intel/auto-round

Pull Requests Commits

feat: enhance max_scale handling in qdq method and add unit tests for scale preservation

xin3he committed 30 days ago

d9854f39

fix model_free saving with single shard and MXFP fake format evaluation (#2391)

xin3he committed 30 days ago

Verified f4d2c162

fix: fake format has no FP32 tensor restoration (#2392)

xin3he committed 30 days ago

Verified 8447db40

fix: update import to include get_model_path for model retrieval (#2388)

xin3he committed 31 days ago

Verified c29847ef

fix:reduce memory usage when loading native fused MoE checkpoints (#2382)

n1ck-guo committed 31 days ago

Verified a5f6ddd8

fix: restore FP32 tensors saved as FP16/BF16 during tensor copying (#2375)

xin3he committed 33 days ago

Verified a6c612fa

fix: share NVFP4 global scales for Wan QKV projections (#2379)

changwangss committed 33 days ago

Verified 36079c49

[ARK] Improve XPU W4A16 WOQ decode with dense S4 DPAS path (#2331)

Zhenzhong1 committed 33 days ago

Verified eaf57a2a

patch triangular_solve for Qwen3.8 model for hpu (#2377)

lkk12014402 committed 33 days ago

Verified 8cd9b878

Remove deprecated files (#2373)

chensuyue committed 34 days ago

Verified fd077878

bug fix: fix NVFP4 evaluation with input_ global_scale handling and testing (#2374)

xin3he committed 34 days ago

Verified 27daf7f8

remove bagel specific constraints (#2372)

WeiweiZhang1 committed 34 days ago

Verified 4b3680b9

add fineweb-edu dataset as calibration bakeup (#2345)

WeiweiZhang1 committed 34 days ago

Verified 3938e152

feat: add NVFP4 static KV cache quantization support (#2366)

yiliu30 committed 34 days ago

Verified 04fca622

docs: clarify diffusion export format (#2367)

changwangss committed 34 days ago

Verified ea838e56

Add blockwise rotation. (#2087)

lkk12014402 committed 35 days ago

Verified 9b746ff9

fix: recognize Modular Diffusers pipelines as diffusion models (#2120)

Ar4ikov committed 35 days ago

Verified 0d76712e

refactor: decouple datatype quantization from algorithms (#2360)

n1ck-guo committed 35 days ago

Verified 2e82448a

fix: save the config of diffusers models on export (#2119)

Ar4ikov committed 35 days ago

Verified 1983b65a

feat(ark): fused Hadamard + MXFP4 activation quantization kernel

jijiaz committed 35 days ago

Verified ecc5b005

perf: avoid unused SVD reconstruction and select the CUDA driver in SVDQuant (#2356)

changwangss committed 35 days ago

Verified 903ba0ce

unify target_bits/options into bits/schemes for AutoScheme (#2333)

n1ck-guo committed 35 days ago

Verified 2b5a3c55

Refactor test timeouts and update dependencies for performance (#2361)

XuehaoSun committed 35 days ago

Verified 1965bdf0

fix: preserve declared FP32 tensors during model dtype conversion (#2359)

changwangss committed 36 days ago

Verified b33eedc9

fix: support list inputs in diffusion tuning cache (#2346)

changwangss committed 36 days ago

Verified c53611fb

fix: update key cache access for compatibility with newer DynamicCache implementation (#2348)

chensuyue committed 37 days ago

Verified f331bb02

fix: reload Diffusers blocks from component checkpoints (#2353)

changwangss committed 37 days ago

Verified 6a8d95f4

fix qwen35 regression (#2352)

wenhuach21 committed 39 days ago

Verified 141e4c99

feat: implement resume functionality for model-free compression (#2330)

xin3he committed 40 days ago

Verified 9e2650f4

Update README.md (#2343)

wenhuach21 committed 40 days ago

Verified 51003909

Older