huggingface/text-generation-inference

Pull Requests Commits
Update the link for qwen2 (#2068)
xianbaoqian committed 1 year ago
Verified 96b7b40c
Add support for GPTQ Marlin (#2052)
danieldk committed 1 year ago
Verified 093a27c5
implement Open Inference Protocol endpoints (#1942)
drbh committed 1 year ago
Verified f433f1f7
PR #2049 CI run (#2054)
drbh committed 1 year ago
Verified 42aa8ee1
fix(layers): fix SuRotaryEmbedding (#2060)
OlivierDehaene committed 1 year ago
Verified 90184df7
fix(server): fix OPT implementation (#2061)
OlivierDehaene committed 1 year ago
Verified 521de6ca
Support chat response format (#2046)
drbh committed 1 year ago
Verified 376a0b7a
Update LLMM1 bound (#2050)
fxmarty committed 1 year ago
Verified a6e4d63c
fix(ci): remove unnecessary permissions (#2045)
McPatate committed 1 year ago
Verified dfca1dfc
feat(ci): add trufflehog secrets detection (#2038)
McPatate committed 1 year ago
Verified 4e74ec09
Add Phi-3 medium support (#2039)
danieldk committed 1 year ago
Verified 85dfc392
ROCm and sliding windows fixes (#2033)
fxmarty committed 1 year ago
Verified 9b3674d9
server: use chunked inputs
danieldk committed 1 year ago
bf3c8137
Xpu gqa (#2013)
sywangyi committed 1 year ago
Verified 4dabddb7
Revert "Enabling CI for AMD with new runner.."
Narsil committed 1 year ago
97656582
Enabling CI for AMD with new runner..
Narsil committed 1 year ago
101ac9a7
Internal runner ? (#2023)
Narsil committed 1 year ago
Verified ed1cfde0
marlin: improve build
danieldk committed 1 year ago
51621439
marlin: support tp>1 when group_size==-1
danieldk committed 1 year ago
0d96468e
Add support for Marlin-quantized models
danieldk committed 1 year ago
4594e6fa
Revert "Less cache misses on cargo build."
Narsil committed 1 year ago
cf0d459a
Less cache misses on cargo build.
Narsil committed 1 year ago
5aec4154
Update __version__ on __init__.py to 0.7.0 (#2017)
andimarafioti committed 1 year ago
Verified 2a48a100
Fix GPTQWeight import (#2020)
danieldk committed 1 year ago
Verified 3f4bcf97
Fixing rocm. (#2021)
Narsil committed 1 year ago
Verified 0a94fad7
feat: move allocation logic to rust (#1835)
OlivierDehaene committed 1 year ago
Verified 8aece3bd
Do not initialize scratch space when there are no ExLlamaV2 layers (#2015)
danieldk committed 1 year ago
Verified 9ffe1f1e
Hotfixing `make install`. (#2008)
Narsil committed 1 year ago
Verified 824edf28
Making `make install` work better by default. (#2004)
Narsil committed 1 year ago
Verified 8390e251
Support GPTQ models with column-packed up/gate tensor (#2006)
danieldk committed 1 year ago
Verified d14eaaca
Terms Of Use Privacy Policy Refund Policy Impressum