huggingface/text-generation-inference

Pull Requests Commits

Tied embeddings in MLP speculator.

Narsil committed 1 year ago

Verified 5838f213

Revert the Cohere tokenizer change (for now using a revision instead).

Narsil committed 1 year ago

Verified 5e293255

Fix disabling prefix caching - Fix windowing checks.

Narsil committed 1 year ago

Verified fc7ea202

Fixing the free algorithm to handle times where the common prefix is

Narsil committed 1 year ago

Verified bef2f6bd

Adding error message when assert is violated.

Narsil committed 1 year ago

Verified 9c839ca5

Revert the integrationt tests change (seem linked to head_size

Narsil committed 1 year ago

Verified e7e03638

Only n_heads / process_group.size() are necessary.

Narsil committed 1 year ago

Verified 8a4df6e1

Update server tests

Narsil committed 1 year ago

Verified 8d018483

Put back default pure shell.

Narsil committed 1 year ago

Verified 12325564

Oops this doesn't belong here.

Narsil committed 1 year ago

Verified f8867479

Truncating left for radix purposes.

Narsil committed 1 year ago

Verified e6ee67f3

Fixing the batching tokenization in flash causal lm.

Narsil committed 1 year ago

Verified 0a609731

Update the chat test.

Narsil committed 1 year ago

Verified c6f1a612

Removing encoder_decoder (seq2seq).

Narsil committed 1 year ago

Verified 8ac1ffa0

Fixing the test.

Narsil committed 1 year ago

Verified ccaf1d00

Fixing the issue with `add_special_tokens` not being passed around.

Narsil committed 1 year ago

Verified 2cf1f5c0

Fixing seqlen with the new vlms.

Narsil committed 1 year ago

Verified e0069a3a

add_special_tokens is internal only

Narsil committed 1 year ago

Verified 9dacac3b

Fixed flashinfer version.

Narsil committed 1 year ago

Verified 55d984d7

Update all models.

Narsil committed 1 year ago

Verified bb9769ed

Fixing prefix caching for flashdecoding.

Narsil committed 1 year ago

Verified 65b94a69

Change `add_special_tokens` in order to have the correct tokens for chat

Narsil committed 1 year ago

Verified 7f1816a4

Don't enable prefix caching on VLM just yet.

Narsil committed 1 year ago

Verified f1c07354

Fixing the default for vlm.

Narsil committed 1 year ago

Verified e30fb254

Downgrade some logs.

Narsil committed 1 year ago

Verified 27b566ba

This seems to be working.

Narsil committed 1 year ago

Verified 26e5037d

Is this enough to make it work ?

Narsil committed 1 year ago

Verified f5182c18

OVerride the env in server tests.

Narsil committed 1 year ago

Verified 1568e825

Handling debugger.

Narsil committed 1 year ago

Verified 682db34b

Remove lambda for cleaner function.

Narsil committed 1 year ago

Verified c53968dc

Older