intel/auto-round

Pull Requests Commits

add fp8 kvcache support

yiliu30 committed 228 days ago

eed46a63

fix cuda hanging issue during packing (#669)

WeiweiZhang1 committed 228 days ago

Verified 7863f69d

fix ut issue (#668)

WeiweiZhang1 committed 228 days ago

Verified 0a8ebdfb

handle zeros in imatrix (#667)

wenhuach21 committed 229 days ago

Verified c1ea0d98

fix bugs of export deepseek gguf format when iters=0 and q3k accuracy (#665)

n1ck-guo committed 229 days ago

Verified 923683ae

fix gguf block-wise issue (#664)

wenhuach21 committed 229 days ago

Verified 313f436e

fix bug of block-wise quant imatrix (#663)

n1ck-guo committed 229 days ago

Verified 5d68aeb7

support vlm models for gguf format

n1ck-guo committed 230 days ago

Verified a2f3f016

fix bug of q4_0 and q5_0 at iters==0 (#658)

n1ck-guo committed 235 days ago

Verified 750a5c10

fix some issues (#655)

wenhuach21 committed 235 days ago

Verified 2f22e29a

support two more calib datasets and fix embedding layer bug (#653)

wenhuach21 committed 236 days ago

Verified 7d724036

speedup mxfp tuning and fix nvfp bug (#647)

wenhuach21 committed 237 days ago

Verified fddcb5d8

Remove cuda UT dependencies version restrictions (#651)

XuehaoSun committed 237 days ago

Verified 35cf154f

Integrate RTN quantization into GGUF packing to enhance robustness (#644)

n1ck-guo committed 237 days ago

Verified 06e0f520

[W4A8]Fix Packing (#648)

yiliu30 committed 238 days ago

Verified a5fa835e

fix norm_bias_tuning (#639)

wenhuach21 committed 238 days ago

Verified 28b01fe3

support llmcompressor format (#646)

xin3he committed 238 days ago

Verified bba62c2b

support packing immediately for gguf to reduce ram usage (#638)

wenhuach21 committed 241 days ago

Verified 3b282dc9

[BUG FIX] fix bug of deepseek gguf:q*k (#637)

n1ck-guo committed 242 days ago

Verified ba5e35c1

fix gguf bug of moe models and lmhead/embedding bits setting regression (#628)

n1ck-guo committed 244 days ago

Verified 20eed5b6

refine gguf rtn algorithm and fix bugs (#630)

wenhuach21 committed 244 days ago

Verified ef6413b2

fix vlm hf config loading issue (#624)

WeiweiZhang1 committed 249 days ago

Verified 51493904

fix not using imatrix for gguf at rtn mode (#623)

wenhuach21 committed 249 days ago

Verified 5181fa02

refine_doc_0624 (#619)

WeiweiZhang1 committed 251 days ago

Verified b6d24385

add auto-round related vllm and transformers UT (#613)

WeiweiZhang1 committed 252 days ago

Verified 898d630d

fix bug of q5_k_s w/ imatrix (#620)

n1ck-guo committed 252 days ago

Verified 3bbddfcc

FP8 WOQ export (#617)

wenhuach21 committed 253 days ago

Verified efff6ca0

fix double quant sym bug (#616)

wenhuach21 committed 257 days ago

Verified 21ecd8b4

support customized mixed bits for gguf (#615)

wenhuach21 committed 257 days ago

Verified e06bdbc3

fix gguf packing bug and moe regression (#614)

wenhuach21 committed 258 days ago

Verified 7d18ddbd

Older