intel/auto-round

Pull Requests Commits

wenhuach21 committed 315 days ago

deccac77

wenhuach21 committed 316 days ago

4d8b622a

[HPU]Fix compile bug when quant layer (#441)

yiliu30 committed 319 days ago

Verified 19601d4e

align auto_quantizer with main branch in Transformers (#437)

WeiweiZhang1 committed 320 days ago

Verified 7d27f670

rm gc collect in packing (#438)

wenhuach21 committed 321 days ago

Verified 082f01e7

fix nblocks issues (#432)

wenhuach21 committed 323 days ago

Verified edc2b034

Fix packing hang, torch compile and force to fp16 at exporting (#430)

wenhuach21 committed 324 days ago

Verified 40ab9825

bump to v0.4.5 (#425)

XuehaoSun committed 338 days ago

Verified 096bc196

fix eval trust_remote_code (#424)

n1ck-guo committed 339 days ago

Verified 5c075a9f

fix cuda UT torch_dtype (#423)

WeiweiZhang1 committed 340 days ago

Verified db8088b2

support to export deepseek v3 gguf format (#416)

n1ck-guo committed 341 days ago

Verified 6a49304e

fix evaluation device bug, relate to issue 413 (#419)

n1ck-guo committed 342 days ago

Verified 22315cc5

enable backup pile dataset loading (#417)

WeiweiZhang1 committed 344 days ago

Verified a9b5ae4f

support bf16 inference (#420)

wenhuach21 committed 345 days ago

Verified 29c5119f

support naive multi-card tuning (#415)

wenhuach21 committed 348 days ago

Verified 09f1ab49

speedup packing stage for autogptq and autoawq format (#407)

wenhuach21 committed 350 days ago

Verified 937d0198

fix log bug and device "auto" bug (#409)

n1ck-guo committed 351 days ago

Verified ba5975c8

update format readme (#411)

wenhuach21 committed 352 days ago

Verified a40ce88d

support to export gguf q4_0 and q4_1 format (#393)

n1ck-guo committed 357 days ago

Verified 86767b0f

support awq with qbits, only support sym (#402)

wenhuach21 committed 358 days ago

Verified ba2426cd

llm_layer_config_bugfix (#406)

WeiweiZhang1 committed 358 days ago

Verified 7714796a

support for deepseek vl2 (#401)

n1ck-guo committed 358 days ago

Verified bf8d68d1

fix dtype converting issue (#403)

wenhuach21 committed 359 days ago

Verified 6ac7a2b2

Fix auto-round install & bump into 0.4.4 (#387)

XuehaoSun committed 363 days ago

Verified 4fd9bbd7

fix device auto issue (#400)

wenhuach21 committed 363 days ago

Verified 73fcf7d7

fix seqlen issue (#399)

WeiweiZhang1 committed 1 year ago

Verified d06546aa

support awq format for vlms (#398)

WeiweiZhang1 committed 1 year ago

Verified c82aa88a

enhance calibration dataset and add awq pre quantization warning (#396)

wenhuach21 committed 1 year ago

Verified 01b779c0

vlm 70B+ in single card (#395)

n1ck-guo committed 1 year ago

Verified 42a6eb91

fix a critic bug of static activation quantization (#392)

wenhuach21 committed 1 year ago

Verified 3dd8ae61

Older