GitLab

* Add links to Adyen blogpost

* Adding to toctree.

* Update external.md

* Update _toctree.yml

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

These should all be cheap assertions.

Also:

* Fixup some comments.
* Delete a `remove` that was done unnecessarily twice.

We need this to ensure that pyright/ruff are part of the same
interpreter/venv.

fix regression caused by attention api change. ipex.varlen_attention does not support paged-cache
format kv input now.
Signed-off-by: Wang, Yi A <yi.a.wang@intel.com>

The minimum batch size logic could cause prefix blocks to be
deallocated without prefill. The next allocation of the same
prefix would then use garbage blocks.

* feat: support lora revisions and qkv_proj weights

* fix: add qkv_proj weights to weight test

* fix: enable chat requests in vertex endpoint

* feat: avoid unwrap and pre allocate future vec

Enables LoRA support.

- Add some test dependencies.
- Install server in venv.
- Install Python client in venv.

* Tied embeddings in MLP speculator.

* Fixing the scale_weight when users decide to not use the speculation as
much as defined in the config.

* Adding scaling support + optimize some ops.

* update doc with intel cpu part
Signed-off-by: Wang, Yi A <yi.a.wang@intel.com>

* Apply suggestions from code review

we do not use latest ever in documentation, it causes too many issues for users. Release number get update on every release.

---------
Signed-off-by: Wang, Yi A <yi.a.wang@intel.com>
Co-authored-by: Nicolas Patry <patry.nicolas@protonmail.com>

* feat: add /v1/models endpoint

* feat: add /v1/models endpoint

* fix: remove unused type import

* fix: revert route typo

* fix: update docs with new endpoint

* fix: add to redocly ignore and lint

* 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>

Updates tgi-nix input:

- Move Torch closer to upstream by building against MKL.
- Remove compute capability 8.7 from Torch (Jetson).
- Sync nixpkgs cumpute capabilities with Torch (avoids
  compiling too mana capabilities for MAGMA).
- Use nixpkgs configuration passed through by `tgi-nix`.

* fix: support tojson and avoid message indexing issue in template

* fix: prefer minijinja native methods and prefer workspace level dependency

* fix: adjust comment typo

* fix[router]: Fix tools not passed in chat template
Signed-off-by: GitHub <noreply@github.com>

* feat: improve default tool serialization and lints

* feat: refactor tool logic to include notify_error in prompt and adjust typing

* fix: adjust non tool template apply

* fix: simplify tool grammar logic and improve schema

* feat: avoid skip tool test and avoid empty tool prompts

* fix: increase test client timeout for grammar compilation tests

---------
Signed-off-by: GitHub <noreply@github.com>
Co-authored-by: Simone Rossi <simone.rossi.93@gmail.com>

* Fix: don't apply post layernorm in SiglipVisionTransformer

This fixes a bug with LLaVA Next when using Siglip as the vision model. LLaVA Next expects the output of the vision model to be the encoder outputs before layernorm (see original transformers implementation here: https://github.com/huggingface/transformers/blob/main/src/transformers/models/llava_next/modeling_llava_next.py#L813).

This also makes Siglip consistent with the existing Clip implementation:

https://github.com/huggingface/text-generation-inference/blob/main/server/text_generation_server/models/custom_modeling/clip.py#L613



* fix: adjust pali gemma for post layer norm and small refactors

---------
Co-authored-by: Travis Addair <tgaddair@gmail.com>

The default package wraps the launcher and puts the server/router in the
path.

As a result, TGI can be started using something like:

```
nix run .# -- \
  --model-id hugging-quants/Meta-Llama-3.1-8B-Instruct-AWQ-INT4 \
  --port 8080
```

nix: add text-generation-benchmark to pure devshell

* nix: pure server and support both pure and impure devShells

* nix: remove unused poetry2nix input

It is not wired up and we now have a pure server.

* nix: add ipdb to impure devshell

* Prefix caching WIP

* Fixing prefix attention.

* Fixing flashinfer import.

* Fixing black.

* Fixing medusa (still wrong outputs, but functional).

* Just medusa values now.

* Fixing medusa without prefix caching.

* Fixing prefix caching.

* Medusa requires reshaping.

* Removing the logs.

* Remove router.nix

* Fixup:

- Remove logs
- Disable VLMs (they do not work)
- Disable prefix caching when user wants prefill logprobs.

* Update flake.lock

---------
Co-authored-by: Daniël de Kok <me@danieldk.eu>

* Update to CUDA 12.4

* poetry2nix: follow tgi-nix nixpkgs

* All integration tests back everywhere (too many failed CI).

* Upgrade integration tests after 12.4

* Attempt to remove the specifed compute cap.

* Common arch list.

* Punica uses raw ASM which is not valid on 9.0 apparently.

* doc: Add metrics documentation and add a 'Reference' section

* doc: Add API reference

* doc: Refactor API reference

* fix: Message API link

* Bad rebase

* Moving the docs.

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

* Further fixes.

* Update the conftest to allow NaN (first logprob).

* Fix the condition.

* Improve the Consuming TGI docs.

* Fix erronous update to .

* add info about Open AI client.

* More updates.

* Apply suggestions from code review
Co-authored-by: Erik Kaunismäki <erik.kaum@gmail.com>

* Suggestions from Lucain.

* Update Gradio snippet.

* Up.

* Apply suggestions from code review
Co-authored-by: Lucain <lucainp@gmail.com>

* Update docs/source/basic_tutorials/consuming_tgi.md
Co-authored-by: Lucain <lucainp@gmail.com>

* Up.

* Apply suggestions from code review
Co-authored-by: Omar Sanseviero <osanseviero@gmail.com>

* Up.

* Up.

* Doc review from Nico.

* Doc review from Nico. x2

* Last nit

---------
Co-authored-by: Erik Kaunismäki <erik.kaum@gmail.com>
Co-authored-by: Lucain <lucainp@gmail.com>
Co-authored-by: Omar Sanseviero <osanseviero@gmail.com>

Try to reduce the number of router/launcher rebuilds by filtering
sources. In this way, recompiles should only be triggered by changes
in Cargo or Rust files.

* Fixing exl2 and other quanize tests again.

* Mark exl2 as non release (so CI tests them, needs to be removed latet).

* Fixing exl2 (by disabling cuda graphs)

* Fix quantization defaults without cuda graphs on exl2 (linked to new
issues with it).

* Removing serde override.

* Go back to released exl2 and remove log.

* Adding warnings for deprecated bitsandbytes + upgrade info to warn.

* (backend) use parking_lot crate for RwLock fairness

* (docker) let's put rust in the TRTLLM folder when building

* (docker) build ompi with SLURM support

* (launcher) default new server::run parameters to false for now

* (chore) fmt ... why?