Skip to content

SOTA Open Source TTS

[fix]:fix future warning: sdp_kernel

Ardent Illuminab3c39d320545
diff --git a/fish_speech/models/text2semantic/inference.py b/fish_speech/models/text2semantic/inference.py
index acf31ee..aeb3973 100644
--- a/fish_speech/models/text2semantic/inference.py
+++ b/fish_speech/models/text2semantic/inference.py
@@ -13,6 +13,7 @@ import torch
import torch._dynamo.config
import torch._inductor.config
from loguru import logger
+from torch.nn.attention import SDPBackend, sdpa_kernel
from tqdm import tqdm
from transformers import AutoTokenizer
@@ -23,7 +24,12 @@ from fish_speech.conversation import (
TextPart,
VQPart,
)
-from fish_speech.models.text2semantic.llama import BaseModelArgs
+from fish_speech.models.text2semantic.llama import (
+ BaseModelArgs,
+ BaseTransformer,
+ DualARTransformer,
+ NaiveTransformer,
+)
from fish_speech.text import clean_text, split_text
from fish_speech.tokenizer import IM_END_TOKEN, FishTokenizer
@@ -36,15 +42,6 @@ if hasattr(torch._inductor.config, "fx_graph_cache"):
torch._inductor.config.fx_graph_cache = True
-from torch.nn.attention import SDPBackend, sdpa_kernel
-
-from fish_speech.models.text2semantic.llama import (
- BaseTransformer,
- DualARTransformer,
- NaiveTransformer,
-)
-
-
def multinomial_sample_one_no_sync(
probs_sort,
): # Does multinomial sampling without a cuda synchronization
@@ -372,8 +369,12 @@ def decode_n_tokens(
window = previous_tokens[:, i - win_size : i]
with (
- torch.backends.cuda.sdp_kernel(
- enable_flash=False, enable_mem_efficient=False, enable_math=True
+ sdpa_kernel(
+ [
+ SDPBackend.FLASH_ATTENTION,
+ SDPBackend.EFFICIENT_ATTENTION,
+ SDPBackend.MATH,
+ ]
)
if torch.cuda.is_available()
else nullcontext()