knox / Llama-Deployment-1.sh
0 likes
0 forks
1 files
Last active
| 1 | sudo apt update && sudo apt upgrade -y |
knox / LLaMA-Factory-22.bib
0 likes
0 forks
1 files
Last active
| 1 | @inproceedings{zheng2024llamafactory, |
| 2 | title={LlamaFactory: Unified Efficient Fine-Tuning of 100+ Language Models}, |
| 3 | author={Yaowei Zheng and Richong Zhang and Junhao Zhang and Yanhan Ye and Zheyan Luo and Zhangchi Feng and Yongqiang Ma}, |
| 4 | booktitle={Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 3: System Demonstrations)}, |
| 5 | address={Bangkok, Thailand}, |
| 6 | publisher={Association for Computational Linguistics}, |
| 7 | year={2024}, |
| 8 | url={http://arxiv.org/abs/2403.13372} |
| 9 | } |
knox / LLaMA-Factory-21.yaml
0 likes
0 forks
1 files
Last active
| 1 | report_to: wandb |
| 2 | run_name: test_run # optional |
knox / LLaMA-Factory-20.sh
0 likes
0 forks
1 files
Last active
| 1 | export USE_OPENMIND_HUB=1 # `set USE_OPENMIND_HUB=1` for Windows |
knox / LLaMA-Factory-19.sh
0 likes
0 forks
1 files
Last active
| 1 | export USE_MODELSCOPE_HUB=1 # `set USE_MODELSCOPE_HUB=1` for Windows |
knox / LLaMA-Factory-18.sh
0 likes
0 forks
1 files
Last active
| 1 | API_PORT=8000 llamafactory-cli api examples/inference/llama3_vllm.yaml |
knox / LLaMA-Factory-17.sh
0 likes
0 forks
1 files
Last active
| 1 | docker build -f ./docker/docker-rocm/Dockerfile \ |
| 2 | --build-arg INSTALL_BNB=false \ |
| 3 | --build-arg INSTALL_VLLM=false \ |
| 4 | --build-arg INSTALL_DEEPSPEED=false \ |
| 5 | --build-arg INSTALL_FLASHATTN=false \ |
| 6 | --build-arg PIP_INDEX=https://pypi.org/simple \ |
| 7 | -t llamafactory:latest . |
| 8 | |
| 9 | docker run -dit \ |
| 10 | -v ./hf_cache:/root/.cache/huggingface \ |
knox / LLaMA-Factory-16.sh
0 likes
0 forks
1 files
Last active
| 1 | # Choose docker image upon your environment |
| 2 | docker build -f ./docker/docker-npu/Dockerfile \ |
| 3 | --build-arg INSTALL_DEEPSPEED=false \ |
| 4 | --build-arg PIP_INDEX=https://pypi.org/simple \ |
| 5 | -t llamafactory:latest . |
| 6 | |
| 7 | # Change `device` upon your resources |
| 8 | docker run -dit \ |
| 9 | -v ./hf_cache:/root/.cache/huggingface \ |
| 10 | -v ./ms_cache:/root/.cache/modelscope \ |
knox / LLaMA-Factory-15.sh
0 likes
0 forks
1 files
Last active
| 1 | docker build -f ./docker/docker-cuda/Dockerfile \ |
| 2 | --build-arg INSTALL_BNB=false \ |
| 3 | --build-arg INSTALL_VLLM=false \ |
| 4 | --build-arg INSTALL_DEEPSPEED=false \ |
| 5 | --build-arg INSTALL_FLASHATTN=false \ |
| 6 | --build-arg PIP_INDEX=https://pypi.org/simple \ |
| 7 | -t llamafactory:latest . |
| 8 | |
| 9 | docker run -dit --gpus=all \ |
| 10 | -v ./hf_cache:/root/.cache/huggingface \ |
knox / LLaMA-Factory-14.sh
0 likes
0 forks
1 files
Last active
| 1 | cd docker/docker-rocm/ |
| 2 | docker compose up -d |
| 3 | docker compose exec llamafactory bash |