_writer.add_head_count(n_head)
self.gguf_writer.add_head_count_kv(n_head_kv)
@@ -2244,48 +3113,122 @@ class Phi3MiniModel(Model):
self.gguf_writer.add_rope_dimension_count(rope_dims)
self.gguf_writer.add_rope_freq_base(self.find_hparam(["rope_theta"]))
self.gguf_writer.add_file_type(self.ftype)
- self.gguf_writer.add_sliding_window(self.find_hparam(["sliding_window"]))
+ sliding_window = self.hparams.get("sliding_window")
+ # use zero value of sliding_window to distinguish Phi-4 from other PHI3 models
+ if sliding_window is None:
+ sliding_window = 0
+ self.gguf_writer.add_sliding_window(sliding_window)
+
+ def generate_extra_tensors(self) -> Iterable[tuple[str, Tensor]]:
+ n_embd = self.find_hparam(["hidden_size", "n_embd"])
+ n_head = self.find_hparam(["num_attention_heads", "n_head"])
+ max_pos_embds = self.find_hparam(
+ ["n_positions", "max_position_embeddings"])
+ orig_max_pos_embds = self.find_hparam(
+ ["original_max_position_embeddings"])
+ rope_dims = n_embd // n_head
+
+ # write rope scaling for long context (128k) model
+ rope_scaling = self.find_hparam(['rope_scaling'], True)
+ if rope_scaling is None:
+ return
+
+ scale = max_pos_embds / orig_max_pos_embds
+
+ rope_scaling_type = rope_scaling.get('type', '').lower()
+ if len(rope_scaling_type) == 0:
+ raise KeyError('Missing the required key rope_scaling.type')
+
+ if rope_scaling_type == 'su' or rope_scaling_type == 'longrope':
+ attn_factor = math.sqrt(
+ 1 + math.log(scale) /
+ math.log(orig_max_pos_embds)) if scale > 1.0 else 1.0
+ elif rope_scaling_type == 'yarn':
+ attn_factor = 0.1 * math.log(scale) + 1.0 if scale > 1.0 else 1.0
+ else:
+ raise NotImplementedError(
+ f'The rope scaling type {rope_scaling_type} is not supported yet'
+ )
+
+ self.gguf_writer.add_rope_scaling_attn_factors(attn_factor)
+
+ long_factors = rope_scaling.get('long_factor', None)
+ short_factors = rope_scaling.get('short_factor', None)
+
+ if long_factors is None or short_factors is None:
+ raise KeyError(
+ 'Missing the required key rope_scaling.long_factor or rope_scaling_short_factor'
+ )
+
+ if len(long_factors) != len(short_factors) or len(
+ long_factors) != rope_dims / 2:
+ raise ValueError(
+ f'The length of rope long and short factors must be {rope_dims / 2}'
+ )
+
+ yield (self.format_tensor_name(gguf.MODEL_TENSOR.ROPE_FACTORS_LONG),
+ torch.tensor(long_factors, dtype=torch.float32))
+ yield (self.format_tensor_name(gguf.MODEL_TENSOR.ROPE_FACTORS_SHORT),
+ torch.tensor(short_factors, dtype=torch.float32))
+
+
[email protected]("PhiMoEForCausalLM")
+class PhiMoeModel(Phi3MiniModel):
+ model_arch = gguf.MODEL_ARCH.PHIMOE
+
+ _experts: list[dict[str, Tensor]] | None = None
+
+ def set_gguf_parameters(self):
+ super().set_gguf_parameters()
+ self.gguf_writer.add_expert_used_count(
+ self.hparams["num_experts_per_tok"])
+ self.gguf_writer.add_expert_count(self.hparams["num_local_experts"])
+
+ def modify_tensors(self, data_torch: Tensor, name: str,
+ bid: int | None) -> Iterable[tuple[str, Tensor]]:
+ # process the experts separately
+ if name.find("block_sparse_moe.experts") != -1:
+ n_experts = self.hparams["num_local_experts"]
+ assert bid is not None
+
+ if self._experts is None:
+ self._experts = [{} for _ in range(self.block_count)]
- def generate_extra_tensors(self) -> Iterable[tuple[str, Tensor]]:
- n_embd = self.find_hparam(["hidden_size", "n_embd"])
- n_head = self.find_hparam(["num_attention_heads", "n_head"])
- max_pos_embds = self.find_hparam(["n_positions", "max_position_embeddings"])
- orig_max_pos_embds = self.find_hparam(["original_max_position_embeddings"])
- rope_dims = n_embd // n_head
+ self._experts[bid][name] = data_torch
- # write rope scaling for long context (128k) model
- rope_scaling = self.find_hparam(['rope_scaling'], True)
- if rope_scaling is None:
- return
+ if len(self._experts[bid]) >= n_experts * 3:
+ tensors: list[tuple[str, Tensor]] = []
- scale = max_pos_embds / orig_max_pos_embds
+ # merge the experts into a single 3d tensor
+ for w_name in ["w1", "w2", "w3"]:
+ data: list[Tensor] = []
- rope_scaling_type = rope_scaling.get('type', '').lower()
- if len(rope_scaling_type) == 0:
- raise KeyError('Missing the required key rope_scaling.type')
+ for xid in range(n_experts):
+ ename = f"model.layers.{bid}.block_sparse_moe.experts.{xid}.{w_name}.weight"
+ data.append(self._experts[bid][ename])
+ del self._experts[bid][ename]
- if rope_scaling_type == 'su' or rope_scaling_type == 'longrope':
- attn_factor = math.sqrt(1 + math.log(scale) / math.log(orig_max_pos_embds)) if scale > 1.0 else 1.0
- elif rope_scaling_type == 'yarn':
- attn_factor = 0.1 * math.log(scale) + 1.0 if scale > 1.0 else 1.0
- else:
- raise NotImplementedError(f'The rope scaling type {rope_scaling_type} is not supported yet')
+ data_torch = torch.stack(data, dim=0)
- self.gguf_writer.add_rope_scaling_attn_factors(attn_factor)
+ merged_name = f"model.layers.{bid}.block_sparse_moe.experts.{w_name}.weight"
- long_factors = rope_scaling.get('long_factor', None)
- short_factors = rope_scaling.get('short_factor', None)
+ new_name = self.map_tensor_name(merged_name)
- if long_factors is None or short_factors is None:
- raise KeyError('Missing the required key rope_scaling.long_factor or rope_scaling_short_factor')
+ tensors.append((new_name, data_torch))
+ return tensors
+ else:
+ return []
- if len(long_factors) != len(short_factors) or len(long_factors) != rope_dims / 2:
- raise ValueError(f'The length of rope long and short factors must be {rope_dims / 2}')
+ return [(self.map_tensor_name(name), data_torch)]
- yield (self.format_tensor_name(gguf.MODEL_TENSOR.ROPE_FACTORS_LONG),
- torch.tensor(long_factors, dtype=torch.float32))
- yield (self.format_tensor_name(gguf.MODEL_TENSOR.ROPE_FACTORS_SHORT),
- torch.tensor(short_factors, dtype=torch.float32))
+ def prepare_tensors(self):
+ super().prepare_tensors()
+
+ if self._experts is not None:
+ # flatten `list[dict[str, Tensor]]` into `list[str]`
+ experts = [k for d in self._experts for k in d.keys()]
+ if len(experts) > 0:
+ raise ValueError(f"Unprocessed experts: {experts}")
@Model.register("PlamoForCausalLM")
@@ -2304,7 +3247,8 @@ class PlamoModel(Model):
self.gguf_writer.add_feed_forward_length(hparams["intermediate_size"])
self.gguf_writer.add_block_count(block_count)
self.gguf_writer.add_head_count(hparams["num_attention_heads"])
- self.gguf_writer.add_head_count_kv(5) # hparams["num_key_value_heads"]) is wrong
+ self.gguf_writer.add_head_count_kv(
+ 5) # hparams["num_key_value_heads"]) is wrong
self.gguf_writer.add_layer_norm_rms_eps(hparams["rms_norm_eps"])
self.gguf_writer.add_file_type(self.ftype)
@@ -2322,7 +3266,8 @@ class PlamoModel(Model):
data_torch = torch.reshape(data_torch, (5120, 5120))
return data_torch
- def modify_tensors(self, data_torch: Tensor, name: str, bid: int | None) -> Iterable[tuple[str, Tensor]]:
+ def modify_tensors(self, data_torch: Tensor, name: str,
+ bid: int | None) -> Iterable[tuple[str, Tensor]]:
del bid # unused
new_name = self.map_tensor_name(name)
@@ -2355,7 +3300,8 @@ class CodeShellModel(Model):
self.gguf_writer.add_rope_scaling_type(gguf.RopeScalingType.LINEAR)
self.gguf_writer.add_rope_scaling_factor(1.0)
- def modify_tensors(self, data_torch: Tensor, name: str, bid: int | None) -> Iterable[tuple[str, Tensor]]:
+ def modify_tensors(self, data_torch: Tensor, name: str,
+ bid: int | None) -> Iterable[tuple[str, Tensor]]:
del bid # unused
new_name = self.map_tensor_name(name)
@@ -2365,9 +3311,12 @@ class CodeShellModel(Model):
if new_name == self.format_tensor_name(gguf.MODEL_TENSOR.TOKEN_EMBD):
assert self.tensor_names is not None
- if all(s not in self.tensor_names for s in ("lm_head.weight", "output.weight")): # pylint: disable=E1135
+ if all(s not in self.tensor_names
+ for s in ("lm_head.weight", "output.weight")):
# copy tok_embd.weight to output.weight
- tensors.append((self.format_tensor_name(gguf.MODEL_TENSOR.OUTPUT), data_torch))
+ tensors.append(
+ (self.format_tensor_name(gguf.MODEL_TENSOR.OUTPUT),
+ data_torch))
return tensors
@@ -2447,7 +3396,8 @@ class InternLM2Model(Model):
if tokenizer_config_file.is_file():
with open(tokenizer_config_file, "r", encoding="utf-8") as f:
tokenizer_config_json = json.load(f)
- added_tokens_decoder = tokenizer_config_json.get("added_tokens_decoder", {})
+ added_tokens_decoder = tokenizer_config_json.get(
+ "added_tokens_decoder", {})
for token_id, foken_data in added_tokens_decoder.items():
token_id = int(token_id)
token = foken_data["content"]
@@ -2457,8 +3407,9 @@ class InternLM2Model(Model):
if toktypes[token_id] != SentencePieceTokenTypes.UNUSED:
if tokens[token_id] != token:
logger.warning(
- f'replacing token {token_id}: {tokens[token_id].decode("utf-8")!r}' \
- +' -> {token.decode("utf-8")!r}')
+ f"replacing token {token_id}: {tokens[token_id].decode('utf-8')!r} "
+ f"-> {token.decode('utf-8')!r}"
+ )
tokens[token_id] = token
scores[token_id] = -1000.0
toktypes[token_id] = SentencePieceTokenTypes.USER_DEFINED
@@ -2479,8 +3430,9 @@ class InternLM2Model(Model):
if toktypes[token_id] != SentencePieceTokenTypes.UNUSED:
if tokens[token_id] != token:
logger.warning(
- f'replacing token {token_id}: {tokens[token_id].decode("utf-8")!r}' \
- +' -> {token.decode("utf-8")!r}')
+ f"replacing token {token_id}: {tokens[token_id].decode('utf-8')!r}"
+ f" -> {token.decode('utf-8')!r}"
+ )
tokens[token_id] = token
scores[token_id] = -1000.0
toktypes[token_id] = SentencePieceTokenTypes.USER_DEFINED
@@ -2501,27 +3453,35 @@ class InternLM2Model(Model):
# TODO: this is a hack, should be fixed
# https://github.com/ggerganov/llama.cpp/pull/6745#issuecomment-2067687048
special_vocab.special_token_ids["eos"] = chat_eos_token_id
- logger.warning(f"Replace eos:{old_eos} with a special token:{chat_eos_token_id}"
- " in chat mode so that the conversation can end normally.")
+ logger.warning(
+ f"Replace eos:{old_eos} with a special token:{chat_eos_token_id}"
+ " in chat mode so that the conversation can end normally.")
special_vocab.add_to_gguf(self.gguf_writer)
def set_gguf_parameters(self):
- self.gguf_writer.add_context_length(self.hparams["max_position_embeddings"])
+ self.gguf_writer.add_context_length(
+ self.hparams["max_position_embeddings"])
self.gguf_writer.add_block_count(self.hparams["num_hidden_layers"])
self.gguf_writer.add_embedding_length(self.hparams["hidden_size"])
- self.gguf_writer.add_feed_forward_length(self.hparams["intermediate_size"])
+ self.gguf_writer.add_feed_forward_length(
+ self.hparams["intermediate_size"])
self.gguf_writer.add_rope_freq_base(self.hparams["rope_theta"])
self.gguf_writer.add_head_count(self.hparams["num_attention_heads"])
self.gguf_writer.add_layer_norm_rms_eps(self.hparams["rms_norm_eps"])
self.gguf_writer.add_head_count_kv(self.hparams["num_key_value_heads"])
self.gguf_writer.add_file_type(self.ftype)
- if self.hparams.get("rope_scaling") is not None and "factor" in self.hparams["rope_scaling"]:
+ if self.hparams.get(
+ "rope_scaling"
+ ) is not None and "factor" in self.hparams["rope_scaling"]:
if self.hparams["rope_scaling"].get("type") == "linear":
- self.gguf_writer.add_rope_scaling_type(gguf.RopeScalingType.LINEAR)
- self.gguf_writer.add_rope_scaling_factor(self.hparams["rope_scaling"]["factor"])
+ self.gguf_writer.add_rope_scaling_type(
+ gguf.RopeScalingType.LINEAR)
+ self.gguf_writer.add_rope_scaling_factor(
+ self.hparams["rope_scaling"]["factor"])
- def modify_tensors(self, data_torch: Tensor, name: str, bid: int | None) -> Iterable[tuple[str, Tensor]]:
+ def modify_tensors(self, data_torch: Tensor, name: str,
+ bid: int | None) -> Iterable[tuple[str, Tensor]]:
num_heads = self.hparams["num_attention_heads"]
num_kv_heads = self.hparams["num_key_value_heads"]
n_embd = self.hparams["hidden_size"]
@@ -2533,11 +3493,13 @@ class InternLM2Model(Model):
qkv = data_torch
qkv = qkv.reshape((num_groups, q_per_kv + 2, head_dim, n_embd))
- q, k, v = qkv[:, : q_per_kv], qkv[:, -2], qkv[:, -1]
+ q, k, v = qkv[:, :q_per_kv], qkv[:, -2], qkv[:, -1]
# The model weights of q and k require additional reshape.
- q = LlamaModel.permute(q.reshape((-1, q.shape[-1])), num_heads, num_heads)
- k = LlamaModel.permute(k.reshape((-1, k.shape[-1])), num_heads, num_kv_heads)
+ q = LlamaModel.permute(q.reshape((-1, q.shape[-1])), num_heads,
+ num_heads)
+ k = LlamaModel.permute(k.reshape((-1, k.shape[-1])), num_heads,
+ num_kv_heads)
v = v.reshape((-1, v.shape[-1]))
return [
@@ -2549,7 +3511,7 @@ class InternLM2Model(Model):
return [(self.map_tensor_name(name), data_torch)]
[email protected]("BertModel", "CamembertModel")
[email protected]("BertModel", "BertForMaskedLM", "CamembertModel")
class BertModel(Model):
model_arch = gguf.MODEL_ARCH.BERT
@@ -2574,14 +3536,16 @@ class BertModel(Model):
# get pooling type
if pooling_path is not None:
- with open(self.dir_model / pooling_path / "config.json", encoding="utf-8") as f:
+ with open(self.dir_model / pooling_path / "config.json",
+ encoding="utf-8") as f:
pooling = json.load(f)
if pooling["pooling_mode_mean_tokens"]:
pooling_type = gguf.PoolingType.MEAN
elif pooling["pooling_mode_cls_token"]:
pooling_type = gguf.PoolingType.CLS
else:
- raise NotImplementedError("Only MEAN and CLS pooling types supported")
+ raise NotImplementedError(
+ "Only MEAN and CLS pooling types supported")
self.gguf_writer.add_pooling_type(pooling_type)
def set_vocab(self):
@@ -2590,7 +3554,9 @@ class BertModel(Model):
# we need this to validate the size of the token_type embeddings
# though currently we are passing all zeros to the token_type embeddings
- self.gguf_writer.add_token_type_count(2) # "Sequence A" or "Sequence B"
+ # "Sequence A" or "Sequence B"
+ self.gguf_writer.add_token_type_count(
+ self.hparams.get("type_vocab_size", 1))
# convert to phantom space vocab
def phantom(tok):
@@ -2599,6 +3565,7 @@ class BertModel(Model):
if tok.startswith("##"):
return tok[2:]
return "\u2581" + tok
+
tokens = list(map(phantom, tokens))
# add vocab to gguf
@@ -2611,16 +3578,81 @@ class BertModel(Model):
special_vocab = gguf.SpecialVocab(self.dir_model, n_vocab=len(tokens))
special_vocab.add_to_gguf(self.gguf_writer)
- def modify_tensors(self, data_torch: Tensor, name: str, bid: int | None) -> Iterable[tuple[str, Tensor]]:
+ def modify_tensors(self, data_torch: Tensor, name: str,
+ bid: int | None) -> Iterable[tuple[str, Tensor]]:
del bid # unused
+ if name.startswith("bert."):
+ name = name[5:]
+
+ if name.endswith(".gamma"):
+ name = name[:-6] + ".weight"
+
+ if name.endswith(".beta"):
+ name = name[:-5] + ".bias"
+
# we are only using BERT for embeddings so we don't need the pooling layer
- if name in ("embeddings.position_ids", "pooler.dense.weight", "pooler.dense.bias"):
- return [] # we don't need these
+ if name in ("embeddings.position_ids", "pooler.dense.weight",
+ "pooler.dense.bias"):
+ return [] # we don't need these
+
+ if name.startswith("cls.predictions"):
+ return []
+
+ if name.startswith("cls.seq_relationship"):
+ return []
return [(self.map_tensor_name(name), data_torch)]
[email protected]("RobertaModel")
+class RobertaModel(BertModel):
+ model_arch = gguf.MODEL_ARCH.BERT
+
+ def __init__(self, *args, **kwargs):
+ super().__init__(*args, **kwargs)
+
+ # we need the pad_token_id to know how to chop down position_embd matrix
+ if (pad_token_id := self.hparams.get("pad_token_id")) is not None:
+ self._position_offset = 1 + pad_token_id
+ if "max_position_embeddings" in self.hparams:
+ self.hparams[
+ "max_position_embeddings"] -= self._position_offset
+ else:
+ self._position_offset = None
+
+ def set_vocab(self):
+ """Support BPE tokenizers for roberta models"""
+ bpe_tok_path = self.dir_model / "tokenizer.json"
+ if bpe_tok_path.exists():
+ self._set_vocab_gpt2()
+ self.gguf_writer.add_add_bos_token(True)
+ self.gguf_writer.add_add_eos_token(True)
+
+ # we need this to validate the size of the token_type embeddings
+ # though currently we are passing all zeros to the token_type embeddings
+ # "Sequence A" or "Sequence B"
+ self.gguf_writer.add_token_type_count(
+ self.hparams.get("type_vocab_size", 1))
+
+ else:
+ return super().set_vocab()
+
+ def modify_tensors(self, data_torch: Tensor, name: str,
+ bid: int | None) -> Iterable[tuple[str, Tensor]]:
+ # if name starts with "roberta.", remove the prefix
+ # e.g. https://huggingface.co/BAAI/bge-reranker-v2-m3/tree/main
+ if name.startswith("roberta."):
+ name = name[8:]
+
+ # position embeddings start at pad_token_id + 1, so just chop down the weight tensor
+ if name == "embeddings.position_embeddings.weight":
+ if self._position_offset is not None:
+ data_torch = data_torch[self._position_offset:, :]
+
+ return super().modify_tensors(data_torch, name, bid)
+
+
@Model.register("NomicBertModel")
class NomicBertModel(BertModel):
model_arch = gguf.MODEL_ARCH.NOMIC_BERT
@@ -2662,7 +3694,8 @@ class XLMRobertaModel(BertModel):
if (pad_token_id := self.hparams.get("pad_token_id")) is not None:
self._position_offset = 1 + pad_token_id
if "max_position_embeddings" in self.hparams:
- self.hparams["max_position_embeddings"] -= self._position_offset
+ self.hparams[
+ "max_position_embeddings"] -= self._position_offset
else:
self._position_offset = None
@@ -2690,7 +3723,9 @@ class XLMRobertaModel(BertModel):
vocab_size = self.hparams.get('vocab_size', tokenizer.vocab_size())
- tokens: list[bytes] = [f"[PAD{i}]".encode("utf-8") for i in range(vocab_size)]
+ tokens: list[bytes] = [
+ f"[PAD{i}]".encode("utf-8") for i in range(vocab_size)
+ ]
scores: list[float] = [-10000.0] * vocab_size
toktypes: list[int] = [SentencePieceTokenTypes.UNUSED] * vocab_size
@@ -2715,7 +3750,9 @@ class XLMRobertaModel(BertModel):
if vocab_size > len(tokens):
pad_count = vocab_size - len(tokens)
- logger.debug(f"Padding vocab with {pad_count} token(s) - [PAD1] through [PAD{pad_count}]")
+ logger.debug(
+ f"Padding vocab with {pad_count} token(s) - [PAD1] through [PAD{pad_count}]"
+ )
for i in range(1, pad_count + 1):
tokens.append(bytes(f"[PAD{i}]", encoding="utf-8"))
scores.append(-1000.0)
@@ -2737,7 +3774,8 @@ class XLMRobertaModel(BertModel):
self.gguf_writer.add_token_scores(scores)
self.gguf_writer.add_token_types(toktypes)
self.gguf_writer.add_add_space_prefix(add_prefix)
- self.gguf_writer.add_token_type_count(self.hparams.get("type_vocab_size", 1))
+ self.gguf_writer.add_token_type_count(
+ self.hparams.get("type_vocab_size", 1))
self.gguf_writer.add_remove_extra_whitespaces(remove_whitespaces)
if precompiled_charsmap:
self.gguf_writer.add_precompiled_charsmap(precompiled_charsmap)
@@ -2748,7 +3786,8 @@ class XLMRobertaModel(BertModel):
self.gguf_writer.add_add_bos_token(True)
self.gguf_writer.add_add_eos_token(True)
- def modify_tensors(self, data_torch: Tensor, name: str, bid: int | None) -> Iterable[tuple[str, Tensor]]:
+ def modify_tensors(self, data_torch: Tensor, name: str,
+ bid: int | None) -> Iterable[tuple[str, Tensor]]:
# if name starts with "roberta.", remove the prefix
# e.g. https://huggingface.co/BAAI/bge-reranker-v2-m3/tree/main
if name.startswith("roberta."):
@@ -2757,7 +3796,7 @@ class XLMRobertaModel(BertModel):
# position embeddings start at pad_token_id + 1, so just chop down the weight tensor
if name == "embeddings.position_embeddings.weight":
if self._position_offset is not None:
- data_torch = data_torch[self._position_offset:,:]
+ data_torch = data_torch[self._position_offset:, :]
return super().modify_tensors(data_torch, name, bid)
@@ -2770,13 +3809,15 @@ class GemmaModel(Model):
self._set_vocab_sentencepiece()
# TODO: these special tokens should be exported only for the CodeGemma family
- special_vocab = gguf.SpecialVocab(self.dir_model, load_merges=False,
- special_token_types = ['prefix', 'suffix', 'middle', 'fsep', 'eot'])
+ special_vocab = gguf.SpecialVocab(
+ self.dir_model,
+ load_merges=False,
+ special_token_types=['prefix', 'suffix', 'middle', 'fsep', 'eot'])
special_vocab._set_special_token("prefix", 67)
special_vocab._set_special_token("suffix", 69)
special_vocab._set_special_token("middle", 68)
- special_vocab._set_special_token("fsep", 70)
- special_vocab._set_special_token("eot", 107)
+ special_vocab._set_special_token("fsep", 70)
+ special_vocab._set_special_token("eot", 107)
special_vocab.chat_template = None # do not add it twice
special_vocab.add_to_gguf(self.gguf_writer)
@@ -2792,19 +3833,23 @@ class GemmaModel(Model):
self.gguf_writer.add_feed_forward_length(hparams["intermediate_size"])
self.gguf_writer.add_head_count(hparams["num_attention_heads"])
self.gguf_writer.add_head_count_kv(
- self.hparams["num_key_value_heads"] if "num_key_value_heads" in hparams else hparams["num_attention_heads"])
+ self.hparams["num_key_value_heads"] if "num_key_value_heads" in
+ hparams else hparams["num_attention_heads"])
self.gguf_writer.add_layer_norm_rms_eps(self.hparams["rms_norm_eps"])
self.gguf_writer.add_key_length(hparams["head_dim"])
self.gguf_writer.add_value_length(hparams["head_dim"])
self.gguf_writer.add_file_type(self.ftype)
- def modify_tensors(self, data_torch: Tensor, name: str, bid: int | None) -> Iterable[tuple[str, Tensor]]:
+ def modify_tensors(self, data_torch: Tensor, name: str,
+ bid: int | None) -> Iterable[tuple[str, Tensor]]:
del bid # unused
# lm_head is not used in llama.cpp, while autoawq will include this tensor in model
# To prevent errors, skip loading lm_head.weight.
if name == "lm_head.weight":
- logger.debug(f"Skipping get tensor {name!r} in safetensors so that convert can end normally.")
+ logger.debug(
+ f"Skipping get tensor {name!r} in safetensors so that convert can end normally."
+ )
return []
if name.endswith("norm.weight"):
@@ -2832,26 +3877,28 @@ class Gemma2Model(Model):
self.gguf_writer.add_feed_forward_length(hparams["intermediate_size"])
self.gguf_writer.add_head_count(hparams["num_attention_heads"])
self.gguf_writer.add_head_count_kv(
- self.hparams["num_key_value_heads"] if "num_key_value_heads" in hparams else hparams["num_attention_heads"])
+ self.hparams["num_key_value_heads"] if "num_key_value_heads" in
+ hparams else hparams["num_attention_heads"])
self.gguf_writer.add_layer_norm_rms_eps(self.hparams["rms_norm_eps"])
self.gguf_writer.add_key_length(hparams["head_dim"])
self.gguf_writer.add_value_length(hparams["head_dim"])
self.gguf_writer.add_file_type(self.ftype)
self.gguf_writer.add_attn_logit_softcapping(
- self.hparams["attn_logit_softcapping"]
- )
+ self.hparams["attn_logit_softcapping"])
self.gguf_writer.add_final_logit_softcapping(
- self.hparams["final_logit_softcapping"]
- )
+ self.hparams["final_logit_softcapping"])
self.gguf_writer.add_sliding_window(self.hparams["sliding_window"])
- def modify_tensors(self, data_torch: Tensor, name: str, bid: int | None) -> Iterable[tuple[str, Tensor]]:
+ def modify_tensors(self, data_torch: Tensor, name: str,
+ bid: int | None) -> Iterable[tuple[str, Tensor]]:
del bid # unused
# lm_head is not used in llama.cpp, while autoawq will include this tensor in model
# To prevent errors, skip loading lm_head.weight.
if name == "lm_head.weight":
- logger.debug(f"Skipping get tensor {name!r} in safetensors so that convert can end normally.")
+ logger.debug(
+ f"Skipping get tensor {name!r} in safetensors so that convert can end normally."
+ )
return []
if name.endswith("norm.weight"):
@@ -2876,13 +3923,17 @@ class Rwkv6Model(Model):
tokens: list[bytes] = ['<s>'.encode("utf-8")]
toktypes: list[int] = [gguf.TokenType.CONTROL]
- with open(self.dir_model / "rwkv_vocab_v20230424.txt", "r", encoding="utf-8") as f:
+ with open(self.dir_model / "rwkv_vocab_v20230424.txt",
+ "r",
+ encoding="utf-8") as f:
lines = f.readlines()
for line in lines:
parts = line.split(' ')
assert len(parts) >= 3
- token, token_len = ast.literal_eval(' '.join(parts[1:-1])), int(parts[-1])
- token = token.encode("utf-8") if isinstance(token, str) else token
+ token, token_len = ast.literal_eval(' '.join(
+ parts[1:-1])), int(parts[-1])
+ token = token.encode("utf-8") if isinstance(token,
+ str) else token
assert isinstance(token, bytes)
assert len(token) == token_len
token_text: str = repr(token)[2:-1] # "b'\xff'" -> "\xff"
@@ -2909,8 +3960,9 @@ class Rwkv6Model(Model):
hidden_size = self.hparams["hidden_size"]
layer_norm_eps = self.hparams["layer_norm_epsilon"]
rescale_every_n_layers = self.hparams["rescale_every"]
- intermediate_size = self.hparams["intermediate_size"] \
- if self.hparams["intermediate_size"] is not None else int((hidden_size * 3.5) // 32 * 32)
+ intermediate_size = self.hparams["intermediate_size"] if self.hparams[
+ "intermediate_size"] is not None else int(
+ (hidden_size * 3.5) // 32 * 32)
time_mix_extra_dim = 64 if hidden_size == 4096 else 32
time_decay_extra_dim = 128 if hidden_size == 4096 else 64
@@ -2929,7 +3981,10 @@ class Rwkv6Model(Model):
# required by llama.cpp, unused
self.gguf_writer.add_head_count(0)
- def modify_tensors(self, data_torch: Tensor, name: str, bid: int | None) -> Iterable[tuple[str, Tensor]]:
+ lerp_weights: dict[int, dict[str, Tensor]] = {}
+
+ def modify_tensors(self, data_torch: Tensor, name: str,
+ bid: int | None) -> Iterable[tuple[str, Tensor]]:
new_name = self.map_tensor_name(name)
if not (new_name.endswith(".weight") or new_name.endswith(".bias")):
@@ -2943,15 +3998,102 @@ class Rwkv6Model(Model):
if new_name.endswith("time_mix_w2.weight"):
data_torch = data_torch.permute(0, 2, 1)
- rescale_every_n_layers = self.hparams["rescale_every"]
- if rescale_every_n_layers > 0:
- if new_name.endswith("time_mix_output.weight") or new_name.endswith("channel_mix_value.weight"):
- data_torch = data_torch.div_(2 ** int(bid // rescale_every_n_layers))
+ if new_name.endswith("time_mix_decay.weight") or "lerp" in new_name:
+ data_torch = data_torch.squeeze()
+
+ try:
+ rescale_every_n_layers = self.hparams["rescale_every"]
+ if rescale_every_n_layers > 0:
+ if new_name.endswith(
+ "time_mix_output.weight") or new_name.endswith(
+ "channel_mix_value.weight"):
+ data_torch = data_torch.div_(2**int(
+ bid // rescale_every_n_layers))
+ except KeyError:
+ pass
+
+ # concat time_mix_lerp weights to reduce some cpu overhead
+ # also reduces the number of tensors in the model
+ if bid is not None and "time_mix_lerp" in new_name and "time_mix_lerp_x" not in new_name:
+ try:
+ self.lerp_weights[bid][new_name] = data_torch
+ except KeyError:
+ self.lerp_weights[bid] = {new_name: data_torch}
+ if all(f"blk.{bid}.time_mix_lerp_{i}.weight" in
+ self.lerp_weights[bid].keys()
+ for i in ["w", "k", "v", "r", "g"]):
+ new_name = f"blk.{bid}.time_mix_lerp_fused.weight"
+ data = torch.stack([
+ self.lerp_weights[bid]
+ [f"blk.{bid}.time_mix_lerp_{i}.weight"].unsqueeze(0)
+ for i in ["w", "k", "v", "r", "g"]
+ ],
+ dim=0).unsqueeze(1)
+ yield (new_name, data)
+ return
yield (new_name, data_torch)
[email protected]("MambaForCausalLM", "MambaLMHeadModel", "FalconMambaForCausalLM")
[email protected]("RWKV6Qwen2ForCausalLM")
+class RWKV6Qwen2Model(Rwkv6Model):
+ model_arch = gguf.MODEL_ARCH.RWKV6QWEN2
+
+ def set_vocab(self):
+ try:
+ self._set_vocab_sentencepiece()
+ except FileNotFoundError:
+ self._set_vocab_gpt2()
+
+ def set_gguf_parameters(self):
+ block_count = self.hparams["num_hidden_layers"]
+ num_attention_heads = self.hparams["num_attention_heads"]
+ num_key_value_heads = self.hparams["num_key_value_heads"]
+ hidden_size = self.hparams["hidden_size"]
+ head_size = hidden_size // num_attention_heads
+ rms_norm_eps = self.hparams["rms_norm_eps"]
+ intermediate_size = self.hparams["intermediate_size"]
+ time_mix_extra_dim = 64 if hidden_size >= 4096 else 32
+ time_decay_extra_dim = 128 if hidden_size >= 4096 else 64
+
+ # RWKV isn't context limited
+ self.gguf_writer.add_context_length(1048576)
+ self.gguf_writer.add_embedding_length(hidden_size)
+ self.gguf_writer.add_block_count(block_count)
+ self.gguf_writer.add_wkv_head_size(head_size)
+ self.gguf_writer.add_time_mix_extra_dim(time_mix_extra_dim)
+ self.gguf_writer.add_time_decay_extra_dim(time_decay_extra_dim)
+ self.gguf_writer.add_feed_forward_length(intermediate_size)
+ self.gguf_writer.add_file_type(self.ftype)
+
+ # special parameters for time_mixing in RWKV6QWEN2
+ self.gguf_writer.add_layer_norm_rms_eps(rms_norm_eps)
+ self.gguf_writer.add_token_shift_count(1)
+ # RWKV6QWEN2 use grouped key/value like GQA
+ self.gguf_writer.add_head_count_kv(num_key_value_heads)
+
+ # required by llama.cpp, unused
+ self.gguf_writer.add_head_count(0)
+
+ def modify_tensors(self, data_torch: Tensor, name: str,
+ bid: int | None) -> Iterable[tuple[str, Tensor]]:
+ for new_name, data in super().modify_tensors(data_torch, name, bid):
+ if "time_mix_w1" in new_name or "time_mix_w2" in new_name:
+ data = data.view(5, -1, data.shape[-1])
+ # rwkv6qwen2 has a different order of rkvwg instead of the original wkvrg
+ # permute them here to avoid code changes
+ data = torch.stack(
+ [data[3], data[1], data[2], data[0], data[4]],
+ dim=0).view(-1, data.shape[-1])
+ if "w2" in new_name:
+ data = data.view(5, -1, data.shape[-1])
+ yield (new_name, data)
+ continue
+ yield (new_name, data)
+
+
[email protected]("MambaForCausalLM", "MambaLMHeadModel",
+ "FalconMambaForCausalLM")
class MambaModel(Model):
model_arch = gguf.MODEL_ARCH.MAMBA
@@ -2973,37 +4115,48 @@ class MambaModel(Model):
self._set_vocab_builtin("gpt-neox", vocab_size)
def set_gguf_parameters(self):
- d_model = self.find_hparam(["hidden_size", "d_model"])
- d_conv = self.find_hparam(["conv_kernel", "d_conv"], optional=True) or 4
- d_inner = self.find_hparam(["intermediate_size", "d_inner"], optional=True) or 2 * d_model
- d_state = self.find_hparam(["state_size", "d_state"], optional=True) or 16
+ d_model = self.find_hparam(["hidden_size", "d_model"])
+ d_conv = self.find_hparam(["conv_kernel", "d_conv"],
+ optional=True) or 4
+ d_inner = self.find_hparam(["intermediate_size", "d_inner"],
+ optional=True) or 2 * d_model
+ d_state = self.find_hparam(["state_size", "d_state"],
+ optional=True) or 16
# ceiling division
- # ref: https://stackoverflow.com/a/17511341/22827863
- dt_rank = self.find_hparam(["time_step_rank", "dt_rank"], optional=True) or -(d_model // -16)
- rms_norm_eps = self.find_hparam(["layer_norm_epsilon", "rms_norm_eps"], optional=True) or 1e-5
+ dt_rank = self.find_hparam(["time_step_rank", "dt_rank"],
+ optional=True) or -(d_model // -16)
+ rms_norm_eps = self.find_hparam(["layer_norm_epsilon", "rms_norm_eps"],
+ optional=True) or 1e-5
use_dt_b_c_norm = False
# For falconmamba we do apply RMS norm on B / DT and C layers
- if self.find_hparam(["model_type"], optional=True) in ("falcon_mamba",):
+ if self.find_hparam(["model_type"],
+ optional=True) in ("falcon_mamba", ):
use_dt_b_c_norm = True
# Fail early for models which don't have a block expansion factor of 2
assert d_inner == 2 * d_model
- self.gguf_writer.add_context_length(2**20) # arbitrary value; for those who use the default
+ self.gguf_writer.add_context_length(
+ 2**20) # arbitrary value; for those who use the default
self.gguf_writer.add_embedding_length(d_model)
- self.gguf_writer.add_feed_forward_length(0) # unused, but seemingly required when loading
- self.gguf_writer.add_head_count(0) # unused, but seemingly required when loading
+ self.gguf_writer.add_feed_forward_length(
+ 0) # unused, but seemingly required when loading
+ self.gguf_writer.add_head_count(
+ 0) # unused, but seemingly required when loading
self.gguf_writer.add_block_count(self.block_count)
self.gguf_writer.add_ssm_conv_kernel(d_conv)
self.gguf_writer.add_ssm_inner_size(d_inner)
self.gguf_writer.add_ssm_state_size(d_state)
self.gguf_writer.add_ssm_time_step_rank(dt_rank)
self.gguf_writer.add_layer_norm_rms_eps(rms_norm_eps)
- self.gguf_writer.add_ssm_dt_b_c_rms(use_dt_b_c_norm)
+ self.gguf_writer.add_ssm_dt_b_c_rms(
+ use_dt_b_c_norm
+ ) # For classic Mamba we don't apply rms norm on B / DT layers
self.gguf_writer.add_file_type(self.ftype)
_tok_embd = None
- def modify_tensors(self, data_torch: Tensor, name: str, bid: int | None) -> Iterable[tuple[str, Tensor]]:
+ def modify_tensors(self, data_torch: Tensor, name: str,
+ bid: int | None) -> Iterable[tuple[str, Tensor]]:
del bid # unused
output_name = self.format_tensor_name(gguf.MODEL_TENSOR.OUTPUT)
@@ -3018,7 +4171,9 @@ class MambaModel(Model):
# assuming token_embd.weight is seen before output.weight
if self._tok_embd is not None and new_name == output_name:
if torch.equal(self._tok_embd, data_torch):
- logger.debug(f"{output_name} is equivalent to {tok_embd_name}, omitting")
+ logger.debug(
+ f"{output_name} is equivalent to {tok_embd_name}, omitting"
+ )
return []
elif new_name == tok_embd_name:
self._tok_embd = data_torch
@@ -3036,11 +4191,31 @@ class CommandR2Model(Model):
# max_position_embeddings = 8192 in config.json but model was actually
# trained on 128k context length
# aya-23 models don't have model_max_length specified
- self.hparams["max_position_embeddings"] = self.find_hparam(["model_max_length", "max_position_embeddings"])
+ self.hparams["max_position_embeddings"] = self.find_hparam(
+ ["model_max_length", "max_position_embeddings"])
+
+ def set_gguf_parameters(self):
+ super().set_gguf_parameters()
+ self.gguf_writer.add_logit_scale(self.hparams["logit_scale"])
+ self.gguf_writer.add_rope_scaling_type(gguf.RopeScalingType.NONE)
+
+
[email protected]("Cohere2ForCausalLM")
+class Cohere2Model(Model):
+ model_arch = gguf.MODEL_ARCH.COHERE2
def set_gguf_parameters(self):
super().set_gguf_parameters()
+
self.gguf_writer.add_logit_scale(self.hparams["logit_scale"])
+ self.gguf_writer.add_sliding_window(self.hparams["sliding_window"])
+ self.gguf_writer.add_vocab_size(self.hparams["vocab_size"])
+
+ rotary_pct = self.hparams["rotary_pct"]
+ hidden_size = self.hparams["hidden_size"]
+ num_attention_heads = self.hparams["num_attention_heads"]
+ self.gguf_writer.add_rope_dimension_count(
+ int(rotary_pct * (hidden_size // num_attention_heads)))
self.gguf_writer.add_rope_scaling_type(gguf.RopeScalingType.NONE)
@@ -3058,7 +4233,8 @@ class OlmoModel(Model):
# Same as super class, but permuting q_proj, k_proj
# Copied from: LlamaModel
- def modify_tensors(self, data_torch: Tensor, name: str, bid: int | None) -> Iterable[tuple[str, Tensor]]:
+ def modify_tensors(self, data_torch: Tensor, name: str,
+ bid: int | None) -> Iterable[tuple[str, Tensor]]:
del bid # unused
n_head = self.hparams["num_attention_heads"]
@@ -3090,7 +4266,8 @@ class OlmoeModel(Model):
_experts: list[dict[str, Tensor]] | None = None
# Copied from: Qwen2MoeModel
- def modify_tensors(self, data_torch: Tensor, name: str, bid: int | None) -> Iterable[tuple[str, Tensor]]:
+ def modify_tensors(self, data_torch: Tensor, name: str,
+ bid: int | None) -> Iterable[tuple[str, Tensor]]:
# process the experts separately
if name.find("experts") != -1:
n_experts = self.hparams["num_experts"]
@@ -3162,7 +4339,9 @@ class JinaBertV2Model(BertModel):
def set_vocab(self):
tokenizer_class = 'BertTokenizer'
- with open(self.dir_model / "tokenizer_config.json", "r", encoding="utf-8") as f:
+ with open(self.dir_model / "tokenizer_config.json",
+ "r",
+ encoding="utf-8") as f:
tokenizer_class = json.load(f)['tokenizer_class']
if tokenizer_class == 'BertTokenizer':
@@ -3171,11 +4350,14 @@ class JinaBertV2Model(BertModel):
self._set_vocab_gpt2()
self.gguf_writer.add_token_type_count(2)
else:
- raise NotImplementedError(f'Tokenizer {tokenizer_class} is not supported for JinaBertModel')
+ raise NotImplementedError(
+ f'Tokenizer {tokenizer_class} is not supported for JinaBertModel'
+ )
self.gguf_writer.add_add_bos_token(True)
self.gguf_writer.add_add_eos_token(True)
- def modify_tensors(self, data_torch: Tensor, name: str, bid: int | None) -> Iterable[tuple[str, Tensor]]:
+ def modify_tensors(self, data_torch: Tensor, name: str,
+ bid: int | None) -> Iterable[tuple[str, Tensor]]:
# if name starts with "bert.", remove the prefix
# e.g. https://huggingface.co/jinaai/jina-reranker-v1-tiny-en
if name.startswith("bert."):
@@ -3205,11 +4387,14 @@ class OpenELMModel(Model):
self._num_kv_heads: list[int] = self.hparams["num_kv_heads"]
self._num_query_heads: list[int] = self.hparams["num_query_heads"]
self._ffn_dims: list[int] = [
- OpenELMModel._make_divisible(multiplier * self._n_embd, ffn_dim_divisor)
+ OpenELMModel._make_divisible(multiplier * self._n_embd,
+ ffn_dim_divisor)
for multiplier in ffn_multipliers
]
- assert isinstance(self._num_kv_heads, list) and isinstance(self._num_kv_heads[0], int)
- assert isinstance(self._num_query_heads, list) and isinstance(self._num_query_heads[0], int)
+ assert isinstance(self._num_kv_heads, list) and isinstance(
+ self._num_kv_heads[0], int)
+ assert isinstance(self._num_query_heads, list) and isinstance(
+ self._num_query_heads[0], int)
# Uses the tokenizer from meta-llama/Llama-2-7b-hf
def set_vocab(self):
@@ -3246,13 +4431,16 @@ class OpenELMModel(Model):
return super().find_hparam(keys, optional)
- def modify_tensors(self, data_torch: Tensor, name: str, bid: int | None) -> Iterable[tuple[str, Tensor]]:
+ def modify_tensors(self, data_torch: Tensor, name: str,
+ bid: int | None) -> Iterable[tuple[str, Tensor]]:
# split ff
if bid is not None and name == f"transformer.layers.{bid}.ffn.proj_1.weight":
ff_dim = self._ffn_dims[bid]
- yield (self.format_tensor_name(gguf.MODEL_TENSOR.FFN_GATE, bid), data_torch[:ff_dim])
- yield (self.format_tensor_name(gguf.MODEL_TENSOR.FFN_UP, bid), data_torch[ff_dim:])
+ yield (self.format_tensor_name(gguf.MODEL_TENSOR.FFN_GATE,
+ bid), data_torch[:ff_dim])
+ yield (self.format_tensor_name(gguf.MODEL_TENSOR.FFN_UP,
+ bid), data_torch[ff_dim:])
return
yield (self.map_tensor_name(name), data_torch)
@@ -3280,7 +4468,9 @@ class ArcticModel(Model):
vocab_size = self.hparams.get('vocab_size', tokenizer.vocab_size())
- tokens: list[bytes] = [f"[PAD{i}]".encode("utf-8") for i in range(vocab_size)]
+ tokens: list[bytes] = [
+ f"[PAD{i}]".encode("utf-8") for i in range(vocab_size)
+ ]
scores: list[float] = [-10000.0] * vocab_size
toktypes: list[int] = [SentencePieceTokenTypes.UNUSED] * vocab_size
@@ -3312,11 +4502,14 @@ class ArcticModel(Model):
tokenizer_config_json = json.load(f)
if "added_tokens_decoder" in tokenizer_config_json:
- added_tokens_decoder = tokenizer_config_json["added_tokens_decoder"]
+ added_tokens_decoder = tokenizer_config_json[
+ "added_tokens_decoder"]
for token_id, token_json in added_tokens_decoder.items():
token_id = int(token_id)
if token_id >= vocab_size:
- logger.debug(f'ignore token {token_id}: id is out of range, max={vocab_size - 1}')
+ logger.debug(
+ f'ignore token {token_id}: id is out of range, max={vocab_size - 1}'
+ )
continue
token_content = token_json["content"]
@@ -3326,15 +4519,17 @@ class ArcticModel(Model):
# Map unk_token to UNKNOWN, other special tokens to CONTROL
# Set the score to 0.0 as in the original tokenizer.model
if ("special" in token_json) and token_json["special"]:
- if token_content == tokenizer_config_json["unk_token"]:
+ if token_content == tokenizer_config_json[
+ "unk_token"]:
token_type = SentencePieceTokenTypes.UNKNOWN
else:
token_type = SentencePieceTokenTypes.CONTROL
token_score = 0.0
logger.info(
- f"Setting added token {token_id} to '{token_content}' (type: {token_type},"
- " score: {token_score:.2f})")
+ f"Setting added token {token_id} to '{token_content}' (type: {token_type}, "
+ f"score: {token_score:.2f})"
+ )
tokens[token_id] = token_content.encode("utf-8")
toktypes[token_id] = token_type
scores[token_id] = token_score
@@ -3352,11 +4547,13 @@ class ArcticModel(Model):
super().set_gguf_parameters()
hparams = self.hparams
self.gguf_writer.add_vocab_size(hparams["vocab_size"])
- self.gguf_writer.add_rope_dimension_count(hparams["hidden_size"] // hparams["num_attention_heads"])
+ self.gguf_writer.add_rope_dimension_count(
+ hparams["hidden_size"] // hparams["num_attention_heads"])
_experts: list[dict[str, Tensor]] | None = None
- def modify_tensors(self, data_torch: Tensor, name: str, bid: int | None) -> Iterable[tuple[str, Tensor]]:
+ def modify_tensors(self, data_torch: Tensor, name: str,
+ bid: int | None) -> Iterable[tuple[str, Tensor]]:
n_head = self.hparams["num_attention_heads"]
n_kv_head = self.hparams.get("num_key_value_heads")
@@ -3411,7 +4608,102 @@ class ArcticModel(Model):
raise ValueError(f"Unprocessed experts: {experts}")
[email protected]("DeepseekForCausalLM")
+class DeepseekModel(Model):
+ model_arch = gguf.MODEL_ARCH.DEEPSEEK
+
+ def set_vocab(self):
+ try:
+ self._set_vocab_sentencepiece()
+ except FileNotFoundError:
+ self._set_vocab_gpt2()
+
+ def set_gguf_parameters(self):
+ super().set_gguf_parameters()
+ hparams = self.hparams
+ if "head_dim" in hparams:
+ rope_dim = hparams["head_dim"]
+ else:
+ rope_dim = hparams["hidden_size"] // hparams["num_attention_heads"]
+
+ self.gguf_writer.add_rope_dimension_count(rope_dim)
+ self.gguf_writer.add_rope_scaling_type(gguf.RopeScalingType.NONE)
+ self.gguf_writer.add_leading_dense_block_count(
+ hparams["first_k_dense_replace"])
+ self.gguf_writer.add_vocab_size(hparams["vocab_size"])
+ self.gguf_writer.add_expert_feed_forward_length(
+ hparams["moe_intermediate_size"])
+ self.gguf_writer.add_expert_weights_scale(1.0)
+ self.gguf_writer.add_expert_count(hparams["n_routed_experts"])
+ self.gguf_writer.add_expert_shared_count(hparams["n_shared_experts"])
+
+ _experts: list[dict[str, Tensor]] | None = None
+
+ @staticmethod
+ def permute(weights: Tensor, n_head: int, n_head_kv: int | None):
+ if n_head_kv is not None and n_head != n_head_kv:
+ n_head = n_head_kv
+ return (weights.reshape(n_head, 2, weights.shape[0] // n_head // 2,
+ *weights.shape[1:]).swapaxes(1, 2).reshape(
+ weights.shape))
+
+ def modify_tensors(self, data_torch: Tensor, name: str,
+ bid: int | None) -> Iterable[tuple[str, Tensor]]:
+ n_head = self.hparams["num_attention_heads"]
+ n_kv_head = self.hparams.get("num_key_value_heads")
+
+ if name.endswith(("q_proj.weight", "q_proj.bias")):
+ data_torch = DeepseekModel.permute(data_torch, n_head, n_head)
+ if name.endswith(("k_proj.weight", "k_proj.bias")):
+ data_torch = DeepseekModel.permute(data_torch, n_head, n_kv_head)
+
+ # process the experts separately
+ if name.find("mlp.experts") != -1:
+ n_experts = self.hparams["n_routed_experts"]
+ assert bid is not None
+
+ if self._experts is None:
+ self._experts = [{} for _ in range(self.block_count)]
+
+ self._experts[bid][name] = data_torch
+
+ if len(self._experts[bid]) >= n_experts * 3:
+ tensors: list[tuple[str, Tensor]] = []
+
+ # merge the experts into a single 3d tensor
+ for w_name in ["down_proj", "gate_proj", "up_proj"]:
+ data: list[Tensor] = []
+
+ for xid in range(n_experts):
+ ename = f"model.layers.{bid}.mlp.experts.{xid}.{w_name}.weight"
+ data.append(self._experts[bid][ename])
+ del self._experts[bid][ename]
+
+ data_torch = torch.stack(data, dim=0)
+
+ merged_name = f"model.layers.{bid}.mlp.experts.{w_name}.weight"
+
+ new_name = self.map_tensor_name(merged_name)
+
+ tensors.append((new_name, data_torch))
+ return tensors
+ else:
+ return []
+
+ return [(self.map_tensor_name(name), data_torch)]
+
+ def prepare_tensors(self):
+ super().prepare_tensors()
+
+ if self._experts is not None:
+ # flatten `list[dict[str, Tensor]]` into `list[str]`
+ experts = [k for d in self._experts for k in d.keys()]
+ if len(experts) > 0:
+ raise ValueError(f"Unprocessed experts: {experts}")
+
+
@Model.register("DeepseekV2ForCausalLM")
[email protected]("DeepseekV3ForCausalLM")
class DeepseekV2Model(Model):
model_arch = gguf.MODEL_ARCH.DEEPSEEK2
@@ -3422,30 +4714,64 @@ class DeepseekV2Model(Model):
super().set_gguf_parameters()
hparams = self.hparams
- self.gguf_writer.add_leading_dense_block_count(hparams["first_k_dense_replace"])
+ self.gguf_writer.add_leading_dense_block_count(
+ hparams["first_k_dense_replace"])
self.gguf_writer.add_vocab_size(hparams["vocab_size"])
if "q_lora_rank" in hparams and hparams["q_lora_rank"] is not None:
self.gguf_writer.add_q_lora_rank(hparams["q_lora_rank"])
self.gguf_writer.add_kv_lora_rank(hparams["kv_lora_rank"])
- self.gguf_writer.add_key_length(hparams["qk_nope_head_dim"] + hparams["qk_rope_head_dim"])
+ self.gguf_writer.add_key_length(hparams["qk_nope_head_dim"] +
+ hparams["qk_rope_head_dim"])
self.gguf_writer.add_value_length(hparams["v_head_dim"])
- self.gguf_writer.add_expert_feed_forward_length(hparams["moe_intermediate_size"])
+ self.gguf_writer.add_expert_feed_forward_length(
+ hparams["moe_intermediate_size"])
self.gguf_writer.add_expert_count(hparams["n_routed_experts"])
self.gguf_writer.add_expert_shared_count(hparams["n_shared_experts"])
- self.gguf_writer.add_expert_weights_scale(hparams["routed_scaling_factor"])
+ self.gguf_writer.add_expert_weights_scale(
+ hparams["routed_scaling_factor"])
+ self.gguf_writer.add_expert_weights_norm(hparams["norm_topk_prob"])
+
+ if hparams["scoring_func"] == "sigmoid":
+ self.gguf_writer.add_expert_gating_func(
+ gguf.ExpertGatingFuncType.SIGMOID)
+ elif hparams["scoring_func"] == "softmax":
+ self.gguf_writer.add_expert_gating_func(
+ gguf.ExpertGatingFuncType.SOFTMAX)
+ else:
+ raise ValueError(
+ f"Unsupported scoring_func value: {hparams['scoring_func']}")
+
self.gguf_writer.add_rope_dimension_count(hparams["qk_rope_head_dim"])
- if self.hparams.get("rope_scaling") is not None and "factor" in self.hparams["rope_scaling"]:
+ if self.hparams.get(
+ "rope_scaling"
+ ) is not None and "factor" in self.hparams["rope_scaling"]:
if self.hparams["rope_scaling"].get("type") == "yarn":
- self.gguf_writer.add_rope_scaling_type(gguf.RopeScalingType.YARN)
- self.gguf_writer.add_rope_scaling_factor(self.hparams["rope_scaling"]["factor"])
+ self.gguf_writer.add_rope_scaling_type(
+ gguf.RopeScalingType.YARN)
+ self.gguf_writer.add_rope_scaling_factor(
+ self.hparams["rope_scaling"]["factor"])
self.gguf_writer.add_rope_scaling_orig_ctx_len(
- self.hparams["rope_scaling"]["original_max_position_embeddings"])
- self.gguf_writer.add_rope_scaling_yarn_log_mul(0.1 * hparams["rope_scaling"]["mscale_all_dim"])
+ self.hparams["rope_scaling"]
+ ["original_max_position_embeddings"])
+ self.gguf_writer.add_rope_scaling_yarn_log_mul(
+ 0.1 * hparams["rope_scaling"]["mscale_all_dim"])
_experts: list[dict[str, Tensor]] | None = None
- def modify_tensors(self, data_torch: Tensor, name: str, bid: int | None) -> Iterable[tuple[str, Tensor]]:
+ def modify_tensors(self, data_torch: Tensor, name: str,
+ bid: int | None) -> Iterable[tuple[str, Tensor]]:
+ # rename e_score_correction_bias tensors
+ if name.endswith("e_score_correction_bias"):
+ name = name.replace("e_score_correction_bias",
+ "e_score_correction.bias")
+
+ # skip Multi-Token Prediction (MTP) layers
+ block_count = self.hparams["num_hidden_layers"]
+ match = re.match(r"model.layers.(\d+)", name)
+ if match and int(match.group(1)) >= block_count:
+ return []
+
# process the experts separately
if name.find("mlp.experts") != -1:
n_experts = self.hparams["n_routed_experts"]
@@ -3538,7 +4864,9 @@ class T5Model(Model):
vocab_size = self.hparams.get('vocab_size', tokenizer.vocab_size())
- tokens: list[bytes] = [f"[PAD{i}]".encode("utf-8") for i in range(vocab_size)]
+ tokens: list[bytes] = [
+ f"[PAD{i}]".encode("utf-8") for i in range(vocab_size)
+ ]
scores: list[float] = [-10000.0] * vocab_size
toktypes: list[int] = [SentencePieceTokenTypes.UNUSED] * vocab_size
@@ -3568,7 +4896,9 @@ class T5Model(Model):
for key in added_tokens_json:
token_id = added_tokens_json[key]
if token_id >= vocab_size:
- logger.warning(f'ignore token {token_id}: id is out of range, max={vocab_size - 1}')
+ logger.warning(
+ f'ignore token {token_id}: id is out of range, max={vocab_size - 1}'
+ )
continue
tokens[token_id] = key.encode("utf-8")
@@ -3577,7 +4907,9 @@ class T5Model(Model):
if vocab_size > len(tokens):
pad_count = vocab_size - len(tokens)
- logger.debug(f"Padding vocab with {pad_count} token(s) - [PAD1] through [PAD{pad_count}]")
+ logger.debug(
+ f"Padding vocab with {pad_count} token(s) - [PAD1] through [PAD{pad_count}]"
+ )
for i in range(1, pad_count + 1):
tokens.append(bytes(f"[PAD{i}]", encoding="utf-8"))
scores.append(-1000.0)
@@ -3601,7 +4933,9 @@ class T5Model(Model):
def set_gguf_parameters(self):
if (n_ctx := self.find_hparam(["n_positions"], optional=True)) is None:
- logger.warning("Couldn't find context length in config.json, assuming default value of 512")
+ logger.warning(
+ "Couldn't find context length in config.json, assuming default value of 512"
+ )
n_ctx = 512
self.gguf_writer.add_context_length(n_ctx)
self.gguf_writer.add_embedding_length(self.hparams["d_model"])
@@ -3611,20 +4945,29 @@ class T5Model(Model):
self.gguf_writer.add_key_length(self.hparams["d_kv"])
self.gguf_writer.add_value_length(self.hparams["d_kv"])
self.gguf_writer.add_layer_norm_eps(self.hparams["layer_norm_epsilon"])
- self.gguf_writer.add_relative_attn_buckets_count(self.hparams["relative_attention_num_buckets"])
- self.gguf_writer.add_layer_norm_rms_eps(self.hparams["layer_norm_epsilon"])
- self.gguf_writer.add_decoder_start_token_id(self.hparams["decoder_start_token_id"])
+ self.gguf_writer.add_relative_attn_buckets_count(
+ self.hparams["relative_attention_num_buckets"])
+ self.gguf_writer.add_layer_norm_rms_eps(
+ self.hparams["layer_norm_epsilon"])
+ self.gguf_writer.add_decoder_start_token_id(
+ self.hparams["decoder_start_token_id"])
self.gguf_writer.add_file_type(self.ftype)
- def modify_tensors(self, data_torch: Tensor, name: str, bid: int | None) -> Iterable[tuple[str, Tensor]]:
+ def modify_tensors(self, data_torch: Tensor, name: str,
+ bid: int | None) -> Iterable[tuple[str, Tensor]]:
del bid # unused
- if name in ["decoder.embed_tokens.weight", "encoder.embed_tokens.weight", "shared.weight"]:
+ if name in [
+ "decoder.embed_tokens.weight", "encoder.embed_tokens.weight",
+ "shared.weight"
+ ]:
if not self.shared_token_embeddings_found:
name = "shared.weight"
self.shared_token_embeddings_found = True
else:
- logger.debug(f"Skipping shared tensor {name!r} in safetensors so that convert can end normally.")
+ logger.debug(
+ f"Skipping shared tensor {name!r} in safetensors so that convert can end normally."
+ )
return []
return [(self.map_tensor_name(name), data_torch)]
@@ -3674,7 +5017,9 @@ class T5EncoderModel(Model):
vocab_size = self.hparams.get('vocab_size', tokenizer.vocab_size())
- tokens: list[bytes] = [f"[PAD{i}]".encode("utf-8") for i in range(vocab_size)]
+ tokens: list[bytes] = [
+ f"[PAD{i}]".encode("utf-8") for i in range(vocab_size)
+ ]
scores: list[float] = [-10000.0] * vocab_size
toktypes: list[int] = [SentencePieceTokenTypes.UNUSED] * vocab_size
@@ -3704,7 +5049,9 @@ class T5EncoderModel(Model):
for key in added_tokens_json:
token_id = added_tokens_json[key]
if token_id >= vocab_size:
- logger.warning(f'ignore token {token_id}: id is out of range, max={vocab_size - 1}')
+ logger.warning(
+ f'ignore token {token_id}: id is out of range, max={vocab_size - 1}'
+ )
continue
tokens[token_id] = key.encode("utf-8")
@@ -3713,7 +5060,9 @@ class T5EncoderModel(Model):
if vocab_size > len(tokens):
pad_count = vocab_size - len(tokens)
- logger.debug(f"Padding vocab with {pad_count} token(s) - [PAD1] through [PAD{pad_count}]")
+ logger.debug(
+ f"Padding vocab with {pad_count} token(s) - [PAD1] through [PAD{pad_count}]"
+ )
for i in range(1, pad_count + 1):
tokens.append(bytes(f"[PAD{i}]", encoding="utf-8"))
scores.append(-1000.0)
@@ -3737,7 +5086,9 @@ class T5EncoderModel(Model):
def set_gguf_parameters(self):
if (n_ctx := self.find_hparam(["n_positions"], optional=True)) is None:
- logger.warning("Couldn't find context length in config.json, assuming default value of 512")
+ logger.warning(
+ "Couldn't find context length in config.json, assuming default value of 512"
+ )
n_ctx = 512
self.gguf_writer.add_context_length(n_ctx)
self.gguf_writer.add_embedding_length(self.hparams["d_model"])
@@ -3747,19 +5098,27 @@ class T5EncoderModel(Model):
self.gguf_writer.add_key_length(self.hparams["d_kv"])
self.gguf_writer.add_value_length(self.hparams["d_kv"])
self.gguf_writer.add_layer_norm_eps(self.hparams["layer_norm_epsilon"])
- self.gguf_writer.add_relative_attn_buckets_count(self.hparams["relative_attention_num_buckets"])
- self.gguf_writer.add_layer_norm_rms_eps(self.hparams["layer_norm_epsilon"])
+ self.gguf_writer.add_relative_attn_buckets_count(
+ self.hparams["relative_attention_num_buckets"])
+ self.gguf_writer.add_layer_norm_rms_eps(
+ self.hparams["layer_norm_epsilon"])
self.gguf_writer.add_file_type(self.ftype)
- def modify_tensors(self, data_torch: Tensor, name: str, bid: int | None) -> Iterable[tuple[str, Tensor]]:
+ def modify_tensors(self, data_torch: Tensor, name: str,
+ bid: int | None) -> Iterable[tuple[str, Tensor]]:
del bid # unused
- if name in ["decoder.embed_tokens.weight", "encoder.embed_tokens.weight", "shared.weight"]:
+ if name in [
+ "decoder.embed_tokens.weight", "encoder.embed_tokens.weight",
+ "shared.weight"
+ ]:
if not self.shared_token_embeddings_found:
name = "shared.weight"
self.shared_token_embeddings_found = True
else:
- logger.debug(f"Skipping shared tensor {name!r} in safetensors so that convert can end normally.")
+ logger.debug(
+ f"Skipping shared tensor {name!r} in safetensors so that convert can end normally."
+ )
return []
return [(self.map_tensor_name(name), data_torch)]
@@ -3789,7 +5148,8 @@ class JaisModel(Model):
self.width_scale = 1.0
if 'mup_output_alpha' in self.hparams:
assert 'mup_width_scale' in self.hparams
- self.width_scale = self.hparams['mup_output_alpha'] * self.hparams['mup_width_scale']
+ self.width_scale = self.hparams['mup_output_alpha'] * self.hparams[
+ 'mup_width_scale']
elif 'width_scale' in self.hparams:
self.width_scale = self.hparams['width_scale']
else:
@@ -3809,7 +5169,8 @@ class JaisModel(Model):
self.gguf_writer.add_layer_norm_eps(self.hparams["layer_norm_epsilon"])
self.gguf_writer.add_file_type(self.ftype)
- def modify_tensors(self, data_torch: Tensor, name: str, bid: int | None) -> Iterable[tuple[str, Tensor]]:
+ def modify_tensors(self, data_torch: Tensor, name: str,
+ bid: int | None) -> Iterable[tuple[str, Tensor]]:
del bid # unused
tensors: list[tuple[str, Tensor]] = []
@@ -3823,13 +5184,16 @@ class JaisModel(Model):
# Some other models has max_alibi_bias spelled out explicitly in the hyperparams,
# but Jais's PyTorch model simply precalculates the slope values and places them
# in relative_pes.slopes
- n_head_closest_log2 = 2 ** math.floor(math.log2(self.hparams["n_head"]))
+ n_head_closest_log2 = 2**math.floor(
+ math.log2(self.hparams["n_head"]))
first_val = float(data_torch[0].item())
- self.max_alibi_bias = -round(math.log2(first_val) * n_head_closest_log2)
+ self.max_alibi_bias = -round(
+ math.log2(first_val) * n_head_closest_log2)
return tensors
- if name.endswith((".c_attn.weight", ".c_proj.weight", ".c_fc.weight", ".c_fc2.weight")):
+ if name.endswith((".c_attn.weight", ".c_proj.weight", ".c_fc.weight",
+ ".c_fc2.weight")):
data_torch = data_torch.transpose(1, 0)
new_name = self.map_tensor_name(name)
@@ -3860,11 +5224,16 @@ class ChatGLMModel(Model):
scores: list[float] = []
from transformers import AutoTokenizer
- tokenizer = AutoTokenizer.from_pretrained(dir_model, trust_remote_code=True)
- vocab_size = hparams.get("padded_vocab_size", len(tokenizer.get_vocab()))
+ tokenizer = AutoTokenizer.from_pretrained(dir_model,
+ trust_remote_code=True)
+ vocab_size = hparams.get("padded_vocab_size",
+ len(tokenizer.get_vocab()))
assert max(tokenizer.get_vocab().values()) < vocab_size
- role_special_tokens = ["<|system|>", "<|user|>", "<|assistant|>", "<|observation|>"]
- special_tokens = ["[MASK]", "[gMASK]", "[sMASK]", "sop", "eop"] + role_special_tokens
+ role_special_tokens = [
+ "<|system|>", "<|user|>", "<|assistant|>", "<|observation|>"
+ ]
+ special_tokens = ["[MASK]", "[gMASK]", "[sMASK]", "sop", "eop"
+ ] + role_special_tokens
for token_id in range(vocab_size):
piece = tokenizer._convert_id_to_token(token_id)
if token_id == 0:
@@ -3876,7 +5245,9 @@ class ChatGLMModel(Model):
text = piece.encode("utf-8")
score = 0.0
- if len(piece) != 0 and token_id < tokenizer.tokenizer.sp_model.vocab_size():
+ if len(
+ piece
+ ) != 0 and token_id < tokenizer.tokenizer.sp_model.vocab_size():
score = tokenizer.tokenizer.sp_model.get_score(token_id)
if token_id >= tokenizer.tokenizer.sp_model.vocab_size():
@@ -3921,10 +5292,13 @@ class ChatGLMModel(Model):
def token_bytes_to_string(b):
from transformers.models.gpt2.tokenization_gpt2 import bytes_to_unicode
byte_encoder = bytes_to_unicode()
- return ''.join([byte_encoder[ord(char)] for char in b.decode('latin-1')])
+ return ''.join(
+ [byte_encoder[ord(char)] for char in b.decode('latin-1')])
@staticmethod
- def bpe(mergeable_ranks: dict[bytes, int], token: bytes, max_rank: int | None = None) -> list[bytes]:
+ def bpe(mergeable_ranks: dict[bytes, int],
+ token: bytes,
+ max_rank: int | None = None) -> list[bytes]:
parts = [bytes([b]) for b in token]
while True:
min_idx = None
@@ -3934,10 +5308,12 @@ class ChatGLMModel(Model):
if rank is not None and (min_rank is None or rank < min_rank):
min_idx = i
min_rank = rank
- if min_rank is None or (max_rank is not None and min_rank >= max_rank):
+ if min_rank is None or (max_rank is not None
+ and min_rank >= max_rank):
break
assert min_idx is not None
- parts = parts[:min_idx] + [parts[min_idx] + parts[min_idx + 1]] + parts[min_idx + 2:]
+ parts = parts[:min_idx] + [parts[min_idx] + parts[min_idx + 1]
+ ] + parts[min_idx + 2:]
return parts
def set_vocab(self):
@@ -3951,7 +5327,8 @@ class ChatGLMModel(Model):
toktypes: list[int] = []
from transformers import AutoTokenizer
- tokenizer = AutoTokenizer.from_pretrained(dir_model, trust_remote_code=True)
+ tokenizer = AutoTokenizer.from_pretrained(dir_model,
+ trust_remote_code=True)
vocab_size = hparams["padded_vocab_size"]
assert max(tokenizer.get_vocab().values()) < vocab_size
@@ -3966,11 +5343,18 @@ class ChatGLMModel(Model):
continue
merged = ChatGLMModel.bpe(mergeable_ranks, token, max_rank=rank)
assert len(merged) >= 2 and len(merged) <= 7
- merges.append(' '.join(map(ChatGLMModel.token_bytes_to_string, merged)))
+ merges.append(' '.join(
+ map(ChatGLMModel.token_bytes_to_string, merged)))
# for this kind of tokenizer, added_vocab is not a subset of vocab, so they need to be combined
added_vocab = tokenizer.get_added_vocab()
- reverse_vocab = {id_ : encoded_tok for encoded_tok, id_ in {**vocab, **added_vocab}.items()}
+ reverse_vocab = {
+ id_: encoded_tok
+ for encoded_tok, id_ in {
+ **vocab,
+ **added_vocab
+ }.items()
+ }
for i in range(vocab_size):
if i not in reverse_vocab:
@@ -3994,23 +5378,33 @@ class ChatGLMModel(Model):
special_vocab = gguf.SpecialVocab(dir_model, load_merges=False)
special_vocab.merges = merges
# only add special tokens when they were not already loaded from config.json
- special_vocab._set_special_token("eos", tokenizer.get_added_vocab()["<|endoftext|>"])
- special_vocab._set_special_token("eot", tokenizer.get_added_vocab()["<|user|>"])
+ special_vocab._set_special_token(
+ "eos",
+ tokenizer.get_added_vocab()["<|endoftext|>"])
+ special_vocab._set_special_token(
+ "eot",
+ tokenizer.get_added_vocab()["<|user|>"])
# this one is usually not in config.json anyway
- special_vocab._set_special_token("unk", tokenizer.get_added_vocab()["<|endoftext|>"])
+ special_vocab._set_special_token(
+ "unk",
+ tokenizer.get_added_vocab()["<|endoftext|>"])
special_vocab.add_to_gguf(self.gguf_writer)
def set_gguf_parameters(self):
n_embed = self.hparams.get("hidden_size", self.hparams.get("n_embed"))
- n_head = self.hparams.get("n_head", self.hparams.get("num_attention_heads"))
+ n_head = self.hparams.get("n_head",
+ self.hparams.get("num_attention_heads"))
n_head_kv = self.hparams.get("multi_query_group_num", n_head)
- self.gguf_writer.add_context_length(self.hparams.get("seq_length", n_embed))
+ self.gguf_writer.add_context_length(
+ self.hparams.get("seq_length", n_embed))
self.gguf_writer.add_embedding_length(n_embed)
- self.gguf_writer.add_feed_forward_length(self.hparams.get("ffn_hidden_size", 4 * n_embed))
+ self.gguf_writer.add_feed_forward_length(
+ self.hparams.get("ffn_hidden_size", 4 * n_embed))
self.gguf_writer.add_block_count(self.hparams["num_layers"])
self.gguf_writer.add_head_count(n_head)
self.gguf_writer.add_head_count_kv(n_head_kv)
- self.gguf_writer.add_layer_norm_rms_eps(self.hparams["layernorm_epsilon"])
+ self.gguf_writer.add_layer_norm_rms_eps(
+ self.hparams["layernorm_epsilon"])
self.gguf_writer.add_file_type(self.ftype)
self.gguf_writer.add_rope_dimension_count(64)
self.gguf_writer.add_add_bos_token(False)
@@ -4019,7 +5413,8 @@ class ChatGLMModel(Model):
rope_freq = rope_freq * self.hparams["rope_ratio"]
self.gguf_writer.add_rope_freq_base(rope_freq)
- def modify_tensors(self, data_torch: Tensor, name: str, bid: int | None) -> Iterable[tuple[str, Tensor]]:
+ def modify_tensors(self, data_torch: Tensor, name: str,
+ bid: int | None) -> Iterable[tuple[str, Tensor]]:
del bid # unused
if name.endswith(".rotary_pos_emb.inv_freq"):
@@ -4043,23 +5438,34 @@ class NemotronModel(Model):
hparams = self.hparams
self.gguf_writer.add_vocab_size(hparams["vocab_size"])
- f_norm_eps = self.find_hparam(["layer_norm_eps", "layer_norm_epsilon", "norm_epsilon", "norm_eps"])
+ f_norm_eps = self.find_hparam([
+ "layer_norm_eps", "layer_norm_epsilon", "norm_epsilon", "norm_eps"
+ ])
self.gguf_writer.add_layer_norm_eps(f_norm_eps)
# * Partial RoPE
- rot_pct = self.find_hparam(["partial_rotary_factor", "rope_pct", "rope_percent"])
+ rot_pct = self.find_hparam(
+ ["partial_rotary_factor", "rope_pct", "rope_percent"])
n_embd = self.find_hparam(["hidden_size", "n_embd"])
n_head = self.find_hparam(["num_attention_heads", "n_head"])
- self.gguf_writer.add_rope_dimension_count(int(rot_pct * n_embd) // n_head)
+ self.gguf_writer.add_rope_dimension_count(
+ int(rot_pct * n_embd) // n_head)
# * RopeScaling for Nemotron
- if "rope_scaling" not in self.hparams or self.hparams["rope_scaling"] is None:
+ if "rope_scaling" not in self.hparams or self.hparams[
+ "rope_scaling"] is None:
self.gguf_writer.add_rope_scaling_type(gguf.RopeScalingType.NONE)
else:
self.gguf_writer.add_rope_scaling_type(gguf.RopeScalingType.LINEAR)
self.gguf_writer.add_rope_scaling_factor(self.hparams["factor"])
- def modify_tensors(self, data_torch: Tensor, name: str, bid: int | None) -> Iterable[tuple[str, Tensor]]:
+ def modify_tensors(self, data_torch: Tensor, name: str,
+ bid: int | None) -> Iterable[tuple[str, Tensor]]:
+ # * Adding +1 to LayerNorm's weights here to implement layernorm1p w/o
+ # changing anything on the GGML engine side
+ # model.layers.{l}.input_layernorm.weight
+ # model.layers.{l}.post_attention_layernorm.weight
+ # model.norm.weight
if name.endswith("norm.weight"):
data_torch = data_torch + 1
@@ -4080,7 +5486,8 @@ class ExaoneModel(Model):
num_heads = hparams["num_attention_heads"]
num_kv_heads = hparams.get("num_key_value_heads", num_heads)
layer_norm_eps = hparams["layer_norm_epsilon"]
- intermediate_size = hparams["intermediate_size"] if "intermediate_size" in hparams else 4 * embed_dim
+ intermediate_size = hparams[
+ "intermediate_size"] if "intermediate_size" in hparams else 4 * embed_dim
num_layers = hparams["num_layers"]
# ignore for now as EXAONE-3.0-7.8B-Instruct attentino_dropout is 0.0
# attention_dropout_rate = hparams["attention_dropout"]
@@ -4097,27 +5504,35 @@ class ExaoneModel(Model):
if (rope_theta := self.hparams.get("rope_theta")) is not None:
self.gguf_writer.add_rope_freq_base(rope_theta)
- rotary_factor = self.find_hparam(["partial_rotary_factor", "rope_pct"], optional=True)
+ rotary_factor = self.find_hparam(["partial_rotary_factor", "rope_pct"],
+ optional=True)
rotary_factor = rotary_factor if rotary_factor is not None else 1.0
self.gguf_writer.add_rope_dimension_count(
int(rotary_factor *
(hparams["hidden_size"] // hparams["num_attention_heads"])))
- if hparams.get("rope_scaling") is not None and "factor" in hparams["rope_scaling"]:
+ if hparams.get("rope_scaling"
+ ) is not None and "factor" in hparams["rope_scaling"]:
if hparams["rope_scaling"].get("type") == "linear":
- self.gguf_writer.add_rope_scaling_type(gguf.RopeScalingType.LINEAR)
- self.gguf_writer.add_rope_scaling_factor(hparams["rope_scaling"]["factor"])
+ self.gguf_writer.add_rope_scaling_type(
+ gguf.RopeScalingType.LINEAR)
+ self.gguf_writer.add_rope_scaling_factor(
+ hparams["rope_scaling"]["factor"])
def generate_extra_tensors(self) -> Iterable[tuple[str, Tensor]]:
if rope_scaling := self.find_hparam(["rope_scaling"], optional=True):
if rope_scaling.get("rope_type", '').lower() == "llama3":
base = self.hparams.get("rope_theta", 10000.0)
- dim = self.hparams.get("head_dim", self.hparams["hidden_size"] // self.hparams["num_attention_heads"])
- freqs = 1.0 / (base ** (torch.arange(0, dim, 2, dtype=torch.float32) / dim))
+ dim = self.hparams.get(
+ "head_dim", self.hparams["hidden_size"] //
+ self.hparams["num_attention_heads"])
+ freqs = 1.0 / (base**(
+ torch.arange(0, dim, 2, dtype=torch.float32) / dim))
factor = rope_scaling.get("factor", 8.0)
low_freq_factor = rope_scaling.get("low_freq_factor", 1.0)
high_freq_factor = rope_scaling.get("high_freq_factor", 4.0)
- old_context_len = self.hparams.get("original_max_position_embeddings", 8192)
+ old_context_len = self.hparams.get(
+ "original_max_position_embeddings", 8192)
low_freq_wavelen = old_context_len / low_freq_factor
high_freq_wavelen = old_context_len / high_freq_factor
@@ -4131,8 +5546,10 @@ class ExaoneModel(Model):
elif wavelen > low_freq_wavelen:
rope_factors.append(factor)
else:
- smooth = (old_context_len / wavelen - low_freq_factor) / (high_freq_factor - low_freq_factor)
- rope_factors.append(1 / ((1 - smooth) / factor + smooth))
+ smooth = (old_context_len / wavelen - low_freq_factor
+ ) / (high_freq_factor - low_freq_factor)
+ rope_factors.append(1 /
+ ((1 - smooth) / factor + smooth))
yield (self.format_tensor_name(gguf.MODEL_TENSOR.ROPE_FREQS),
torch.tensor(rope_factors, dtype=torch.float32))
@@ -4154,16 +5571,19 @@ class GraniteModel(LlamaModel):
- logits_scaling
"""
if head_dim := self.hparams.pop("head_dim", None):
- logger.warning("Ignoring head_dim (%s) from config for Granite", head_dim)
+ logger.warning("Ignoring head_dim (%s) from config for Granite",
+ head_dim)
super().set_gguf_parameters()
# NOTE: Convert _multiplier params to _scale params for naming
# consistency
if attention_scale := self.hparams.get("attention_multiplier"):
self.gguf_writer.add_attention_scale(attention_scale)
- logger.info("gguf: (granite) attention_scale = %s", attention_scale)
+ logger.info("gguf: (granite) attention_scale = %s",
+ attention_scale)
if embedding_scale := self.hparams.get("embedding_multiplier"):
self.gguf_writer.add_embedding_scale(embedding_scale)
- logger.info("gguf: (granite) embedding_scale = %s", embedding_scale)
+ logger.info("gguf: (granite) embedding_scale = %s",
+ embedding_scale)
if residual_scale := self.hparams.get("residual_multiplier"):
self.gguf_writer.add_residual_scale(residual_scale)
logger.info("gguf: (granite) residual_scale = %s", residual_scale)
@@ -4177,7 +5597,8 @@ class GraniteMoeModel(GraniteModel):
"""Conversion for IBM's GraniteMoeForCausalLM"""
model_arch = gguf.MODEL_ARCH.GRANITE_MOE
- def modify_tensors(self, data_torch: Tensor, name: str, bid: int | None) -> Iterable[tuple[str, Tensor]]:
+ def modify_tensors(self, data_torch: Tensor, name: str,
+ bid: int | None) -> Iterable[tuple[str, Tensor]]:
"""In modeling_granitemoe, the JetMoe implementation of parallel experts
is used. This essentially merges w1 and w3 into a single tensor with 2x
the hidden size that is then split during forward. To keep compatibility
@@ -4186,11 +5607,15 @@ class GraniteMoeModel(GraniteModel):
if name.endswith("block_sparse_moe.input_linear.weight"):
ffn_dim = self.hparams["intermediate_size"]
- assert data_torch.shape[-2] == 2 * ffn_dim, "Merged FFN tensor size must be 2 * intermediate_size"
- gate, up = data_torch[..., :ffn_dim, :], data_torch[..., ffn_dim:, :]
+ assert data_torch.shape[
+ -2] == 2 * ffn_dim, "Merged FFN tensor size must be 2 * intermediate_size"
+ gate, up = data_torch[..., :ffn_dim, :], data_torch[...,
+ ffn_dim:, :]
return [
- (self.format_tensor_name(gguf.MODEL_TENSOR.FFN_GATE_EXP, bid), gate),
- (self.format_tensor_name(gguf.MODEL_TENSOR.FFN_UP_EXP, bid), up),
+ (self.format_tensor_name(gguf.MODEL_TENSOR.FFN_GATE_EXP,
+ bid), gate),
+ (self.format_tensor_name(gguf.MODEL_TENSOR.FFN_UP_EXP,
+ bid), up),
]
return super().modify_tensors(data_torch, name, bid)
@@ -4208,7 +5633,8 @@ class ChameleonModel(Model):
def set_vocab(self):
self._set_vocab_gpt2()
- def modify_tensors(self, data_torch: Tensor, name: str, bid: int | None) -> Iterable[tuple[str, Tensor]]:
+ def modify_tensors(self, data_torch: Tensor, name: str,
+ bid: int | None) -> Iterable[tuple[str, Tensor]]:
# ignore image tokenizer for now
# TODO: remove this once image support is implemented for Chameleon
if name.startswith("model.vqmodel"):
@@ -4223,9 +5649,11 @@ class ChameleonModel(Model):
if name.endswith(("k_proj.weight", "k_proj.bias")):
data_torch = LlamaModel.permute(data_torch, n_head, n_kv_head)
if name.endswith(("q_norm.weight", "q_norm.bias")):
- data_torch = ChameleonModel._reverse_hf_permute(data_torch, n_head, hidden_dim)
+ data_torch = ChameleonModel._reverse_hf_permute(
+ data_torch, n_head, hidden_dim)
if name.endswith(("k_norm.weight", "k_norm.bias")):
- data_torch = ChameleonModel._reverse_hf_permute(data_torch, n_kv_head, hidden_dim)
+ data_torch = ChameleonModel._reverse_hf_permute(
+ data_torch, n_kv_head, hidden_dim)
return [(self.map_tensor_name(name), data_torch)]
@@ -4253,7 +5681,6 @@ class LazyTorchTensor(gguf.LazyBase):
torch.float32: np.float32,
}
- # used for safetensors slices
_dtype_str_map: dict[str, torch.dtype] = {
"F64": torch.float64,
"F32": torch.float32,
@@ -4275,20 +5702,23 @@ class LazyTorchTensor(gguf.LazyBase):
def numpy(self) -> gguf.LazyNumpyTensor:
dtype = self._dtype_map[self.dtype]
return gguf.LazyNumpyTensor(
- meta=gguf.LazyNumpyTensor.meta_with_dtype_and_shape(dtype, self.shape),
- args=(self,),
- func=(lambda s: s.numpy())
- )
+ meta=gguf.LazyNumpyTensor.meta_with_dtype_and_shape(
+ dtype, self.shape),
+ args=(self, ),
+ func=(lambda s: s.numpy()))
@classmethod
- def meta_with_dtype_and_shape(cls, dtype: torch.dtype, shape: tuple[int, ...]) -> Tensor:
+ def meta_with_dtype_and_shape(cls, dtype: torch.dtype,
+ shape: tuple[int, ...]) -> Tensor:
return torch.empty(size=shape, dtype=dtype, device="meta")
@classmethod
def from_safetensors_slice(cls, st_slice: Any) -> Tensor:
dtype = cls._dtype_str_map[st_slice.get_dtype()]
shape: tuple[int, ...] = tuple(st_slice.get_shape())
- lazy = cls(meta=cls.meta_with_dtype_and_shape(dtype, shape), args=(st_slice,), func=lambda s: s[:])
+ lazy = cls(meta=cls.meta_with_dtype_and_shape(dtype, shape),
+ args=(st_slice, ),
+ func=lambda s: s[:])
return cast(torch.Tensor, lazy)
@classmethod
@@ -4308,11 +5738,13 @@ def parse_args() -> argparse.Namespace:
parser = argparse.ArgumentParser(
description="Convert a huggingface model to a GGML compatible file")
parser.add_argument(
- "--vocab-only", action="store_true",
+ "--vocab-only",
+ action="store_true",
help="extract only the vocab",
)
parser.add_argument(
- "--outfile", type=Path,
+ "--outfile",
+ type=Path,
help="path to write to; default: based on input. {ftype} will be replaced by the outtype.",
)
parser.add_argument(
@@ -4320,57 +5752,74 @@ def parse_args() -> argparse.Namespace:
type=str,
choices=["f32", "f16", "bf16", "q8_0", "tq1_0", "tq2_0", "auto"],
default="f16",
- help=
- "output format - use f32 for float32, f16 for float16, bf16 for bfloat16, q8_0 for Q8_0,"
- " tq1_0 or tq2_0 for ternary, and auto for the highest-fidelity 16-bit float type depending"
- " on the first loaded tensor type",
- )
+ help="output format - use f32 for float32, f16 for float16, bf16 for bfloat16, q8_0 for Q8_0,"
+ "tq1_0 or tq2_0 for ternary, and auto for the highest-fidelity 16-bit float type depending "
+ "on the first loaded tensor type"),
parser.add_argument(
- "--bigendian", action="store_true",
+ "--bigendian",
+ action="store_true",
help="model is executed on big endian machine",
)
parser.add_argument(
- "model", type=Path,
+ "model",
+ type=Path,
help="directory containing model file",
+ nargs="?",
)
parser.add_argument(
- "--use-temp-file", action="store_true",
+ "--use-temp-file",
+ action="store_true",
help="use the tempfile library while processing (helpful when running out of memory, process killed)",
)
parser.add_argument(
- "--no-lazy", action="store_true",
+ "--no-lazy",
+ action="store_true",
help="use more RAM by computing all outputs before writing (use in case lazy evaluation is broken)",
)
parser.add_argument(
- "--model-name", type=str, default=None,
+ "--model-name",
+ type=str,
+ default=None,
help="name of the model",
)
parser.add_argument(
- "--verbose", action="store_true",
+ "--verbose",
+ action="store_true",
help="increase output verbosity",
)
parser.add_argument(
- "--split-max-tensors", type=int, default=0,
+ "--split-max-tensors",
+ type=int,
+ default=0,
help="max tensors in each split",
)
parser.add_argument(
- "--split-max-size", type=str, default="0",
+ "--split-max-size",
+ type=str,
+ default="0",
help="max size per split N(M|G)",
)
parser.add_argument(
- "--dry-run", action="store_true",
+ "--dry-run",
+ action="store_true",
help="only print out a split plan and exit, without writing any new files",
)
parser.add_argument(
- "--no-tensor-first-split", action="store_true",
- help="do not add tensors to the first split (disabled by default)"
- )
+ "--no-tensor-first-split",
+ action="store_true",
+ help="do not add tensors to the first split (disabled by default)")
parser.add_argument(
- "--metadata", type=Path,
- help="Specify the path for an authorship metadata override file"
- )
+ "--metadata",
+ type=Path,
+ help="Specify the path for an authorship metadata override file")
+ parser.add_argument("--print-supported-models",
+ action="store_true",
+ help="Print the supported models")
- return parser.parse_args()
+ args = parser.parse_args()
+ if not args.print_supported_models and args.model is None:
+ parser.error("the following arguments are required: model")
+ return args
def split_str_to_n_bytes(split_str: str) -> int:
@@ -4383,7 +5832,9 @@ def split_str_to_n_bytes(split_str: str) -> int:
elif split_str.isnumeric():
n = int(split_str)
else:
- raise ValueError(f"Invalid split size: {split_str}, must be a number, optionally followed by K, M, or G")
+ raise ValueError(
+ f"Invalid split size: {split_str}, must be a number, optionally followed by K, M, or G"
+ )
if n < 0:
raise ValueError(f"Invalid split size: {split_str}, must be positive")
@@ -4394,6 +5845,11 @@ def split_str_to_n_bytes(split_str: str) -> int:
def main() -> None:
args = parse_args()
+ if args.print_supported_models:
+ logger.error("Supported models:")
+ Model.print_registered_models()
+ sys.exit(0)
+
if args.verbose:
logging.basicConfig(level=logging.DEBUG)
else:
@@ -4439,21 +5895,26 @@ def main() -> None:
logger.error(f"Model {model_architecture} is not supported")
sys.exit(1)
- from transformers import AutoModelForCausalLM
- model = AutoModelForCausalLM.from_pretrained(dir_model)
- model_instance = model_class(model, dir_model=dir_model, ftype=output_type, fname_out=fname_out,
- is_big_endian=args.bigendian, use_temp_file=args.use_temp_file,
- eager=args.no_lazy,
- metadata_override=args.metadata, model_name=args.model_name,
- split_max_tensors=args.split_max_tensors,
- split_max_size=split_str_to_n_bytes(args.split_max_size),
- dry_run=args.dry_run,
- small_first_shard=args.no_tensor_first_split)
+ model_instance = model_class(
+ dir_model=dir_model,
+ ftype=output_type,
+ fname_out=fname_out,
+ is_big_endian=args.bigendian,
+ use_temp_file=args.use_temp_file,
+ eager=args.no_lazy,
+ metadata_override=args.metadata,
+ model_name=args.model_name,
+ split_max_tensors=args.split_max_tensors,
+ split_max_size=split_str_to_n_bytes(args.split_max_size),
+ dry_run=args.dry_run,
+ small_first_shard=args.no_tensor_first_split)
if args.vocab_only:
logger.info("Exporting model vocab...")
model_instance.write_vocab()
- logger.info(f"Model vocab successfully exported to {model_instance.fname_out}")
+ logger.info(
+ f"Model vocab successfully exported to {model_instance.fname_out}"
+ )
else:
logger.info("Exporting model...")
model_instance.write()