Reduce the memory usage of logits from O(context_length) to O(1) #4688
Dr.CI classification results
{"FAILED":[{"workflowId":10517220132,"workflowUniqueId":71492439,"id":29140915340,"runnerName":"i-014044f03b0f32206","authorEmail":"myuan@meta.com","name":"trunk / test-custom-ops-macos (cmake) / macos-job","jobName":"test-custom-ops-macos (cmake) / macos-job","conclusion":"failure","completed_at":"2024-08-23T00:18:54Z","html_url":"https://github.com/pytorch/executorch/actions/runs/10517220132/job/29140915340","head_branch":"ciflow/trunk/4688","pr_number":4688,"head_sha":"be438eb1f28c893f5e814d4402cf539966448af8","head_sha_timestamp":"2024-08-22T16:34:12-07:00","failure_captures":["##[error]The operation was canceled."],"failure_lines":["##[error]The operation was canceled."],"failure_context":["+ pip install --progress-bar off -r requirements-ci.txt","+ pushd .ci/docker","+ install_pip_dependencies","+ popd","+ rm buck2-aarch64-apple-darwin-2024-05-15.zst","+ mv buck2 /opt/homebrew/bin","+ chmod +x buck2","+ zstd -d buck2-aarch64-apple-darwin-2024-05-15.zst -o buck2","+ curl -s https://ossci-macos.s3.amazonaws.com/buck2-aarch64-apple-darwin-2024-05-15.zst -o buck2-aarch64-apple-darwin-2024-05-15.zst","+ BUCK2=buck2-aarch64-apple-darwin-2024-05-15.zst","+ BUCK2_VERSION=2024-05-15","++ cat ci_commit_pins/buck2.txt"],"time":"2024-08-23T00:19:05.754622Z"}],"FLAKY":[],"BROKEN_TRUNK":[],"UNSTABLE":[]}