Skip to content
nv-pre-compile-ops

adding 8bit dequantization kernel for asym fine-grained block quantization in zero-inference #3122

Sign in to view logs

adding 8bit dequantization kernel for asym fine-grained block quantization in zero-inference

adding 8bit dequantization kernel for asym fine-grained block quantization in zero-inference #3122

Workflow file for this run

.github/workflows/nv-pre-compile-ops.yml at 571ab83
name: nv-pre-compile-ops
on:
pull_request:
branches:
'**'
paths-ignore:
- 'docs/**'
- 'blogs/**'
merge_group:
branches: [ master ]
schedule:
- cron: "0 0 * * *"
concurrency:
group: ${{ github.workflow }}-${{ github.ref }}
cancel-in-progress: true
jobs:
build-ops:
runs-on: ubuntu-20.04
container:
image: deepspeed/gh-builder:ubuntu1804-py38-torch1131-cu116
steps:
- uses: actions/checkout@v3
- name: environment
run: |
which python
python --version
python -c "import torch; print('torch:', torch.__version__, torch)"
#python -c "import torch; print('CUDA available:', torch.cuda.is_available())"
- name: Compile DeepSpeed Ops
run: |
TORCH_CUDA_ARCH_LIST="7.0;7.5;8.0" DS_BUILD_OPS=1 DS_BUILD_SPARSE_ATTN=0 DS_BUILD_EVOFORMER_ATTN=0 pip3 install .
- name: DS Report
run: |
ds_report