Convert hf (#225) #573
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 compile tests | |
| on: | |
| pull_request: | |
| push: | |
| branches: | |
| - main | |
| workflow_dispatch: | |
| jobs: | |
| test-cuda: | |
| uses: pytorch/test-infra/.github/workflows/linux_job.yml@main | |
| with: | |
| runner: linux.g5.4xlarge.nvidia.gpu | |
| gpu-arch-type: cuda | |
| gpu-arch-version: "12.1" | |
| script: | | |
| 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::Download checkpoints" | |
| # Install requirements | |
| pip install --pre torch torchvision torchaudio --index-url https://download.pytorch.org/whl/nightly/cu121 | |
| pip install -r requirements.txt | |
| echo "::endgroup::" | |
| echo "::group::Download checkpoints" | |
| mkdir -p checkpoints/stories15M | |
| pushd checkpoints/stories15M | |
| wget https://huggingface.co/karpathy/tinyllamas/resolve/main/stories15M.pt | |
| wget 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 cuda --checkpoint-path ${MODEL_PATH} --temperature 0 > ./output_eager | |
| cat ./output_eager | |
| python generate.py --device cuda --compile --checkpoint-path ${MODEL_PATH} --temperature 0 > ./output_compiled | |
| cat ./output_compiled | |
| python export.py --device cuda --checkpoint-path ${MODEL_PATH} --output-dso-path ${MODEL_DIR}/${MODEL_NAME}.so | |
| python generate.py --device cuda --checkpoint-path ${MODEL_PATH} --temperature 0 --dso-path ${MODEL_DIR}/${MODEL_NAME}.so > ./output_aoti | |
| cat ./output_aoti | |
| echo "******************************************" | |
| echo "******* Emb: channel-wise quantized ******" | |
| echo "******************************************" | |
| python generate.py --device cuda --quant '{"embedding" : {"bitwidth": 8, "groupsize": 0}}' --checkpoint-path ${MODEL_PATH} --temperature 0 > ./output_eager | |
| cat ./output_eager | |
| python generate.py --device cuda --compile --quant '{"embedding" : {"bitwidth": 8, "groupsize": 0}}' --checkpoint-path ${MODEL_PATH} --temperature 0 > ./output_compiled | |
| cat ./output_compiled | |
| python export.py --device cuda --quant '{"embedding" : {"bitwidth": 8, "groupsize": 0}}' --checkpoint-path ${MODEL_PATH} --output-dso-path ${MODEL_DIR}/${MODEL_NAME}.so | |
| python generate.py --device cuda --checkpoint-path ${MODEL_PATH} --temperature 0 --dso-path ${MODEL_DIR}/${MODEL_NAME}.so > ./output_aoti | |
| cat ./output_aoti | |
| echo "******************************************" | |
| echo "******** Emb: group-wise quantized *******" | |
| echo "******************************************" | |
| python generate.py --device cuda --quant '{"embedding" : {"bitwidth": 8, "groupsize": 8}}' --checkpoint-path ${MODEL_PATH} --temperature 0 > ./output_eager | |
| cat ./output_eager | |
| python generate.py --device cuda --compile --quant '{"embedding" : {"bitwidth": 8, "groupsize": 8}}' --checkpoint-path ${MODEL_PATH} --temperature 0 > ./output_compiled | |
| cat ./output_compiled | |
| python export.py --device cuda --quant '{"embedding" : {"bitwidth": 8, "groupsize": 8}}' --checkpoint-path ${MODEL_PATH} --output-dso-path ${MODEL_DIR}/${MODEL_NAME}.so | |
| python generate.py --device cuda --checkpoint-path ${MODEL_PATH} --temperature 0 --dso-path ${MODEL_DIR}/${MODEL_NAME}.so > ./output_aoti | |
| cat ./output_aoti | |
| echo "******************************************" | |
| echo "******* INT8 channel-wise quantized ******" | |
| echo "******************************************" | |
| python generate.py --device cuda --quant '{"linear:int8" : {"bitwidth": 8, "groupsize": 0}}' --checkpoint-path ${MODEL_PATH} --temperature 0 > ./output_eager | |
| cat ./output_eager | |
| python generate.py --device cuda --compile --quant '{"linear:int8" : {"bitwidth": 8, "groupsize": 0}}' --checkpoint-path ${MODEL_PATH} --temperature 0 > ./output_compiled | |
| cat ./output_compiled | |
| python export.py --device cuda --quant '{"linear:int8" : {"bitwidth": 8, "groupsize": 0}}' --checkpoint-path ${MODEL_PATH} --output-dso-path ${MODEL_DIR}/${MODEL_NAME}.so | |
| python generate.py --device cuda --checkpoint-path ${MODEL_PATH} --temperature 0 --dso-path ${MODEL_DIR}/${MODEL_NAME}.so > ./output_aoti | |
| cat ./output_aoti | |
| echo "******************************************" | |
| echo "******** INT8 group-wise quantized *******" | |
| echo "******************************************" | |
| python generate.py --device cuda --quant '{"linear:int8" : {"bitwidth": 8, "groupsize": 8}}' --checkpoint-path ${MODEL_PATH} --temperature 0 > ./output_eager | |
| cat ./output_eager | |
| python generate.py --device cuda --compile --quant '{"linear:int8" : {"bitwidth": 8, "groupsize": 8}}' --checkpoint-path ${MODEL_PATH} --temperature 0 > ./output_compiled | |
| cat ./output_compiled | |
| python export.py --device cuda --quant '{"linear:int8" : {"bitwidth": 8, "groupsize": 8}}' --checkpoint-path ${MODEL_PATH} --output-dso-path ${MODEL_DIR}/${MODEL_NAME}.so | |
| python generate.py --device cuda --checkpoint-path ${MODEL_PATH} --temperature 0 --dso-path ${MODEL_DIR}/${MODEL_NAME}.so > ./output_aoti | |
| cat ./output_aoti | |
| echo "******************************************" | |
| echo "******** INT4 group-wise quantized *******" | |
| echo "******************************************" | |
| python generate.py --device cuda --quant '{"linear:int4" : {"groupsize": 32}}' --checkpoint-path ${MODEL_PATH} --temperature 0 > ./output_eager | |
| cat ./output_eager | |
| python generate.py --device cuda --compile --quant '{"linear:int4" : {"groupsize": 32}}' --checkpoint-path ${MODEL_PATH} --temperature 0 > ./output_compiled | |
| cat ./output_compiled | |
| python export.py --device cuda --quant '{"linear:int4" : {"groupsize": 32}}' --checkpoint-path ${MODEL_PATH} --output-dso-path ${MODEL_DIR}/${MODEL_NAME}.so | |
| python generate.py --device cuda --checkpoint-path ${MODEL_PATH} --temperature 0 --dso-path ${MODEL_DIR}/${MODEL_NAME}.so > ./output_aoti | |
| cat ./output_aoti | |
| echo "tests complete" | |
| echo "******************************************" | |
| echo "::endgroup::" | |