huggingface/text-generation-inference

Pull Requests Commits

mht-sharma committed 1 year ago

1fa9ca2f

TensorRT-LLM backend bump to latest version + misc fixes (#2791)

mfuntowicz committed 1 year ago

Verified ea7f4082

Fixup opt to reduce the amount of odd if statements. (#2833)

Narsil committed 1 year ago

Verified 3bb3fd19

fix facebook/opt-125m not working issue (#2824)

sywangyi committed 1 year ago

Verified bf59118a

Fixing latest flavor by disabling it. (#2831)

Narsil committed 1 year ago

Verified c3bd7212

docs(README): supported hardware links TGI AMD GPUs (#2814)

guspan-tanadi committed 1 year ago

Verified f01f2fb6

Prepare patch release. (#2829)

Narsil committed 1 year ago

Verified 07b01293

Update README.md (#2827)

RodriMora committed 1 year ago

Verified cc66dccb

Using both value from config as they might not be correct. (#2817)

Narsil committed 1 year ago

Verified 82c24f74

Small update to docs (#2816)

Narsil committed 1 year ago

Verified a2d878fa

Hotfix link2 (#2812)

Narsil committed 1 year ago

Verified b2fac5d9

Hotfixing the link. (#2811)

Narsil committed 1 year ago

Verified a70dd299

Prep new version (#2810)

Narsil committed 1 year ago

Verified 042791fb

Narsil committed 1 year ago

Verified 27fa83ca

Attempt for cleverer auto batch_prefill values (some simplifications). (#2808)

Narsil committed 1 year ago

Verified a04356fb

Enable paligemma2 (#2807)

drbh committed 1 year ago

Verified 9f5c9a5e

Removing experimental to prefill chunking.

Narsil committed 1 year ago

Verified 08f6fa0b

Adding A100 compute. (#2806)

Narsil committed 1 year ago

Verified d96dcb17

Auto max prefill (#2797)

Narsil committed 1 year ago

Verified 5df80590

feat: auto max_new_tokens (#2803)

OlivierDehaene committed 1 year ago

Verified 8c3669b2

use oneapi 2024 docker image directly for xpu (#2793)

sywangyi committed 1 year ago

Verified 6685e8fc

fix: avoid setting use_sgmv if no kernels present (#2796)

drbh committed 1 year ago

Verified e0db6333

Saving some VRAM. (#2790)

Narsil committed 1 year ago

Verified b57f3703

Sync (most) server dependencies with Nix (#2782)

danieldk committed 1 year ago

Verified 2003d8be

fix: only use eos_token_id as pad_token_id if int (#2774)

dvrogozh committed 1 year ago

Verified 535149d8

fix: add merge-lora arg for model id (#2788)

drbh committed 1 year ago

Verified 2c74c556

Removing ../ that broke the link (#2789)

Getty committed 1 year ago

Verified a35d1e6f

Fix doc. (#2792)

Narsil committed 1 year ago

Verified 1d2cb356

Support continue final message (#2733)

drbh committed 1 year ago

Verified d4718051

Fix: docs typo (#2777)

jp1924 committed 1 year ago

Verified caff779d

Older