Local UI to run and train LLMs and diffusion models. Supports GGUF, MLX, Qwen3.8, DeepSeek-V4.1, MiniMax-H3, Gemma 4, FLUX and more.
Python agentaichatgptdeepseekfine-tuninggemmaimage-generationllamallmllmsopenaipythonqwenreinforcement-learningself-hostedstable-diffusiontext-to-speechttsuiunsloth
"""CPU-only checks of the compressed-tensors registry and save_method normalization; no model load."""
from __future__ import annotations
import pytest
from unsloth.save import COMPRESSED_EXPORT_SCHEMES, _normalize_compressed_method
def test_registry_entries_are_well_formed():
assert COMPRESSED_EXPORT_SCHEMES, "compressed export registry must not be empty"
for alias, value in COMPRESSED_EXPORT_SCHEMES.items():
assert (
isinstance(alias, str) and alias == alias.lower()
), f"alias must be a lowercase str: {alias!r}"
assert (
isinstance(value, tuple) and len(value) == 3
), f"{alias!r} must map to a (scheme, needs_calib, suffix) tuple"
scheme, needs_calib, suffix = value
assert isinstance(scheme, str) and scheme, f"{alias!r}: scheme must be a non-empty str"
assert isinstance(needs_calib, bool), f"{alias!r}: needs_calibration must be a bool"
assert isinstance(suffix, str) and suffix, f"{alias!r}: suffix must be a non-empty str"
# The suffix builds the sibling dir "<save_dir>-<suffix>", so it must be path-safe.
assert not (
set(suffix) & set("/\\ ")
), f"{alias!r}: suffix {suffix!r} must be filesystem-safe"
def test_every_alias_round_trips_case_and_separator_insensitive():
for alias, value in COMPRESSED_EXPORT_SCHEMES.items():
assert _normalize_compressed_method(alias) == value
assert _normalize_compressed_method(alias.upper()) == value
assert _normalize_compressed_method(f" {alias.replace('_', '-')} ") == value
@pytest.mark.parametrize(
"method", ["merged_16bit", "16bit", "merged_4bit", "lora", "", None, 123, ["fp8"]]
)
def test_standard_save_methods_are_not_treated_as_compressed(method):
assert _normalize_compressed_method(method) is None
@pytest.mark.parametrize(
"method", ["fp8_turbo", "nvfp4_xl", "w4a99", "mxfp3", "int8_banana", "fp4_max"]
)
def test_near_miss_compressed_names_raise(method):
# Unsupported compressed names must fail loudly, not fall through to "unknown save_method".
with pytest.raises(RuntimeError):
_normalize_compressed_method(method)
def test_calibration_flags_match_known_schemes():
# Only static FP8 and NVFP4 need calibration data.
assert _normalize_compressed_method("fp8")[1] is False
assert _normalize_compressed_method("fp8_static")[1] is True
assert _normalize_compressed_method("nvfp4")[1] is True
assert _normalize_compressed_method("mxfp4")[1] is False
def test_core_aliases_present():
for alias in ("fp8", "fp8_dynamic", "fp8_static", "mxfp4", "nvfp4", "int8", "w4a16", "w8a8"):
assert alias in COMPRESSED_EXPORT_SCHEMES, f"expected core alias {alias!r} in registry"