* transformers flash llm/vlm enabling in xpu
Signed-off-by: Wang, Yi A <yi.a.wang@intel.com>
* ipex cpu could also support in function
Signed-off-by: Wang, Yi A <yi.a.wang@intel.com>
---------
Signed-off-by: Wang, Yi A <yi.a.wang@intel.com>
* clean cuda/rocm code in hpu backend, enable flat_hpu
Signed-off-by: Wang, Yi A <yi.a.wang@intel.com>
* fix TP in pageattn
Signed-off-by: Wang, Yi A <yi.a.wang@intel.com>
* adjust block table in hpu to improve performance
Signed-off-by: Wang, Yi A <yi.a.wang@intel.com>
* enable all the model. not testet yet
Signed-off-by: Wang, Yi A <yi.a.wang@intel.com>
* use tensor cache in hpu graph to avoid replay issue
Signed-off-by: Wang, Yi A <yi.a.wang@intel.com>
* add moe support, fix qwen/mistral/mixtral crash
Signed-off-by: Wang, Yi A <yi.a.wang@intel.com>
* fix phimoe issue
Signed-off-by: Wang, Yi A <yi.a.wang@intel.com>
* gpt_bigcode could also go pageattn
Signed-off-by: Wang, Yi A <yi.a.wang@intel.com>
* enable dbrx remove some unused code
Signed-off-by: Wang, Yi A <yi.a.wang@intel.com>
* multi-modality initial PR
Signed-off-by: Wang, Yi A <yi.a.wang@intel.com>
* adjust warmup and enable vlm
Signed-off-by: Wang, Yi A <yi.a.wang@intel.com>
* fix incorrect output in qwen2 idefics if hpu graph is used
Signed-off-by: Wang, Yi A <yi.a.wang@intel.com>
* remove unused quantization code and enable awq/gptq int4
Signed-off-by: Wang, Yi A <yi.a.wang@intel.com>
* fix gptq issue
Signed-off-by: Wang, Yi A <yi.a.wang@intel.com>
* enable fp8
Signed-off-by: Wang, Yi A <yi.a.wang@intel.com>
* warmup prefill
remove model where pageattn is not used, set block table to None since it's not used
Signed-off-by: Wang, Yi A <yi.a.wang@intel.com>
* add warmup_decode
Signed-off-by: Wang, Yi A <yi.a.wang@intel.com>
* warmup decode
Signed-off-by: Wang, Yi A <yi.a.wang@intel.com>
* remove block_tables and prefill_cache_indices which will lead to dynamic shape
Signed-off-by: Wang, Yi A <yi.a.wang@intel.com>
* fix comment
Signed-off-by: Wang, Yi A <yi.a.wang@intel.com>
* missing gptj change...
Signed-off-by: Wang, Yi A <yi.a.wang@intel.com>
* fix some issue
Signed-off-by: Wang, Yi A <yi.a.wang@intel.com>
* remove torch.where to fix incorrect output in hpu graph model
Signed-off-by: Wang, Yi A <yi.a.wang@intel.com>
* match the latest vllm_extension ops
Signed-off-by: Wang, Yi A <yi.a.wang@intel.com>
---------
Signed-off-by: Wang, Yi A <yi.a.wang@intel.com>
* update transformres
* Upgrading the nix deps too.
* Forcing torchvision to be in there.
* Fixing bug in mllama.
* Those tests cannot be run in CI.
* Lint.
---------
Co-authored-by: Pedro Cuenca <pedro@huggingface.co>
Co-authored-by: Nicolas Patry <patry.nicolas@protonmail.com>
* initial changes
* Add support for other vlm
* cleanup comment
* Improve attn_implementation
* Add comments for support of models
* add model
* add model
* fixes and improvements
* update docker
* Add cache position
* Add tests
* remove redundant changes
* remove tr version
* Upgrade doc + fix linting.
* Fixing the CI.
---------
Co-authored-by: Nicolas Patry <patry.nicolas@protonmail.com>
* Publish nix docker image.
* Run during PR.
* Something else.
* Forgot to push.
* Build zstd.
* Pushing with skopeo
* Testing the PR.
* Runnign from nix.
* Cleaner tags.
* launcher: correctly get the head dimension for VLMs
For most (?) VLMs, the head dimension is in the `text_config`
configuration section. However, since we only queried the top-level
`head_dim` (which typically doesn't exist in VLMs), we would never use
flashinfer. This change adds a method that gets the head dimension from
the top-level `Config` struct or `text_config` when that fails.
* fix: bump org name in gemma3 test
---------
Co-authored-by: drbh <david.richard.holtz@gmail.com>
* xpu 2.6 update
Signed-off-by: Wang, Yi A <yi.a.wang@intel.com>
* install whl
Signed-off-by: Wang, Yi A <yi.a.wang@intel.com>
* update get xpu memory api
Signed-off-by: Wang, Yi A <yi.a.wang@intel.com>
* int
Signed-off-by: Wang, Yi A <yi.a.wang@intel.com>
* fix awq crash if modules_to_not_convert is None
Signed-off-by: Wang, Yi A <yi.a.wang@intel.com>
---------
Signed-off-by: Wang, Yi A <yi.a.wang@intel.com>
On NixOS, the CUDA driver shim gets mounted on /run/opengl-driver,
where Nix packages expect the shim to be. However, on other
distributions, some FHS paths are mounted. This is a small change
to make the dynamic loader find the shim.
* feat(gaudi): release ready (docs, docker image and vlm ready)
* fix(gaudi): add default argument for the dockerfile
* fix(gaudi): remove use of latest for gaudi docker image + redid gaudi benchmarking section to include best practices
* Update to `kernels` 0.2.1
The package was renamed from `hf-kernels` to `kernels`. The new version
also updates the lockfile format.
* Download kernels in `install-cuda` target
* feat(neuron): use AWS Neuron SDK 2.21.1
* feat(neuron): bump optimum-neuron version
* feat(neuron): tag latest image for local tests
* test(neuron): simplify sampling test
* Fixing the tool calling convention.
* Update tehe doc.
* Fixing some corner cases.
* Fixing the tool call id.
* Fmt.
* Snapshot update with the new updated tool_call_id.
* More qwen2.
* change ChatCompletionChunk to align with "OpenAI Chat Completions streaming API"
Moving after tool_calls2
Signed-off-by: Wang, Yi A <yi.a.wang@intel.com>
add in Buffering..
Signed-off-by: Wang, Yi A <yi.a.wang@intel.com>
fix: handle usage outside of stream state and add tests
Simplifying everything quite a bit.
Remove the unused model_dump.
Clippy.
Clippy ?
Ruff.
Uppgrade the flake for latest transformers.
Upgrade after rebase.
Remove potential footgun.
Fix completion test.
* Clippy.
* Tweak for multi prompt.
* Ruff.
* Update the snapshot a bit.
---------
Co-authored-by: Nicolas Patry <patry.nicolas@protonmail.com>
