intel/auto-round

Pull Requests Commits

[pre-commit.ci] auto fixes from pre-commit.com hooks

pre-commit-ci[bot] committed 26 days ago

6121924b

Merge branch 'main' into lvl/cpu_ram_optimization

lvliang-intel committed 26 days ago

Verified 483809f9

Extensible Module Weight Type Conversion Framework. (#1375)

xin3he committed 26 days ago

d4f74c72

Fix FP8 MLA (#1343)

yiliu30 committed 26 days ago

857c7ce8

feat: handle per-tensor FP8 dequantization for Devstral models (#1356)

SwekeR-463 committed 26 days ago

a37363ca

fix copy python file issue (#1372)

wenhuach21 committed 26 days ago

8c41dcbf

add support for w4a16_mixed (#1326)

n1ck-guo committed 26 days ago

7f97ab82

fix cuda ut fail for transformers v5.0 (#1357)

n1ck-guo committed 26 days ago

3118fee5

support qwen3-next/glm4-moe/deepseekv3 monkey patch for transformers 5 (#1367)

wenhuach21 committed 26 days ago

f2685073

support qwen3-moe/glm-4.7-flash monkey patch for transformers 5 (#1354)

wenhuach21 committed 26 days ago

6a3a9003

fix bug of exporting fp8 static (#1361)

n1ck-guo committed 26 days ago

89177a15

Refactor FP8 dequantization and detection using registry pattern (#1348)

scopophobic committed 26 days ago

a9d32cda

refactor auto_round api, using local symbol table to pass all args (#1358)

n1ck-guo committed 26 days ago

a55b0427

Autoround in vLLM Office Hours (#1322)

yiliu30 committed 26 days ago

63e5a10e

Fix cpu ut for transformers v5 (#1333)

Kaihui-intel committed 26 days ago

0d557d6a

Fix CPU CI (#1344)

XuehaoSun committed 26 days ago

048261cb

add copilot-instructions (#1341)

xin3he committed 26 days ago

67b5537e

refactor eval and add UT (#1324)

xin3he committed 26 days ago

409ceef8

Add Chinese README (#1335)

ZaneMark committed 26 days ago

dd86e355

Fix issue where predefined ignore layers incorrectly apply to GGUF (#1340)

wenhuach21 committed 26 days ago

331b5eb5

Fix attribute bug for diffusion model (#1338)

mengniwang95 committed 26 days ago

7aa787fc

update release package and bump version (#1327)

chensuyue committed 26 days ago

f92b0a59

format add check for scheme args (#1332)

n1ck-guo committed 26 days ago

ccf9496d

Delay materializing the replaced model weights until quantization (#1307)

yiliu30 committed 26 days ago

56518473

(feat): add support for g2 fp8 on cpu with LUT (#1254)

SwekeR-463 committed 26 days ago

92681af3

fix tied weights keys bug (#1330)

wenhuach21 committed 26 days ago

2f4427d6

Update publication dates in README and add recipe configuration detai… (#1329)

xin3he committed 26 days ago

6e89d03e

[GGUF] using quant_nontext_module to control whether quant vision model (#1317)

n1ck-guo committed 26 days ago

13ecc941

simplify pull_request_template.md (#1328)

xin3he committed 26 days ago

cc0065aa

add blockwise way in pure rtn to ease module replacements (#1320)

wenhuach21 committed 26 days ago

cf6b178b

Older