Code

Add an HLO backend for LLM models#775

Code

Merged

dacorvo merged 9 commits intomainhuggingface/optimum-neuron:mainfrom add_hlo_backendhuggingface/optimum-neuron:add_hlo_backendCopy head branch name to clipboardFeb 6, 2025

+7,151-1,394

Commits

Commits on Feb 4, 2025

• fix(examples): use max_new_tokens for generation

  dacorvo

  committed
  1eabddd

  Copy full SHA for 1eabddd

• feat: add new HLO backend

  Show description for 53e8a2d
  dacorvo

  committed
  53e8a2d

  Copy full SHA for 53e8a2d

• chore: bump dev version

  Show description for e018422
  dacorvo

  committed
  e018422

  Copy full SHA for e018422

• refactor(exporter): rename NeuronConfig

  Show description for 0cc691f
  dacorvo

  committed
  0cc691f

  Copy full SHA for 0cc691f

• refactor(models): use new HLO backend

  dacorvo

  committed
  4c624fc

  Copy full SHA for 4c624fc

Commits on Feb 5, 2025

• ci: isolate LLM tests

  Show description for 1128d75
  dacorvo

  committed
  1128d75

  Copy full SHA for 1128d75

• ci: bump upload-artifact action

  dacorvo

  committed
  c510221

  Copy full SHA for c510221

Commits on Feb 6, 2025

• test: do not use staging for cache tests

  Show description for 830c553
  dacorvo

  committed
  830c553

  Copy full SHA for 830c553

• review: avoid error if transformers-neuronx is not installed

  dacorvo

  committed
  9299f94

  Copy full SHA for 9299f94