Support loading Quark quantized models in Transformers #36372
SunMarc
merged
22 commits into
huggingface:main
from
fxmarty-amd:quark-quantizer-upstream
add quark quantizer
1f87b7dd
add quark doc
c405adb4
clean up doc
eb189de5
fix tests
36d18cfe
make style
8d233b46
more style fixes
5f24cee8
cleanup imports
d275c873
cleaning
f5e18172
fxmarty-amd
commented
on 2025-02-24
precise install
70e30fae
Merge branch 'main' into quark-quantizer-upstream
05efcb0a
Merge branch 'quark-quantizer-upstream' of https://github.com/fxmarty…
ea2b62ea
Update docs/source/en/quantization/quark.md
c2e5ba0c
Update tests/quantization/quark_integration/test_quark.py
9ee20b1f
Update src/transformers/utils/quantization_config.py
9b0c135e
remove import guard as suggested
a1b2c8b1
update copyright headers
93d84803
add quark to transformers-quantization-latest-gpu Dockerfile
2be83a1b
make tests pass on transformers main + quark==0.7
3f76848c
fxmarty-amd
commented
on 2025-03-10
fxmarty-amd
commented
on 2025-03-10
add missing F8_E4M3 and F8_E5M2 keys from str_to_torch_dtype
fda836f9
fxmarty-amd
commented
on 2025-03-10
Merge remote-tracking branch 'origin/main' into quark-quantizer-upstream
d8ca5e56
Merge remote-tracking branch 'origin/main' into quark-quantizer-upstream
7da2a57e
Merge branch 'main' into quark-quantizer-upstream
f6dbb795
SunMarc
approved these changes
on 2025-03-20
Login to write a write a comment.
Login via GitHub