huggingface/optimum-amd

Issues

• can't build rocm enabled onnx with podman/docker

  #41 opened 10 months ago by Tokariew
  2

• Ability to patch transformers / diffusers with FlashAttention for AMD

  #24 opened a year ago by mfuntowicz
  1

• [VERNE] AMD Instinct MI250 setup

  #1 opened a year ago by mfuntowicz
  1

• [VERNE] AMD Instinct MI250 servers racked up

  #2 opened a year ago by mfuntowicz
  1

• [VERNE] AMD Instinct MI250 servers online

  #3 opened a year ago by mfuntowicz
  1

• [INSTINCT] Slides for model optimizations

  #9 opened a year ago by mfuntowicz
  1

• AMD FlashAttention v1 Integration

  #19 opened a year ago by mfuntowicz
  3

• TGI integration for AMD Instinct GPUs

  #21 opened a year ago by mfuntowicz
  1

• Make the 🤗 documentation more GPU vendor agnostic

  #23 opened a year ago by mfuntowicz
  1

• Need DeepSpeed patch release for RoCM 5.6/5.7

  #18 opened a year ago by mfuntowicz
  0

• Securing Google Dana deployment internally

  #22 opened a year ago by mfuntowicz
  0

• Ensure DeepSpeed tests are running

  #17 opened a year ago by mfuntowicz
  3

• Integrate `optimum-benchmarks` to validate model performances

  #15 opened a year ago by mfuntowicz
  0

• [INSTINCT] Move all the `*-amdgpu` workflow inside it's own workflow YAML file

  #4 opened a year ago by mfuntowicz
  0

• [INSTINCT] Enable transformers CI to run nightly

  #10 opened a year ago by mfuntowicz
  0

• [RYZEN AI] Ensure Whisper can be exported to ONNX IR

  #5 opened a year ago by mfuntowicz
  1

• [RYZEN AI] Ensure Starcoder can be exported to ONNX IR

  #8 opened a year ago by mfuntowicz
  0

• [RYZEN AI] Ensure Llama2 can be exported to ONNX IR

  #7 opened a year ago by mfuntowicz
  0

• [RYZEN AI] Ensure Stable Diffusion can be exported to ONNX IR

  #6 opened a year ago by mfuntowicz
  0