intel/auto-round

Pull Requests Commits

Add envs.py for environment checks and update imports in test files

xin3he committed 15 hours ago

1d2fcca9

Update BackendInfos for AutoGPTQ based on transformer version and save g_idx for gptqmodel backend (#1429)

xin3he committed 1 day ago

Verified a1876107

Fix cuda model ut [glm4][Molmo] (#1428)

Kaihui-intel committed 1 day ago

Verified 16a18d7c

remove require_intel_extension_for_pytorch and fix TypeError: unhashable type: 'set' (#1425)

xin3he committed 1 day ago

Verified 319da040

fix bug in PR-1244 (#1422)

xin3he committed 1 day ago

Verified 81383e92

Refactor evaluation in tests to use evaluate_accuracy function (#1402)

xin3he committed 2 days ago

Verified 70a6bcc2

fix: update cleanup logic for quantized model path in tests (#1419)

xin3he committed 2 days ago

Verified ebb534a2

fix test_evaluation and AutoTuner new issue (#1417)

xin3he committed 2 days ago

Verified 520214a8

Update torch to 2.9.1 in CI (#1313)

XuehaoSun committed 2 days ago

Verified 0cac8490

adapt vllm_ext to new extra config (#1409)

mengniwang95 committed 2 days ago

Verified d2b1fed1

update auto-round-kernel package name to auto-round-lib (#1410)

chensuyue committed 5 days ago

Verified bab09935

[fix]AttributeError: 'Autotuner' object has no attribute '_cache_lock' (#1414)

xin3he committed 5 days ago

Verified 6093abc2

refine qwen3_vl_moe experts forward (#1413)

WeiweiZhang1 committed 5 days ago

Verified 2824c37b

fix Qwen3-VL model auto_awq export, add auto_awq vllm ut (#1384)

WeiweiZhang1 committed 5 days ago

Verified 40bd6410

better warning for not support mllm while quant_non_text_module (#1411)

n1ck-guo committed 5 days ago

Verified 96cfbc95

[LLMC] Wrap block for DDP (#1407)

yiliu30 committed 6 days ago

Verified 99204417

fix: update model imports from AutoModelForVision2Seq to AutoModelFor… (#1412)

xin3he committed 6 days ago

Verified b2bf11df

Enable xpu test for CI (#1352)

chensuyue committed 6 days ago

Verified 4bcebe7a

fix cuda ut fail: gptqmodel and gguf format with fp8 input model (#1399)

n1ck-guo committed 6 days ago

Verified 4dc04f4b

support llm dynamic wint8aint8 export (#1376)

mengniwang95 committed 7 days ago

Verified c9f6155b

align scale for MLA kv (#1395)

yiliu30 committed 7 days ago

Verified a267625e

Support quant for `meituan-longcat/LongCat-Flash-Lite` (#1388)

Kaihui-intel committed 7 days ago

Verified a12e3f5a

Fix version (#1393)

XuehaoSun committed 7 days ago

Verified 68d07bfa

Implement MXFP4Handler MXFP8Handler NVFP4Handler (#1385)

xin3he committed 7 days ago

Verified f7db867a

fix layer config (#1373)

wenhuach21 committed 7 days ago

Verified 4064c4f1

[Regression] Detach scale tensor to prevent holding computation graph references (#1389)

xin3he committed 7 days ago

Verified 1c758f45

fix cuda ut fail (#1370)

n1ck-guo committed 7 days ago

Verified 6cdef7a2

Support general MOE replacement for MOE models (Transformers 5.0 compatible) (#1374)

lvliang-intel committed 8 days ago

Verified 28839a9a

Disable replace `FP8Expert` (#1379)

yiliu30 committed 8 days ago

Verified 137da551

fix --device_map cuda and xpu issue (#1383)

wenhuach21 committed 8 days ago

Verified f132c0c4

Older