GitLab

* Remove vLLM dependency for CUDA

This change adds `attention-kernels` as a dependency for paged
attention and cache reshaping. With that, we don't use vLLM
anywhere for CUDA.

Tested run (since we don't have paged attention in CI):

```
❯ ATTENTION=paged python -m pytest integration-tests -k "llama and awq" --release
[...]
5 snapshots passed.
```

* Fix clippy warning

compressed-tensors is a safetensors extension for sparse, quantized
tensors. The format is more powerful than earlier AWQ/GPTQ/FP8
quantization, because

- Different quantizer configurations can be used for different targets.
- The format can specify input/output quantizers in addition to weight
  quantizers.
- Configurable exclusions for quantization.

This change adds a dependency on the `compressed-tensors` package for
its configuration parsing and layer matching functionality.

The following types of quantization are supported in this PR:

- W8A16 and W4A16 INT using GPTQ-Marlin kernels.
- W8A8 and W8A16 FP using FP8-Marlin and cutlass kernels.

Support for other quantization types will be added in subsequent PRs.

* Switch from fbgemm-gpu w8a8 scaled matmul to vLLM/marlin-kernels

Performance and accuracy of these kernels are on par (tested with Llama
70B and 405B). Removes a dependency and resolves some stability issues
we have been seeing.

* Update test snapshots

* Making prefix/flashinfer the default and testing the full release tests.

* Include flashinfer in the docker.

* Using prebuilt.

* Allowing window_left_size (dummy version).

* Disabling flashinfer/prefix caching on odd head_dim

* Disable prefix caching for lora.

* More specific codes.

* Update lock

* Updating integration tests with new values with FI/FD.

Remove paged as a default too, and using FD everywhere.

* Update cargo lock ?

* Upgrade to 1.80 because of bitstream...

* Everywhere 1.80

* Forgot last default place.

* Apply suggestions from code review
Co-authored-by: drbh <david.richard.holtz@gmail.com>

* Updated flake lock

* Tmp

* Upgrade resolution system for less errors in resolution.

* Remove lambda for cleaner function.

* Handling debugger.

* OVerride the env in server tests.

* Is this enough to make it work ?

* This seems to be working.

* Downgrade some logs.

* Fixing the default for vlm.

* Don't enable prefix caching on VLM just yet.

* Change `add_special_tokens` in order to have the correct tokens for chat
input and not (since it's super important with the prefixing now)

* Fixing prefix caching for flashdecoding.

* Update all models.

* Fixed flashinfer version.

* add_special_tokens is internal only

* Fixing seqlen with the new vlms.

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

* Fixing the test.

* Removing encoder_decoder (seq2seq).

* Update the chat test.

* Fixing the batching tokenization in flash causal lm.

* Truncating left for radix purposes.

* Oops this doesn't belong here.

* Put back default pure shell.

* Update server tests

- Default to throughput test in k6
- Use TGI_WIGGLE_ROOM to adjust wiggle room

* Only n_heads / process_group.size() are necessary.

* Revert the integrationt tests change (seem linked to head_size
modification).

* Adding error message when assert is violated.

* Fixing the free algorithm to handle times where the common prefix is
smaller.

* Apply suggestions from code review
Co-authored-by: OlivierDehaene <olivier@huggingface.co>

* Update server/text_generation_server/layers/attention/common.py
Co-authored-by: OlivierDehaene <olivier@huggingface.co>

* Fix disabling prefix caching - Fix windowing checks.

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

* Fmt.

---------
Co-authored-by: drbh <david.richard.holtz@gmail.com>
Co-authored-by: OlivierDehaene <olivier@huggingface.co>

* Upgrading exl2.

* Fixing the other pathways.

* Fix idefics.

* chore: update to torch 2.4

* remove un-necessary patch

* fix

* feat(fp8): add support for fbgemm

* allow loading fp8 weights directly

* update outlines

* fix makefile

* build fbgemm

* avoid circular import and fix dockerfile

* add default dtype

* refactored weights loader

* fix auto conversion

* fix quantization config parsing

* force new nccl on install

* missing get_weights implementation

* increase timeout

Remove bitsandbytes installation when running cpu-only install

* fix nccl issue

* add note in dockerfile

* use v2.22.3 that also fixes @samsamoa's repro

* poetry actually can't handle the conflict between torch and nccl

* set LD_PRELOAD

* feat: first draft load multiple lora

* feat: load weights within layer and refactor lora pass

* fix: refactor and reduce lora math

* feat: baseline impl single request multi lora support

* feat: prefer lorax implementation and port loading logic

* fix: prefer adapter_data and refactors

* feat: perfer loraxs custom punica kernels and add mlp loras

* fix: adjust batch for bgmv

* fix: adjust adapter_segments logic when in batch

* fix: refactor and move changes to v3 proto

* fix: pass model_id for all flash causal lms

* fix: pass model_id for all causal and seq2seq lms

* fix: add model_id to model test

* feat: add lora support to mistral and refactors

* feat: prefer model id in request

* fix: include rust code for adapter id

* feat: bump launcher and add new lora docs

* feat: support base model generation and refactors

* fix: rename doc to retry ci build

* feat: support if vlm models

* fix: add adapter_data param and avoid missing layers

* fix: add adapter_data param to phi and neox

* fix: update all models forwards to include adapter_data

* fix: add model_id to IdeficsCausalLM

* Update lora.md

Fixed a typo

* Update lora.md

Fixing spam image

* fix: add lora kernel to dockerfile, support running without kernels and refactors

* fix: avoid dockerfile conflict

* fix: refactors and adjust flash llama lora logic

* fix: skip llama test due to CI issue (temp)

* fix: skip llama test CI (temp) 2

* fix: revert skips and prefer updated ci token for tests

* fix: refactors and helpful comments

* fix: add noop in TensorParallelAdapterRowLinear too

* fix: refactor and move shard_lora_weights logic

* fix: exit early if no adapter_data

---------
Co-authored-by: Derek <datavistics@gmail.com>

Add support for GPTQ Marlin kernels

GPTQ Marlin extends the Marlin kernels to support common GPTQ
configurations:

- bits: 4 or 8
- groupsize: -1, 32, 64, or 128
- desc_act: true/false

Using the GPTQ Marlin kernels requires repacking the parameters in the
Marlin quantizer format.

The kernels were contributed by Neural Magic to VLLM. We vendor them
here for convenience.

This change adds support for Marlin-quantized models. Marlin is an
FP16xINT4 matmul kernel, which provides good speedups decoding batches
of 16-32 tokens. It supports quantized models with symmetric
quantization, groupsize -1 or 128, and 4-bit.

Tested with:

- Llama 2
- Llama 3
- Phi 3

# What does this PR do?

Making `make install` a much better sane default to start local dev
environments.

<!--
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

 -->

- Refactor code to allow supporting multiple versions of the
generate.proto at the same time
- Add v3/generate.proto (ISO to generate.proto for now but allow for
future changes without impacting v2 backends)
- Add Schedule trait to abstract queuing and batching mechanisms that
will be different in the future
- Add SchedulerV2/V3 impl

@Narsil @drbh this will update flash attention v2 and vllm.
You will need to re-install them.

This draft PR is a work in progress implementation of the mamba model.
This PR currently loads weights, and produces correct logits after a
single pass.

This PR still needs to correctly integrate this model so it produces
tokens as expected, and apply optimization to avoid all copies during
runtime/unnecessary operations.

#### Helpful resources
[Mamba: Linear-Time Sequence Modeling with Selective State Spaces
(Albert Gu and Tri Dao)](https://arxiv.org/abs/2312.00752)
https://github.com/johnma2006/mamba-minimal

https://github.com/huggingface/candle/blob/main/candle-examples/examples/mamba-minimal/model.rs
https://github.com/huggingface/transformers/pull/28094



Notes: this dev work is currently targeting `state-spaces/mamba-130m`,
so if you want to test please use that model. Additionally when starting
the router the prefill needs to be limited: `cargo run --
--max-batch-prefill-tokens 768 --max-input-length 768`


## Update / Current State

Integration tests have been added and basic functionality such as model
loading is supported.

```bash
cd integration-tests
pytest -vv models/test_fused_kernel_mamba.py
```
- [x] add tests
- [x] load model
- [x] make simple request 
- [ ] resolve warmup issue
- [ ] resolve output issues


fetching models tested during dev
```bash
text-generation-server download-weights state-spaces/mamba-130m
text-generation-server download-weights state-spaces/mamba-1.4b
text-generation-server download-weights state-spaces/mamba-2.8b
```

The server can be run 
```bash
cd server
 MASTER_ADDR=127.0.0.1 MASTER_PORT=5555 python text_generation_server/cli.py serve state-spaces/mamba-2.8b
```

router
```bash
cargo run
```

make a request
```bash
curl -s localhost:3000/generate \
    -X POST \
    -d '{"inputs":"What is Deep Learning?","parameters":{"max_new_tokens":20}}' \
    -H 'Content-Type: application/json' | jq
```

response
```json
{
  "generated_text": "\n\nDeep learning is a machine learning technique that uses a deep neural network to learn from data."
}
```

---------
Co-authored-by: Nicolas Patry <patry.nicolas@protonmail.com>

This PR adds support for AMD Instinct MI210 & MI250 GPUs, with paged
attention and FAv2 support.

Remaining items to discuss, on top of possible others:
* Should we have a
`ghcr.io/huggingface/text-generation-inference:1.1.0+rocm` hosted image,
or is it too early?
* Should we set up a CI on MI210/MI250? I don't have access to the
runners of TGI though.
* Are we comfortable with those changes being directly in TGI, or do we
need a fork?

---------
Co-authored-by: Felix Marty <felix@hf.co>
Co-authored-by: OlivierDehaene <olivier@huggingface.co>
Co-authored-by: Your Name <you@example.com>

Close #1142

# 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

 -->

---------
Co-authored-by: zhaosida <zhaosida@corp.netease.com>

# Add AWQ quantization inference support

Fixes
https://github.com/huggingface/text-generation-inference/issues/781

This PR (partially) adds support for AWQ quantization for inference.
More information on AWQ [here](https://arxiv.org/abs/2306.00978). In
general, AWQ is faster and more accurate than GPTQ, which is currently
supported by TGI.

This PR installs 4-bit GEMM custom CUDA kernels released by AWQ authors
(in `requirements.txt`, just one line change).

Quick way to test this PR would be bring up TGI as follows:

```
text-generation-server download-weights abhinavkulkarni/codellama-CodeLlama-7b-Python-hf-w4-g128-awq

text-generation-launcher \
--huggingface-hub-cache ~/.cache/huggingface/hub/ \
--model-id abhinavkulkarni/codellama-CodeLlama-7b-Python-hf-w4-g128-awq \
--trust-remote-code --port 8080 \
--max-input-length 2048 --max-total-tokens 4096 --max-batch-prefill-tokens 4096 \
--quantize awq
```

Please note:
* This PR was tested with FlashAttention v2 and vLLM.
* This PR adds support for AWQ inference, not quantizing the models.
That needs to be done outside of TGI, instructions

[here](https://github.com/mit-han-lab/llm-awq/tree/f084f40bd996f3cf3a0633c1ad7d9d476c318aaa).
* This PR only adds support for `FlashLlama` models for now.
* Multi-GPU setup has not been tested. 
* No integration tests have been added so far, will add later if
maintainers are interested in this change.
* This PR can be tested on any of the models released

[here](https://huggingface.co/abhinavkulkarni?sort_models=downloads#models).

Please refer to the linked issue for benchmarks for

[abhinavkulkarni/meta-llama-Llama-2-7b-chat-hf-w4-g128-awq](https://huggingface.co/abhinavkulkarni/meta-llama-Llama-2-7b-chat-hf-w4-g128-awq)
vs

[TheBloke/Llama-2-7b-Chat-GPTQ](https://huggingface.co/TheBloke/Llama-2-7b-Chat-GPTQ).

Please note, AWQ has released faster (and in case of Llama, fused)
kernels for 4-bit GEMM, currently at the top of the `main` branch at
https://github.com/mit-han-lab/llm-awq, but this PR uses an older commit
that has been tested to work. We can switch to latest commit later on.

## Who can review?

@OlivierDehaene OR @Narsil

---------



# 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

 -->

---------
Co-authored-by: Abhinav M Kulkarni <abhinavkulkarni@gmail.com>
Co-authored-by: Abhinav Kulkarni <abhinav@concentric.ai>

# What does this PR do?

For consistency and ease of use (you can just run `make` to install vllm
without any extra steps).

<!--
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

 -->

# 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>

Fixes #333

---------
Co-authored-by: Nicolas Patry <patry.nicolas@protonmail.com>

Generates .pyi files for protobuf stubs which provide strong typing
information. Very helpful for IDE auto-completion, etc.