huggingface/text-generation-inference

Pull Requests Commits

Dummy fix for medusa.

Narsil committed 1 year ago

10dd0150

Narsil committed 1 year ago

b24bdb9f

Narsil committed 1 year ago

c31cb32d

Forgot to update docs.

Narsil committed 1 year ago

8cd198a6

Marking the flag as really not the fastest and BETA.

Narsil committed 1 year ago

eb40f8cc

Narsil committed 1 year ago

6568e481

Narsil committed 1 year ago

be59a6bc

Dummy but working version.

Narsil committed 1 year ago

e1e9a184

Narsil committed 1 year ago

50d5a3c1

hotfix: mixtral

OlivierDehaene committed 1 year ago

30620a9a

fix: fix CohereForAI/c4ai-command-r-plus (#1707)

OlivierDehaene committed 1 year ago

Verified ad9d6288

Adding Llava-Next (Llava 1.6) with full support. (#1709)

Narsil committed 1 year ago

Verified 4634b00c

Automatic quantization config. (#1719)

Narsil committed 1 year ago

Verified 106d8ee8

Revert license to Apache 2.0 (#1714)

OlivierDehaene committed 1 year ago

Verified ff42d33e

Regenerate ld.so.cache (#1708)

oOraph committed 1 year ago

Verified 53c2c3db

Force weights_only (before fully breaking pickle files anyway). (#1710)

Narsil committed 1 year ago

Verified 8dca3b04

Fixing cohere tokenizer. (#1697)

Narsil committed 1 year ago

Verified f9958ee1

Push users to streaming in the readme. (#1698)

Narsil committed 1 year ago

Verified 5062fda4

Pickle conversion now requires `--trust-remote-code`. (#1704)

Narsil committed 1 year ago

Verified c7e570e5

Add cuda graphs sizes and make it default. (#1703)

Narsil committed 1 year ago

Verified 99874eae

OlivierDehaene committed 1 year ago

Verified 4ee0a0c4

feat: Add dbrx support (#1685)

OlivierDehaene committed 1 year ago

Verified f04255c6

fix: handle batches with and without grammars (#1676)

drbh committed 1 year ago

Verified 762dbf3f

fix: adjust logprob response logic (#1682)

drbh committed 1 year ago

Verified 818aee37

OlivierDehaene committed 1 year ago

Verified 6c4496a1

feat: cohere (#1660)

OlivierDehaene committed 1 year ago

Verified 1e9bcd9d

Inline images for multimodal models. (#1666)

Narsil committed 1 year ago

Verified f171bdc8

fix: LlamaTokenizerFast to AutoTokenizer at flash_mistral.py (#1637)

SeongBeomLEE committed 1 year ago

Verified 66914f7b

feat: update client to 0.7 (#1667)

OlivierDehaene committed 1 year ago

Verified 08e91814

Repair idefics integration tests. (#1663)

Narsil committed 1 year ago

Verified deb440b3

Older