Commit Graph

175 Commits

Author SHA1 Message Date
OlivierDehaene e496c9ba5b
feat(server): optimize dist ops (#434) 2023-06-09 11:55:29 +02:00
Nicolas Patry abd58ff82c
feat(server): Rework model loading (#344)
# What does this PR do?

Reworked the loading logic. Idea is to use cleaner loading code:

- Remove need for `no_init_weights`
- Remove all weird `bnb_linear` and `load_weights` and
`post_load_weights`.

New code layout:

- New class `Weights` in charge of handling loading the weights from
multiple files into appropiate tensors (potentially sharded)
- TP layers now are "shells", they contain the code to know what kind of
sharding we need + eventual `all_reduce`. They do not inherit from
linear, but they contain some kind of Linear instead
- the contained linear can be either FastLinear, BnbLinear or GPTq
Linear next.
- All modeling code is explictly made for sharding, process group is
just no-ops for non sharded code (removes a lot of test cases)

![Screenshot from 2023-05-19
23-19-59](https://github.com/huggingface/text-generation-inference/assets/204321/9a802654-74a3-488c-87a8-073743a6143f)

---------

Co-authored-by: Ubuntu <ubuntu@ip-172-31-41-161.taildb5d.ts.net>
Co-authored-by: Ubuntu <ubuntu@ip-172-31-41-161.ec2.internal>
Co-authored-by: OlivierDehaene <olivier@huggingface.co>
Co-authored-by: OlivierDehaene <23298448+OlivierDehaene@users.noreply.github.com>
2023-06-08 14:51:52 +02:00
OlivierDehaene 6abec14a7e
feat(server): batch tokenization for flash causal lm (#411) 2023-06-05 16:09:41 +02:00
OlivierDehaene 87dc034b59
feat(server): add retry on download (#384) 2023-05-31 10:57:53 +02:00
OlivierDehaene b8b950b37c
feat(server): support RefinedWeb models (#379) 2023-05-30 18:25:19 +02:00
OlivierDehaene 62f91f78ac
feat(server): support vectorized warpers in flash causal lm (#317)
Co-authored-by: Joel Lamy-Poirier <joel.lamy-poirier@servicenow.com>
2023-05-26 12:30:27 +02:00
OlivierDehaene e71471bec9
feat: add snapshot testing (#282) 2023-05-15 23:36:30 +02:00
Nicolas Patry f58f0a0364
Single place for TP layers + Dropout Layer Norm + FastLinear (#329)
# What does this PR do?

<!--
Congratulations! You've made it this far! You're not quite done yet
though.

Once merged, your PR is going to appear in the release notes with the
title you set, so make sure it's a great title that fully reflects the
extent of your awesome contribution.

Then, please replace this with a description of the change and which
issue is fixed (if applicable). Please also include relevant motivation
and context. List any dependencies (if any) that are required for this
change.

Once you're done, someone will review your PR shortly (see the section
"Who can review?" below to tag some potential reviewers). They may
suggest changes to make the code even better. If no one reviewed your PR
after a week has passed, don't hesitate to post a new comment
@-mentioning the same persons---sometimes notifications get lost.
-->

<!-- Remove if not applicable -->

Fixes # (issue)


## Before submitting
- [ ] This PR fixes a typo or improves the docs (you can dismiss the
other checks if that's the case).
- [ ] Did you read the [contributor
guideline](https://github.com/huggingface/transformers/blob/main/CONTRIBUTING.md#start-contributing-pull-requests),
      Pull Request section?
- [ ] Was this discussed/approved via a Github issue or the
[forum](https://discuss.huggingface.co/)? Please add a link
      to it if that's the case.
- [ ] Did you make sure to update the documentation with your changes?
Here are the
[documentation
guidelines](https://github.com/huggingface/transformers/tree/main/docs),
and
[here are tips on formatting
docstrings](https://github.com/huggingface/transformers/tree/main/docs#writing-source-documentation).
- [ ] Did you write any new necessary tests?


## Who can review?

Anyone in the community is free to review the PR once the tests have
passed. Feel free to tag
members/contributors who may be interested in your PR.

<!-- Your PR will be replied to more quickly if you can figure out the
right person to tag with @


@OlivierDehaene OR @Narsil

 -->
2023-05-15 17:30:47 +02:00
OlivierDehaene 4f6d038c0b fix(server): fix multinomial implem in Sampling 2023-05-11 13:30:38 +02:00
OlivierDehaene a6c18c39bb
feat(server): use cuda graph in logits warping (#302) 2023-05-10 19:08:54 +02:00
OlivierDehaene 68e9d6ab33
feat(server): shard token decode (#303) 2023-05-10 15:48:21 +02:00
Nicolas Patry b4aa87db58
fea(server): decrease convert RAM requirements (#286) 2023-05-05 17:57:02 +02:00
Nicolas Patry 690fc31757
fix(server): fix convert (#284) 2023-05-05 15:28:08 +02:00
Nicolas Patry f08343d44d
fix(server): Removes the parallelism in file convertion (during download) (#275) 2023-05-04 15:22:54 +02:00
OlivierDehaene 85aa7e2e7b
feat(server): support hf endpoint weight layout (#266) 2023-05-03 11:36:24 +02:00
OlivierDehaene f26dfd0dc1
feat(server): support OPT models (#55)
OPT models do not all have a `tokenizer.json` file on the hub at the
moment. Can't merge for now.
2023-04-11 19:16:41 +02:00
OlivierDehaene 3f2542bb6a
fix(server): fix escape characters in stop sequence (#155) 2023-04-05 19:37:41 +02:00
OlivierDehaene 610bb1f978
feat(benchmark): tui based benchmarking tool (#149) 2023-03-30 15:26:27 +02:00
OlivierDehaene d6a93fe992
fix(server): fix flash-neox scores warping (#137) 2023-03-24 18:21:41 +01:00
OlivierDehaene 05e9a796cc
feat(server): flash neoX (#133) 2023-03-24 14:02:14 +01:00
OlivierDehaene 8ad60b752f
fix(server): add position ids to neox (#126) 2023-03-15 13:12:49 +01:00
OlivierDehaene c0795de2f2
fix(server): do not warp prefill logits (#116) 2023-03-09 13:00:10 +01:00
OlivierDehaene 1a2d68250a
feat: support typical sampling (#114)
closes #112
2023-03-09 11:33:57 +01:00
OlivierDehaene 941cd42e0c
fix(server): fix index out of range for watermarking (#110) 2023-03-08 18:29:08 +01:00
OlivierDehaene 3fef90d50f
feat(clients): Python client (#103) 2023-03-07 18:52:22 +01:00