Update cli.py with inline comments #416
Workflow file for this run
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| name: Run eager tests on MPS | |
| on: | |
| pull_request: | |
| push: | |
| branches: | |
| - main | |
| workflow_dispatch: | |
| jobs: | |
| test-mps: | |
| uses: pytorch/test-infra/.github/workflows/macos_job.yml@main | |
| with: | |
| runner: macos-m1-stable | |
| script: | | |
| set -eou pipefail | |
| echo "::group::Print machine info" | |
| uname -a | |
| if [ $(uname -s) == Darwin ]; then | |
| sysctl machdep.cpu.brand_string | |
| sysctl machdep.cpu.core_count | |
| fi | |
| echo "::endgroup::" | |
| echo "::group::Install requirements" | |
| # Install requirements | |
| pip install --pre torch torchvision torchaudio --index-url https://download.pytorch.org/whl/nightly/cpu | |
| ls -la | |
| pwd | |
| pip install -r requirements.txt | |
| echo "::endgroup::" | |
| echo "::group::Download checkpoints" | |
| ( | |
| mkdir -p checkpoints/stories15M | |
| pushd checkpoints/stories15M | |
| curl -fsSL -O https://huggingface.co/karpathy/tinyllamas/resolve/main/stories15M.pt | |
| curl -fsSL -O https://github.com/karpathy/llama2.c/raw/master/tokenizer.model | |
| popd | |
| ) | |
| echo "::endgroup::" | |
| echo "::group::Run inference" | |
| export MODEL_PATH=checkpoints/stories15M/stories15M.pt | |
| export MODEL_NAME=stories15M | |
| export MODEL_DIR=/tmp | |
| python generate.py --device mps --checkpoint-path ${MODEL_PATH} --temperature 0 > ./output_eager | |
| cat ./output_eager | |
| echo "************************************************************" | |
| echo "*** embedding" | |
| echo "************************************************************" | |
| python generate.py --device mps --quant '{"embedding" : {"bitwidth": 8, "groupsize": 0}}' --checkpoint-path ${MODEL_PATH} --temperature 0 > ./output_eager | |
| cat ./output_eager | |
| python generate.py --device mps --quant '{"embedding" : {"bitwidth": 8, "groupsize": 8}}' --checkpoint-path ${MODEL_PATH} --temperature 0 > ./output_eager | |
| cat ./output_eager | |
| echo "************************************************************" | |
| echo "*** linear int8" | |
| echo "************************************************************" | |
| python generate.py --device mps --quant '{"linear:int8" : {"bitwidth": 8, "groupsize": 0}}' --checkpoint-path ${MODEL_PATH} --temperature 0 > ./output_eager | |
| cat ./output_eager | |
| python generate.py --device mps --quant '{"linear:int8" : {"bitwidth": 8, "groupsize": 8}}' --checkpoint-path ${MODEL_PATH} --temperature 0 > ./output_eager | |
| cat ./output_eager | |
| echo "************************************************************" | |
| echo "*** linear int4" | |
| echo "************************************************************" | |
| PYTORCH_ENABLE_MPS_FALLBACK=1 python generate.py --device mps --quant '{"linear:int4" : {"groupsize": 32}}' --checkpoint-path ${MODEL_PATH} --temperature 0 > ./output_eager | |
| cat ./output_eager | |