huggingface/text-generation-inference

Pull Requests Commits

Add a stale bot. (#1313)

Narsil committed 2 years ago

Verified 3238c491

Fix AMD documentation (#1307)

fxmarty committed 2 years ago

Verified 25b5f819

OlivierDehaene committed 2 years ago

ccd5725a

Make GPTQ test less flaky (#1295)

Narsil committed 2 years ago

Verified 624800c4

Let each model resolve their own default dtype. (#1287)

Narsil committed 2 years ago

Verified ba552e1a

`make install-flash-attn-v2-cuda` should work like `make install-flash-attn-v2` used to work. (#1294)

Narsil committed 2 years ago

Verified 3c71c656

Add RoCm support (#1243)

fxmarty committed 2 years ago

Verified b2b5df0e

Exllama v2 (#1211)

Narsil committed 2 years ago

Verified ed2a3f61

Reduce race condition on file system for test

Narsil committed 2 years ago

3c02262f

Fix IDEFICS dtype (#1214)

vakker committed 2 years ago

Verified c6bb7670

chore: update to torch 2.1.0 (#1182)

OlivierDehaene committed 2 years ago

Verified 35509ff5

Load PEFT weights from local directory (#1260)

tleyden committed 2 years ago

Verified e12c34bd

Fix missing `trust_remote_code` flag for AutoTokenizer in utils.peft (#1270)

creatorrr committed 2 years ago

Verified 91111a0d

Update README.md (#1272)

Narsil committed 2 years ago

Verified b226e469

fix: do not leak inputs on error (#1228)

OlivierDehaene committed 2 years ago

Verified 3dbc649b

OlivierDehaene committed 2 years ago

8acdc1fa

OlivierDehaene committed 2 years ago

457e72c3

Fix link in quantization guide (#1246)

osanseviero committed 2 years ago

Verified a5def7c2

Update README.md (#1242)

Narsil committed 2 years ago

Verified 7b5c1674

Narsil patch 1 (#1241)

Narsil committed 2 years ago

Verified b9184093

Adding the video -> moving the architecture picture lower (#1239)

Narsil committed 2 years ago

Verified 414a911b

fix: better warmup error

OlivierDehaene committed 2 years ago

96a982ad

feat: remove flume (#1184)

OlivierDehaene committed 2 years ago

Verified f9910d13

feat: paged attention v2 (#1183)

OlivierDehaene committed 2 years ago

Verified 12590fdc

Fix link to quantization page in preparing_model.md (#1187)

aasthavar committed 2 years ago

Verified 63fa5346

#1049 CI (#1178)

OlivierDehaene committed 2 years ago

Verified 5e28f44a

fix: remove useless token (#1179)

rtrompier committed 2 years ago

Verified 72b8f88b

fix: EETQLinear with bias in layers.py (#1176)

SidaZh committed 2 years ago

Verified 648ea064

Fix: Replace view() with reshape() in neox_modeling.py to resolve RuntimeError (#1155)

Mario928 committed 2 years ago

Verified 9179605e

Fix calling cuda() on load_in_8bit (#1153)

mmngays committed 2 years ago

Verified 7402a355

Newer Older