DeepSpeed
40e6764e - Merge branch 'master' into loadams/rocm6

Comment changes are shownComment changes are hidden
Commit
1 year ago
Merge branch 'master' into loadams/rocm6
References
Author
  • .github/workflows
    • File
      cpu-inference.yml
    • File
      nv-inference.yml
    • File
      nv-sd.yml
  • accelerator
    • File
      hpu_accelerator.py
    • File
      real_accelerator.py
    • File
      xpu_accelerator.py
  • csrc
    • adam
      • File
        multi_tensor_adam.cu
    • aio/py_lib
      • File
        deepspeed_aio_thread.cpp
    • cpu/comm
      • File
        ccl.cpp
    • xpu
      • adagrad
        • File
          cpu_adagrad.cpp
      • adam
        • File
          cpu_adam.cpp
        • File
          cpu_adam_impl.cpp
        • File
          fused_adam_frontend.cpp
        • File
          multi_tensor_adam.dp.cpp
        • File
          multi_tensor_apply.dp.hpp
      • common
        • File
          custom_cuda_kernel.dp.cpp
      • includes
        • File
          compat.h
        • File
          cpu_adagrad.h
        • File
          cpu_adam.h
        • File
          simd.h
        • File
          type_shim.h
  • deepspeed
    • comm
      • File
        ccl.py
    • inference/v2
      • model_implementations
        • falcon
          • File
            model.py
        • File
          inference_transformer_base.py
        • llama_v2
          • File
            model.py
        • mistral
          • File
            model.py
        • mixtral
          • File
            model.py
        • opt
          • File
            model.py
        • phi
          • File
            model.py
      • modules/implementations/attention
        • File
          dense_blocked_attention.py
    • module_inject
      • File
        auto_tp.py
      • containers
        • File
          vae.py
      • File
        fusedqkv_utils.py
      • File
        replace_module.py
      • File
        tp_shard.py
    • runtime
      • File
        engine.py
      • File
        sparse_tensor.py
      • swap_tensor
        • File
          optimizer_utils.py
        • File
          partitioned_optimizer_swapper.py
        • File
          pipelined_optimizer_swapper.py
      • zero
        • File
          mics.py
        • File
          stage3.py
  • docs
    • _tutorials
      • File
        accelerator-abstraction-interface.md
    • code-docs/source
      • File
        schedulers.rst
  • op_builder
    • File
      builder.py
    • npu
      • File
        __init__.py
      • File
        async_io.py
    • xpu
      • File
        __init__.py
      • File
        async_io.py
      • File
        builder.py
      • File
        cpu_adagrad.py
      • File
        cpu_adam.py
      • File
        fused_adam.py
  • requirements
    • File
      requirements-inf.txt
    • File
      requirements-sd.txt
  • tests/unit
    • inference
      • File
        test_inference.py
      • File
        test_inference_config.py
    • runtime/zero
      • File
        test_nvme_checkpointing.py
      • File
        test_zero.py
      • File
        test_zero_context.py
      • File
        test_zero_offloadpp.py