 AlpinDale
|
22422d962b
feat: add cuda sampling kernels for top_k and top_p
|
2 月之前 |
|
AlpinDale
|
22427602eb
feat: add top-nsigma sampling method
|
2 月之前 |
|
AlpinDale
|
22429e4a10
fix: sampler test with new transformers version
|
2 月之前 |
 AlpinDale
|
2f61644f6e
SPMD optimizations (#824)
|
2 月之前 |
|
AlpinDale
|
32a37e8107
tests: partially fix tensorizer and logprobs tests
|
2 月之前 |
|
AlpinDale
|
7f1c9af5e2
fix: fp8 quant test
|
2 月之前 |
|
AlpinDale
|
173ac23399
fix: experts int8 quant test
|
2 月之前 |
|
AlpinDale
|
68f050129d
fix: lora worker manager test import
|
2 月之前 |
|
AlpinDale
|
3661de812d
fix: lora layer test
|
2 月之前 |
|
AlpinDale
|
0a369f9171
feat: support chunked prefill with LoRA (#823)
|
2 月之前 |
|
AlpinDale
|
e5b1afe625
feat: add chat method for LLM class (#822)
|
2 月之前 |
|
AlpinDale
|
262cbc63b7
fix: vision api test template path
|
2 月之前 |
|
AlpinDale
|
b0113a1eaa
fix: tokenization api test (#821)
|
2 月之前 |
|
AlpinDale
|
c6c91edab7
ci: update & overhaul test units (#769)
|
2 月之前 |
|
AlpinDale
|
f088ea81c7
fix: --max-seq-len-to-capture arg (#818)
|
2 月之前 |
 50h100a
|
a5346b2ea5
Merge pull request #814 from PygmalionAI/50h100a-temp-fix
|
2 月之前 |
|
50h100a
|
273c61d406
guard against nan temperature from dynatemp (or anywhere else).
|
2 月之前 |
|
50h100a
|
a22e887319
why we don't use the github website editor to make changes
|
2 月之前 |
|
50h100a
|
54a8320638
logit shenanigans to prevent even worse shenanigans
|
2 月之前 |
|
50h100a
|
b6a897d2a1
fix temperature, and address those pernicious dynatemp NaNs
|
2 月之前 |
|
50h100a
|
a61d00fad7
Merge pull request #813 from PygmalionAI/50h100a-patch-1
|
2 月之前 |
|
50h100a
|
83040c6389
Mask dynatemp using min/max, rather than exp
|
2 月之前 |
|
AlpinDale
|
2fa112f86b
feat: update to serviceinfo v0.2 (#808)
|
2 月之前 |
|
AlpinDale
|
72fbfa1b5b
feat: add serviceinfo endpoint (#807)
|
2 月之前 |
|
AlpinDale
|
6e25b03f25
ci: docker build and upload script
|
2 月之前 |
|
AlpinDale
|
6145deab4a
frontend: enable kobold api by default (#803)
|
2 月之前 |
|
AlpinDale
|
f0e00f1b43
ci: bump to 0.6.3.post1 (#801)
|
2 月之前 |
|
AlpinDale
|
5b0eabe0e8
fix: compilation of gptq_marlin_gemm object (#800)
|
2 月之前 |
 dependabot[bot]
|
7e2d2e7ae7
build(deps): bump rollup from 4.21.0 to 4.24.3 in /docs (#796)
|
2 月之前 |
|
AlpinDale
|
76c05c5591
ci: bump version to 0.6.3 (#799)
|
2 月之前 |
