ray/release/sgd_tests/sgd_tests.yaml

# Test multi-node, multi-GPU Ray SGD example.
- name: sgd_gpu
  team: ml
  cluster:
    app_config: sgd_gpu/sgd_gpu_app_config.yaml
    compute_template: sgd_gpu/sgd_gpu_compute.yaml

  run:
    timeout: 3000
    prepare: python wait_cluster.py 2 600
    script: python sgd_gpu/sgd_gpu_test.py --num-workers=2 --use-gpu --address=auto
[Release] Add multi-node, multi-GPU SGD release test (#16046) 2021-05-31 16:23:04 -07:00			`# Test multi-node, multi-GPU Ray SGD example.`
			`- name: sgd_gpu`
[Test Infra] Unrevert team col (#21700) This fixes the previous problems from team column revert. This has 2 additional changes; alert handler receives the team argument, which was the root cause of breakage; https://github.com/ray-project/ray/pull/21289 Previously, tests without a team column were raising an exception, but I made the condition weaker (warning logs). I will eventually change it to raise an exception, but for smoother transition, we will log warning instead for a short time 2022-01-20 06:29:53 +09:00			`team: ml`
[Release] Add multi-node, multi-GPU SGD release test (#16046) 2021-05-31 16:23:04 -07:00			`cluster:`
[SGD] Add support for native Torch AMP in SGD (#16382) * SGD native AMP initial commit * SGD native amp second pass * Update docs * Update TorchTrainer doc * Temp fix release test * Update release/sgd_tests/sgd_gpu/sgd_gpu_app_config.yaml Co-authored-by: Amog Kamsetty <amogkam@users.noreply.github.com> 2021-06-16 02:48:21 +02:00			`app_config: sgd_gpu/sgd_gpu_app_config.yaml`
[Release] Add multi-node, multi-GPU SGD release test (#16046) 2021-05-31 16:23:04 -07:00			`compute_template: sgd_gpu/sgd_gpu_compute.yaml`

			`run:`
			`timeout: 3000`
			`prepare: python wait_cluster.py 2 600`
			`script: python sgd_gpu/sgd_gpu_test.py --num-workers=2 --use-gpu --address=auto`