intel/auto-round

Pull Requests Commits

init performance test

XuehaoSun committed 3 days ago

Verified 020731d3

Fix sglang ut (#1541)

mengniwang95 committed 3 days ago

Verified 53f2b0e0

support MTP module WOQ quantization and step3p5 model quantization (#1526)

xin3he committed 4 days ago

Verified 42fdb069

fix missing trust_remote_code in unfused_moe (#1532)

xin3he committed 5 days ago

Verified 837888dd

fix dynamic int8 w8a8 export issue with tuning (#1525)

thuang6 committed 5 days ago

Verified a42fd85d

fix broken link (#1530)

xin3he committed 5 days ago

Verified f8e8ae9e

Optimize CPU RAM peak memory during quantization (#1386)

lvliang-intel committed 6 days ago

Verified 0d372151

Support diffusion model saving (#1519)

mengniwang95 committed 6 days ago

Verified a0ad7780

support gpt-oss mxfp4 directly loading (#1401)

xin3he committed 6 days ago

Verified e14e7db6

enable --eval for diffusion model (#1522)

xin3he committed 6 days ago

Verified eec1bed6

Fix transformers 5.2.0 transpose conversion issue (#1475)

lvliang-intel committed 6 days ago

Verified 0127590c

Refactor build process to use 'uv build' instead of 'python setup.py' (#1495)

XuehaoSun committed 6 days ago

Verified 50246114

Fix Inference tensors do not track version counter error (#1502)

mengniwang95 committed 6 days ago

Verified e2b88a5c

Enable more xpu instance for xpu CI (#1521)

chensuyue committed 6 days ago

Verified 30e4043f

fix bug of quantizing Z-image (#1516)

xin3he committed 6 days ago

Verified 8a404ec0

Fix CUDA UT and HPU UT (#1514)

XuehaoSun committed 6 days ago

Verified 89f99ae0

release CUDA memory in WeightConverter and avoid meaningless print (#1498)

xin3he committed 6 days ago

Verified cd10eab0

fix low_gpu default value, refine doc (#1520)

WeiweiZhang1 committed 7 days ago

Verified 279637de

[BUG] update moe check logic and make .gate ignore general (#1517)

xin3he committed 7 days ago

Verified 2eaa3b14

reduce ram&vram usage for vlm calib stage (#1488)

WeiweiZhang1 committed 7 days ago

Verified be387131

support minimax_m2 ignore layer: block_sparse_moe.gate (#1508)

xin3he committed 7 days ago

Verified a32224d3

Fix FP8 quantizer for Transformers v4 (#1504)

yiliu30 committed 7 days ago

Verified 6ab2db2b

Update README.md: Add nightly installation instructions for auto-round (#1505)

chensuyue committed 7 days ago

Verified 1bca8f72

[bug fix] force ignore mlp.gate (#1501)

xin3he committed 11 days ago

Verified f5f6b059

Add handle_generation_config function to manage model generation_config saving failure (#1448)

xin3he committed 12 days ago

Verified 32c1e3c1

Update MOE linear_loop implementation for speedup and matching name in vLLM (#1478)

xin3he committed 12 days ago

Verified d1501180

Update torch to 2.10.0 in CPU CI (#1492)

XuehaoSun committed 12 days ago

Verified 1e72e6a9

gguf better support for transformers5.0 and fix bug of Qwen3Next (#1474)

n1ck-guo committed 13 days ago

Verified ab14698c

Grep core dumped issue in UT test (#1484)

chensuyue committed 13 days ago

Verified c2152e48

add qwen35 ut (#1482)

wenhuach21 committed 14 days ago

Verified 2b5d1175

Older