Revert "[V1] [Bugfix] eagle bugfix and enable correct lm_head for multimodal (#18034)" (#18600)

Cyrus Leung committed over 1 year ago

Migrate docs from Sphinx to MkDocs (#18145)

Harry Mellor committed over 1 year ago

[Bugfix] Add half type support in reshape_and_cache_cpu_impl on x86 cpu platform (#18430)

Yuqi Zhang committed over 1 year ago

[Feature][V1]: suupports cached_tokens in response usage (#18149)

Chauncey committed over 1 year ago

[Doc] Update quickstart and install for cu128 using `--torch-backend=auto` (#18505)

Michael Goin committed over 1 year ago

[Feature]Add async tensor parallelism using compilation pass (#17882)

cascade committed over 1 year ago

[V1] [Bugfix] eagle bugfix and enable correct lm_head for multimodal (#18034)

RonaldBXu committed over 1 year ago

[Hardware] correct method signatures for HPU,ROCm,XPU (#18551)

Ning Xie committed over 1 year ago

[Misc] Replace `cuda` hard code with `current_platform` (#16983)

Shanshan Shen committed over 1 year ago

[Bugfix][Benchmarks] Fix a benchmark of deepspeed-mii backend to use api_key (#17291)

Teruaki Ishizaki committed over 1 year ago

[Neuron] Remove bypass on EAGLEConfig and add a test (#18514)

aws-elaineyz committed over 1 year ago

[Spec Decode] Make EAGLE3 draft token ID mapping optional (#18488)

Benjamin Chislett committed over 1 year ago

[Bugfix] Set `KVTransferConfig.engine_id` in post_init (#18576)

lkchen committed over 1 year ago

[BugFix] Increase TP execute_model timeout (#18558)

Nick Hill committed over 1 year ago

[Misc] refactor: simplify input validation and num_requests handling in _convert_v1_inputs (#18482)

CYJiang committed over 1 year ago

Enable hybrid attention models for Transformers backend (#18494)

Harry Mellor committed over 1 year ago

[V1][Spec Decoding] Use model_loader.get_model() to load models (#18273)

Mark McLoughlin committed over 1 year ago

Re-submit: Fix: Proper RGBA -> RGB conversion for PIL images. (#18569)

Chenheli Hua committed over 1 year ago

[AMD] [P/D] Compute num gpus for ROCm correctly in run_accuracy_test.sh (#18568)

rasmith committed over 1 year ago

[Frontend] [Core] Add Tensorizer support for V1, LoRA adapter serialization and deserialization (#17926)

Sanger Steel committed over 1 year ago