Historique des commits

Auteur SHA1 Message Date
  AlpinDale 6ac658b0d6 some small performance improvements il y a 6 mois
  AlpinDale b7a2d52e47 fix: allow using mp executor for pipeline parallel il y a 6 mois
  AlpinDale cf381a0c54 OpenAI API Refactor (#591) il y a 6 mois
  AlpinDale ddb28a80a3 fix: bump torch for rocm, unify CUDA_VISIBLE_DEVICES for cuda and rocm il y a 6 mois
  AlpinDale a5fafaa9ce chore: add more tuning for the CPU backend via intel-openmp il y a 6 mois
  AlpinDale 5257ebce8c fix: device >= 0 && device < num_gpus INTERNAL_ASSERT FAILED il y a 6 mois
  AlpinDale cda0e93a10 abstract away the platform for device capability il y a 6 mois
  AlpinDale 7d79c0e726 chore: use nvml query to avoid accidental cuda initialization il y a 6 mois
  AlpinDale 0886c361f4 feat: OpenVINO CPU backend (#576) il y a 7 mois
  AlpinDale 2c321ce1f2 chore: upgrade to rocm 6.1, update docker il y a 7 mois
  AlpinDale 25feb1d592 chore: add support for pinning lora adapters in the lru cache il y a 7 mois
  AlpinDale 6a57861fca feat: initial XPU support via intel_extension_for_pytorch (#571) il y a 7 mois
  AlpinDale a89c9a0e92 fix: device ordinal issues with world_size and stuff il y a 7 mois
  AlpinDale fe21123a1c feat: TPU support (#570) il y a 7 mois
  AlpinDale 156f577f79 feat: switch from `PYBIND11_MODULE` to `TORCH_LIBRARY` (#569) il y a 7 mois
  AlpinDale b029a544ff optimize eager mode host time with numpy il y a 7 mois
  AlpinDale f2b7a42c4e fix: async cancels in merge_async_iterators for python>=3.9 il y a 7 mois
  AlpinDale 7194047318 remove vllm-nccl il y a 7 mois
  AlpinDale 90ceab32ff refactor: consolidate prompt args to LLM engines il y a 7 mois
  AlpinDale 656459fd84 make fp8_e4m3 work on nvidia il y a 7 mois
  AlpinDale 251568470e initial nvidia fp8 e4m3 for kv cache il y a 7 mois
  AlpinDale 4476d2d885 remove cuda version check il y a 7 mois
  AlpinDale 2351a0e2cd feat: FlashInfer backend for decoding phase (#548) il y a 7 mois
  AlpinDale 2656df543b why was this removed? weird il y a 7 mois
  AlpinDale 2e0b115ce1 move func tracing to utils il y a 8 mois
  AlpinDale 46159b107a formatting: pt1 il y a 8 mois
  AlpinDale fca911ee0a vLLM Upstream Sync (#526) il y a 8 mois
  AlpinDale f894f7b176 Revert "reduce dedupe by wrapping in general worker class" il y a 9 mois
  AlpinDale 9fff6fb507 reduce dedupe by wrapping in general worker class il y a 9 mois
  AlpinDale 9d81716bfd [v0.5.3] Release Candidate (#388) il y a 10 mois