Skip to content

Add support for data parallel QLoRA training via DeepSpeed Zero stages 0, 1 and 2. #11361

Add support for data parallel QLoRA training via DeepSpeed Zero stages 0, 1 and 2.

Add support for data parallel QLoRA training via DeepSpeed Zero stages 0, 1 and 2. #11361

Triggered via pull request October 17, 2023 20:37
@arnavgarg1arnavgarg1
synchronize #3728
ds-stage2
Status Cancelled
Total duration 15m 3s
Artifacts

pytest.yml

on: pull_request
LLM Tests
0s
LLM Tests
Combinatorial Tests
0s
Combinatorial Tests
Test Minimal Install
0s
Test Minimal Install
Event File
0s
Event File
Matrix: integration-tests
Matrix: pytest
Fit to window
Zoom out
Zoom in

Annotations

15 errors
integration_tests_c
Canceling since a higher priority waiting request for 'pytest-ds-stage2' exists
integration_tests_a
Canceling since a higher priority waiting request for 'pytest-ds-stage2' exists
integration_tests_d
Canceling since a higher priority waiting request for 'pytest-ds-stage2' exists
integration_tests_b
Canceling since a higher priority waiting request for 'pytest-ds-stage2' exists
integration_tests_e
Canceling since a higher priority waiting request for 'pytest-ds-stage2' exists
LLM Tests
Canceling since a higher priority waiting request for 'pytest-ds-stage2' exists
Combinatorial Tests
Canceling since a higher priority waiting request for 'pytest-ds-stage2' exists
py3.9, torch-2.0.0, not distributed, ubuntu-latest, ray 2.3.0
Canceling since a higher priority waiting request for 'pytest-ds-stage2' exists
py3.8, torch-1.13.0, distributed, ubuntu-latest, ray 2.2.0
Canceling since a higher priority waiting request for 'pytest-ds-stage2' exists
py3.10, torch-nightly, not distributed, ubuntu-latest, ray 2.3.1
Canceling since a higher priority waiting request for 'pytest-ds-stage2' exists
py3.8, torch-1.13.0, not distributed, ubuntu-latest, ray 2.2.0
Canceling since a higher priority waiting request for 'pytest-ds-stage2' exists
py3.10, torch-nightly, distributed, ubuntu-latest, ray 2.3.1
Canceling since a higher priority waiting request for 'pytest-ds-stage2' exists
Event File
Canceling since a higher priority waiting request for 'pytest-ds-stage2' exists
py3.9, torch-2.0.0, distributed, ubuntu-latest, ray 2.3.0
Canceling since a higher priority waiting request for 'pytest-ds-stage2' exists
Test Minimal Install
Canceling since a higher priority waiting request for 'pytest-ds-stage2' exists