benchmarks.yml 1.4 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748
  1. name: Build and Test
  2. on:
  3. push:
  4. branches: [ '*' ]
  5. tags: [ '*' ]
  6. pull_request:
  7. branches: [ '*' ]
  8. jobs:
  9. # test-m4-cluster:
  10. # strategy:
  11. # matrix:
  12. # model: ['llama-3.2-1b', 'llama-3.2-3b']
  13. # # Optional: add fail-fast: false if you want all matrix jobs to continue even if one fails
  14. # fail-fast: false
  15. # uses: ./.github/workflows/bench_job.yml
  16. # with:
  17. # config: '{"M4PRO_GPU16_24GB": 2}'
  18. # model: ${{ matrix.model }}
  19. # calling_job_name: 'test-m4-cluster'
  20. # secrets: inherit
  21. three-m4-pro-cluster:
  22. strategy:
  23. matrix:
  24. model: ['llama-3.2-1b', 'llama-3.2-3b', 'llama-3.1-8b', 'llama-3.3-70b']
  25. # Optional: add fail-fast: false if you want all matrix jobs to continue even if one fails
  26. fail-fast: false
  27. uses: ./.github/workflows/bench_job.yml
  28. with:
  29. config: '{"M4PRO_GPU16_24GB": 3}'
  30. model: ${{ matrix.model }}
  31. calling_job_name: 'three-m4-pro-cluster'
  32. secrets: inherit
  33. # test-m3-single-node:
  34. # strategy:
  35. # matrix:
  36. # model: ['llama-3.2-1b']
  37. # # Optional: add fail-fast: false if you want all matrix jobs to continue even if one fails
  38. # fail-fast: false
  39. # uses: ./.github/workflows/bench_job.yml
  40. # with:
  41. # config: '{"M3MAX_GPU40_128GB": 1}'
  42. # model: ${{ matrix.model }}
  43. # calling_job_name: 'test-m3-cluster'
  44. # secrets: inherit