ms-swift 2.4.2.post2

Swift: Scalable lightWeight Infrastructure for Fine-Tuning

SWIFT (Scalable lightWeight Infrastructure for Fine-Tuning)

ModelScope Community Website
中文   ｜   English  

📖 Table of Contents

• Introduction
• Groups
• News
• Installation
• Getting Started
• Classroom
• License
• Citation

📝 Introduction

SWIFT supports training(PreTraining/Fine-tuning/RLHF), inference, evaluation and deployment of 350+ LLMs and 90+ MLLMs (multimodal large models). Developers can directly apply our framework to their own research and production environments to realize the complete workflow from model training and evaluation to application. In addition to supporting the lightweight training solutions provided by PEFT, we also provide a complete Adapters library to support the latest training techniques such as NEFTune, LoRA+, LLaMA-PRO, etc. This adapter library can be used directly in your own custom workflow without our training scripts.

To facilitate use by users unfamiliar with deep learning, we provide a Gradio web-ui for controlling training and inference, as well as accompanying deep learning courses and best practices for beginners. SWIFT web-ui is available both on Huggingface space and ModelScope studio, please feel free to try!

SWIFT has rich documentations for users, please feel free to check our documentation website:

Paper   ｜ English Documentation   ｜   中文文档  

☎ Groups

You can contact us and communicate with us by adding our group:

Discord Group	微信群

🎉 News

• 2024.09.23: Support for training and deploying pixtral-12b. Experience it using swift infer --model_type pixtral-12b --dtype fp16.
• 🔥2024.09.19: Supports the qwen2.5, qwen2.5-math, and qwen2.5-coder series models. Supports the qwen2-vl-72b series models. Best practices can be found here.
• 2024.09.07: Support the Reflection-llama3-70b model, use by swift sft/infer --model_type reflection-llama_3_1-70b.
• 2024.09.06: Support fine-tuning and inference for mplug-owl3. Best practices can be found here.
• 2024.09.05: Support for the minicpm3-4b model. Experience it using swift infer --model_type minicpm3-4b.
• 2024.09.05: Support for the yi-coder series models. Experience it using swift infer --model_type yi-coder-1_5b-chat.
• 🔥2024.08.30: Support for inference and fine-tuning of the qwen2-vl series models: qwen2-vl-2b-instruct, qwen2-vl-7b-instruct. The best practices can be found here.
• 🔥2024.08.26: Support Liger, which supports models like LLaMA、Qwen、Mistral and so on, and reduce memory usage by 10%~60%, use --use_liger true to begin.
• 🔥2024.08.22: Support reft tuner from ReFT to achieve 15×–65× more parameter-efficient than LoRA, use --sft_type reft to begin!
• 🔥2024.08.21: Support for phi3_5-mini-instruct, phi3_5-moe-instruct, and phi3_5-vision-instruct. The best practices for fine-tuning Latex OCR using phi3_5-vision-instruct can be found here.
• 2024.08.21: Support for idefics3-8b-llama3, llava-onevision-qwen2-0_5b-ov, llava-onevision-qwen2-7b-ov, and llava-onevision-qwen2-72b-ov.
• 🔥2024.08.20: Support fine-tuning of multimodal large models using DeepSpeed-Zero3.
• 2024.08.20: Supported models: longwriter-glm4-9b, longwriter-llama3_1-8b. Supported dataset: longwriter-6k.
• 🔥2024.08.12: 🎉 SWIFT paper has been published to arXiv. Check this link to read.
• 🔥2024.08.12: Support packing with flash-attention without the contamination of attention_mask, use --packing to begin. CheckPR.
• 🔥2024.08.09: Support for inference and fine-tuning of the qwen2-audio model. Best practice can be found here.
• 🔥2024.08.08: Supports the qwen2-math series models: 1.5B, 7B, 72B. Use swift infer --model_type qwen2-math-1_5b-instruct for an experience.
• 🔥2024.08.07: Support for using vLLM for accelerating inference and deployment of multimodal large models such as the llava series and phi3-vision models. You can refer to the Multimodal & vLLM Inference Acceleration Documentation for more information.
• 2024.08.06: Support for minicpm-v-v2_6-chat is available. You can use swift infer --model_type minicpm-v-v2_6-chat for inference experience. Best practices can be found here.
• 2024.08.06: Supports internlm2.5 series of 1.8b and 20b. Experience it using swift infer --model_type internlm2_5-1_8b-chat.
• 🔥2024.08.05: Support evaluation for multi-modal models! Same command with new datasets.
• 🔥2024.08.02: Support Fourier Ft. Use --sft_type fourierft to begin, Check parameter documentation here.
• 🔥2024.07.29: Support the use of lmdeploy for inference acceleration of LLM and VLM models. Documentation can be found here.
• 🔥2024.07.24: Support DPO/ORPO/SimPO/CPO alignment algorithm for vision MLLM, training scripts can be find in Document. support RLAIF-V dataset.
• 🔥2024.07.24: Support using Megatron for CPT and SFT on the Qwen2 series. You can refer to the Megatron training documentation.
• 🔥2024.07.24: Support for the llama3.1 series models, including 8b, 70b, and 405b. Support for openbuddy-llama3_1-8b-chat.

More

• 2024.07.20: Support mistral-nemo series models. Use --model_type mistral-nemo-base-2407 and --model_type mistral-nemo-instruct-2407 to begin.

• 2024.07.19: Support Q-Galore, this algorithm can reduce the training memory cost by 60% (qwen-7b-chat, full, 80G -> 35G), use swift sft --model_type xxx --use_galore true --galore_quantization true to begin!

• 2024.07.17: Support newly released InternVL2 models: model_type are internvl2-1b, internvl2-40b, internvl2-llama3-76b. For best practices, refer to here.

• 2024.07.17: Support the training and inference of NuminaMath-7B-TIR. Use with model_type numina-math-7b.

• 🔥2024.07.16: Support exporting for ollama and bitsandbytes. Use swift export --model_type xxx --to_ollama true or swift export --model_type xxx --quant_method bnb --quant_bits 4

• 2024.07.08: Support cogvlm2-video-13b-chat. You can check the best practice here.

• 2024.07.08: Support internlm-xcomposer2_5-7b-chat. You can check the best practice here.

• 🔥2024.07.06: Support for the llava-next-video series models: llava-next-video-7b-instruct, llava-next-video-7b-32k-instruct, llava-next-video-7b-dpo-instruct, llava-next-video-34b-instruct. You can refer to llava-video best practice for more information.

• 🔥2024.07.06: Support InternVL2 series: internvl2-2b, internvl2-4b, internvl2-8b, internvl2-26b.

• 2024.07.06: Support codegeex4-9b-chat.

• 2024.07.04: Support internlm2_5-7b series: internlm2_5-7b, internlm2_5-7b-chat, internlm2_5-7b-chat-1m.

• 2024.07.02: Support for llava1_6-vicuna-7b-instruct, llava1_6-vicuna-13b-instruct and other llava-hf models. For best practices, refer to here.

• 🔥2024.06.29: Support eval-scope&open-compass for evaluation! Now we have supported over 50 eval datasets like BoolQ, ocnli, humaneval, math, ceval, mmlu, gsk8k, ARC_e, please check our Eval Doc to begin! Next sprint we will support Multi-modal and Agent evaluation, remember to follow us : )

• 🔥2024.06.28: Support for Florence series model! See document

• 🔥2024.06.28: Support for Gemma2 series models: gemma2-9b, gemma2-9b-instruct, gemma2-27b, gemma2-27b-instruct.

• 🔥2024.06.18: Supports DeepSeek-Coder-v2 series model! Use model_type deepseek-coder-v2-instruct and deepseek-coder-v2-lite-instruct to begin.

• 🔥2024.06.16: Supports KTO and CPO training! See document to start training!

• 2024.06.11: Support for tool-calling agent deployment that conform to the OpenAI interface.You can refer to Agent deployment best practice

• 🔥2024.06.07: Support Qwen2 series LLM, including Base and Instruct models of 0.5B, 1.5B, 7B, and 72B, as well as corresponding quantized versions gptq-int4, gptq-int8, and awq-int4. The best practice for self-cognition fine-tuning, inference and deployment of Qwen2-72B-Instruct using dual-card 80GiB A100 can be found here.

• 🔥2024.06.05: Support for glm4 series LLM and glm4v-9b-chat MLLM. You can refer to glm4v best practice.

• 🔥2024.06.01: Supports SimPO training! See document to start training!

• 🔥2024.06.01: Support for deploying large multimodal models, please refer to the Multimodal Deployment Documentation for more information.

• 2024.05.31: Supports Mini-Internvl model, Use model_type mini-internvl-chat-2b-v1_5 and mini-internvl-chat-4b-v1_5to train.

• 2024.05.24: Supports Phi3-vision model, Use model_type phi3-vision-128k-instruct to train.

• 2024.05.22: Supports DeepSeek-V2-Lite series models, model_type are deepseek-v2-lite and deepseek-v2-lite-chat

• 2024.05.22: Supports TeleChat-12B-v2 model with quantized version, model_type are telechat-12b-v2 and telechat-12b-v2-gptq-int4

• 🔥2024.05.21: Inference and fine-tuning support for MiniCPM-Llama3-V-2_5 are now available. For more details, please refer to minicpm-v-2.5 Best Practice.

• 🔥2024.05.20: Support for inferencing and fine-tuning cogvlm2-llama3-chinese-chat-19B, cogvlm2-llama3-chat-19B. you can refer to cogvlm2 Best Practice.

• 🔥2024.05.17: Support peft=0.11.0. Meanwhile support 3 new tuners: BOFT, Vera and Pissa. use --sft_type boft/vera to use BOFT or Vera, use --init_lora_weights pissa with --sft_type lora to use Pissa.

• 2024.05.16: Supports Llava-Next (Stronger) series models. For best practice, you can refer to here.

• 🔥2024.05.13: Support Yi-1.5 series models，use --model_type yi-1_5-9b-chat to begin!

• 2024.05.11: Support for qlora training and quantized inference using hqq and eetq. For more information, see the LLM Quantization Documentation.

• 2024.05.10: Support split a sequence to multiple GPUs to reduce memory usage. Use this feature by pip install .[seq_parallel], then add --sequence_parallel_size n to your DDP script to begin!

• 2024.05.08: Support DeepSeek-V2-Chat model, you can refer to this script.Support InternVL-Chat-V1.5-Int8 model, for best practice, you can refer to here.

• 🔥2024.05.07: Supoprts ORPO training! See document to start training!

• 2024.05.07: Supports Llava-Llama3 model from xtuner，model_type is llava-llama-3-8b-v1_1.

• 2024.04.29: Supports inference and fine-tuning of InternVL-Chat-V1.5 model. For best practice, you can refer to here.

• 🔥2024.04.26: Support LISA and unsloth training! Specify --lisa_activated_layers=2 to use LISA(to reduce the memory cost to 30 percent!), specify --tuner_backend unsloth to use unsloth to train a huge model(full or lora) with lesser memory(30 percent or lesser) and faster speed(5x)!

• 🔥2024.04.26: Support the fine-tuning and inference of Qwen1.5-110B and Qwen1.5-110B-Chat model, use this script to start training!

• 2024.04.24: Support for inference and fine-tuning of Phi3 series models. Including: phi3-4b-4k-instruct, phi3-4b-128k-instruct.

• 2024.04.22: Support for inference, fine-tuning, and deployment of chinese-llama-alpaca-2 series models. This includes：chinese-llama-2-1.3b, chinese-llama-2-7b, chinese-llama-2-13b, chinese-alpaca-2-1.3b, chinese-alpaca-2-7b and chinese-alpaca-2-13b along with their corresponding 16k and 64k long text versions.

• 2024.04.22: Support for inference and fine-tuning of Llama3 GPTQ-Int4, GPTQ-Int8, and AWQ series models. Support for inference and fine-tuning of chatglm3-6b-128k, Openbuddy-Llama3.

• 2024.04.20: Support for inference, fine-tuning, and deployment of Atom series models. This includes: Atom-7B and Atom-7B-Chat. use this script to train.

• 2024.04.19: Support for single-card, DDP, ZeRO2, and ZeRO3 training and inference with NPU, please refer to NPU Inference and Fine-tuning Best Practice.

• 2024.04.19: Support for inference, fine-tuning, and deployment of Llama3 series models. This includes: Llama-3-8B, Llama-3-8B-Instruct, Llama-3-70B, and Llama-3-70B-Instruct. use this script to train.

• 2024.04.18: Supported models: wizardlm2-7b-awq, wizardlm2-8x22b, yi-6b-chat-awq, yi-6b-chat-int8, yi-34b-chat-awq, yi-34b-chat-int8. Supported --deepspeed zero3-offload and provided default zero3-offload configuration file for zero3+cpu offload usage.

• 2024.04.18: Supported compatibility with HuggingFace ecosystem using the environment variable USE_HF, switching to use models and datasets from HF. Please refer to the HuggingFace ecosystem compatibility documentation.

• 2024.04.17: Support the evaluation for OpenAI standard interfaces. Check the parameter documentation for details.

• 🔥2024.04.17: Support CodeQwen1.5-7B series: CodeQwen1.5-7B, CodeQwen1.5-7B-Chat,CodeQwen1.5-7B-Chat-AWQ, use this script to train.

• 2024.04.16: Supports inference and fine-tuning of llava-v1.6-34b model. For best practice, you can refer to here.

• 2024.04.13: Support the fine-tuning and inference of Mixtral-8x22B-v0.1 model, use this script to start training!

• 2024.04.13: Support the newly launched MiniCPM series: MiniCPM-V-2.0、MiniCPM-2B-128k、MiniCPM-MoE-8x2B and MiniCPM-1B.use this script to start training!

• 🔥2024.04.11: Support Model Evaluation with MMLU/ARC/CEval datasets(also user custom eval datasets) with one command! Check this documentation for details. Meanwhile, we support a trick way to do multiple ablation experiments, check this documentation to use.

• 🔥2024.04.11: Support c4ai-command-r series: c4ai-command-r-plus, c4ai-command-r-v01, use this script to train.

• 2024.04.10: Use SWIFT to fine-tune the qwen-7b-chat model to enhance its function call capabilities, and combine it with Modelscope-Agent for best practices, which can be found here.

• 🔥2024.04.09: Support ruozhiba dataset. Search ruozhiba in this documentation to begin training!

• 2024.04.08: Support the fine-tuning and inference of XVERSE-MoE-A4.2B model, use this script to start training!

• 2024.04.04: Support QLoRA+FSDP to train a 70B model with two 24G memory GPUs, use this script to train.

• 🔥2024.04.03: Support Qwen1.5-32B series: Qwen1.5-32B, Qwen1.5-32B-Chat, Qwen1.5-32B-Chat-GPTQ-Int4.use this script to start training!

• 🔥2024.04.02: Support the fine-tuning and inference of Mengzi3-13B-Base model, use this script to start training!

• 🔥2024.04.01: Support dbrx series: dbrx-base and dbrx-instruct, use this script to start training!

• 🔥2024.03.29: Support Qwen1.5-MoE series: Qwen1.5-MoE-A2.7B, Qwen1.5-MoE-A2.7B-Chat, Qwen1.5-MoE-A2.7B-Chat-GPTQ-Int4.

• 🔥2024.03.29: Support the fine-tuning and inference of Grok-1 300B MoE, please view details here.

• 🔥2024.03.25: Supports inference and fine-tuning of TeleChat-7b and TeleChat-12b model, use this script to start training!

• 🔥2024.03.20: Supports inference and fine-tuning for the llava series. For best practice, you can refer to here.

• 🔥2024.03.12: Support inference and fine-tuning for deepseek-vl series. Best practices can be found here.

• 🔥2024.03.11: Support GaLore for effectively reducing memory usage to 1/2 of the original in full-parameter training.

• 🔥2024.03.10: End-to-end best practices from fine-tuning to deployment for Qwen1.5-7B-Chat and Qwen1.5-72B-Chat.

• 🔥2024.03.09: Support training and inference of MAMBA model, use this script to start training!

• 2024.03.09: Support training and inference of AQLM quantized model, use this script to start training!

• 2024.03.06: Support training and inference of AWQ quantized model, use this Qwen1.5-AWQ model script to start training, and support training and inference of yi-9b.

• 🔥2024.02.29: Support LLaMA PRO, simply use this script to start training.

• 🔥2024.02.29: Support LoRA+, simply use this script to start training.

• 2024.02.25: Support swift export to quantize models using AWQ/GPTQ and push to ModelScope Hub. See documentation: LLM Quantization.

• 2024.02.22: Support gemma series: gemma-2b, gemma-2b-instruct, gemma-7b, gemma-7b-instruct.

• 2024.02.16: Support deepseek-math series: deepseek-math-7b, deepseek-math-7b-instruct, deepseek-math-7b-chat.

• 🔥2024.02.05: Support Qwen1.5 series models, see model list for all supported Qwen1.5 models. Provide fine-tuning scripts for qwen1half-7b-chat, qwen1half-7b-chat-int8.

• 2024.02.05: Support training of diffusion models such as SDXL, SD, ControlNet, as well as DreamBooth training. See corresponding training scripts for details.

• 2024.02.01: Support minicpm series: minicpm-2b-sft-chat, minicpm-2b-chat.

• 🔥2024.02.01: Support dataset mixing to reduce catastrophic forgetting. Use --train_dataset_mix_ratio 2.0 to enable training! We also open sourced the general knowledge dataset ms-bench.

• 🔥2024.02.01: Support Agent training! Agent training algorithm is derived from this paper. We also added ms-agent, a high-quality agent dataset. Use this script to start Agent training!

• 🔥2024.02.01: Support adding SFT loss in DPO training to reduce repetitive generation caused by KL divergence loss.

• 2024.02.01: Support using AdaLoRA and IA3 adapters in training.

• 2024.02.01: Support --merge_lora parameter in AnimateDiff training.

• 2024.01.30: Support internlm-xcomposer2-7b-chat.

• 🔥2024.01.30: Support ZeRO-3, simply specify --deepspeed default-zero3.

• 2024.01.29: Support internlm2-math series: internlm2-math-7b, internlm2-math-7b-chat, internlm2-math-20b, internlm2-math-20b-chat.

• 🔥2024.01.26: Support yi-vl-6b-chat, yi-vl-34b-chat.

• 2024.01.24: Support codefuse-codegeex2-6b-chat, codefuse-qwen-14b-chat.

• 2024.01.23: Support orion series: orion-14b, orion-14b-chat.

• 2024.01.20: Support xverse-13b-256k, xverse-65b-v2, xverse-65b-chat.

• 🔥2024.01.17: Support internlm2 series: internlm2-7b-base, internlm2-7b, internlm2-7b-sft-chat, internlm2-7b-chat, internlm2-20b-base, internlm2-20b, internlm2-20b-sft-chat, internlm2-20b-chat.

• 2024.01.15: Support yuan series: yuan2-2b-instruct, yuan2-2b-janus-instruct, yuan2-51b-instruct, yuan2-102b-instruct.

• 🔥2024.01.12: Support deepseek-moe series: deepseek-moe-16b, deepseek-moe-16b-chat.

• 🔥2024.01.04: Support VLLM deployment, compatible with OpenAI API style, see VLLM Inference Acceleration and Deployment for details.

• 2024.01.04: Update Benchmark for convenient viewing of training speed and memory usage of different models.

• 🔥2023.12.29: Support web-ui for sft training and inference, use swift web-ui after installing ms-swift to start.

• 🔥2023.12.29: Support DPO RLHF (Reinforcement Learning from Human Feedback) and three datasets for this task: AI-ModelScope/stack-exchange-paired, AI-ModelScope/hh-rlhf and AI-ModelScope/hh_rlhf_cn. See documentation to start training!

• 🔥2023.12.28: Support SCEdit! This tuner can significantly reduce memory usage in U-Net and support low-memory controllable image generation (replacing ControlNet), read the section below to learn more.

• 2023.12.23: Support codegeex2-6b.

• 2023.12.19: Support phi2-3b.

• 2023.12.18: Support VLLM for inference acceleration.

• 2023.12.15: Support deepseek, deepseek-coder series: deepseek-7b, deepseek-7b-chat, deepseek-67b, deepseek-67b-chat, openbuddy-deepseek-67b-chat, deepseek-coder-1_3b, deepseek-coder-1_3b-instruct, deepseek-coder-6_7b, deepseek-coder-6_7b-instruct, deepseek-coder-33b, deepseek-coder-33b-instruct.

• 2023.12.13: Support mistral-7b-instruct-v2, mixtral-moe-7b, mixtral-moe-7b-instruct.

• 2023.12.09: Support freeze_parameters_ratio parameter as a compromise between lora and full-parameter training. Corresponding sh can be found in full_freeze_ddp. Support disable_tqdm, lazy_tokenize, preprocess_num_proc parameters, see command line arguments for details.

• 2023.12.08: Support sus-34b-chat, support yi-6b-200k, yi-34b-200k.

• 2023.12.07: Support Multi-Node DDP training.

• 2023.12.05: Support models: zephyr-7b-beta-chat, openbuddy-zephyr-7b-chat. Support datasets: hc3-zh, hc3-en.

• 🔥2023.12.02: Self-cognition fine-tuning best practices, 10 minutes to fine-tune a large model for self-cognition, create your own unique large model.

• 🔥2023.11.30: Support training and inference of qwen-1_8b, qwen-72b, qwen-audio series models. Corresponding sh scripts can be found in qwen_1_8b_chat, qwen_72b_chat, qwen_audio_chat

• 🔥2023.11.29: Support training and inference of AnimateDiff

• 🔥2023.11.24: Support yi-34b-chat, codefuse-codellama-34b-chat models. Corresponding sh scripts can be found in yi_34b_chat, codefuse_codellama_34b_chat.

• 🔥2023.11.18: Support tongyi-finance-14b series models: tongyi-finance-14b, tongyi-finance-14b-chat, tongyi-finance-14b-chat-int4. Corresponding sh scripts can be found in tongyi_finance_14b_chat_int4.

• 2023.11.16: Support flash attn for more models: qwen series, qwen-vl series, llama series, openbuddy series, mistral series, yi series, ziya series. Please use use_flash_attn parameter.

• 🔥2023.11.11: Support NEFTune, simply use Swift.prepare_model(model, NEFTuneConfig()) to enable.

• 🔥2023.11.11: Support training and inference by command line and inference by Web-UI, see Usage with Swift CLI section below for details.

• 🔥2023.11.10: Support bluelm series models: bluelm-7b, bluelm-7b-chat, bluelm-7b-32k, bluelm-7b-chat-32k. Corresponding sh scripts can be found in bluelm_7b_chat.

• 🔥2023.11.08: Support training and inference of xverse-65b model, script at xverse_65b.

• 🔥2023.11.07: Support training and inference of yi-6b, yi-34b models, scripts at yi_6b, yi_34b.

• 🔥2023.10.30: Support two new tuners: QA-LoRA and LongLoRA.

• 🔥2023.10.30: Support editing models using ROME (Rank One Model Editing) to infuse new knowledge into models without training!

• 2023.10.30: Support skywork-13b series models: skywork-13b, skywork-13b-chat. Corresponding sh scripts can be found in skywork_13b.

• 🔥2023.10.27: Support chatglm3 series models: chatglm3-6b-base, chatglm3-6b, chatglm3-6b-32k. Corresponding sh scripts can be found in chatglm3_6b.

• 🔥2023.10.17: Support SFT of int4, int8 models: qwen-7b-chat-int4, qwen-14b-chat-int4, qwen-vl-chat-int4, baichuan2-7b-chat-int4, baichuan2-13b-chat-int4, qwen-7b-chat-int8, qwen-14b-chat-int8.

• 2023.10.15: Support ziya2-13b series models: ziya2-13b, ziya2-13b-chat.

• 2023.10.12: Support mistral-7b series models: openbuddy-mistral-7b-chat, mistral-7b, mistral-7b-instruct.

• 🔥2023.10.07: Support DeepSpeed ZeRO-2, enabling lora (not just qlora) to run DDP on dual A10 cards.

• 2023.10.04: Support more math, law, SQL, code domain datasets: blossom-math-zh, school-math-zh, text2sql-en, sql-create-context-en, lawyer-llama-zh, tigerbot-law-zh, leetcode-python-en.

• 🔥2023.09.25: Support qwen-14b series: qwen-14b, qwen-14b-chat.

• 2023.09.18: Support internlm-20b series: internlm-20b, internlm-20b-chat.

• 2023.09.12: Support MP+DDP to accelerate full-parameter training.

• 2023.09.05: Support openbuddy-llama2-70b-chat.

• 2023.09.03: Support baichuan2 series: baichuan2-7b, baichuan2-7b-chat, baichuan2-13b, baichuan2-13b-chat.

🛠️ Installation

SWIFT runs in the Python environment. Please ensure your Python version is higher than 3.8.

• Method 1: Install SWIFT using pip command:

# Full capabilities
pip install 'ms-swift[all]' -U
# LLM only
pip install 'ms-swift[llm]' -U
# AIGC only
pip install 'ms-swift[aigc]' -U
# Adapters only
pip install ms-swift -U

• Method 2: Install SWIFT through source code (convenient for running training and inference scripts), please run the following commands:

git clone https://github.com/modelscope/swift.git
cd swift
pip install -e '.[llm]'

SWIFT depends on torch>=1.13, recommend torch>=2.0.0.

• Method 3: Use SWIFT in our Docker image

🚀 Getting Started

This section introduces basic usage, see the Documentation section for more ways to use.

Web-UI

Web-UI is a gradio-based interface for zero-threshold training and deployment. It is easy to use and perfectly supports multi-GPU training and deployment:

SWIFT_UI_LANG=en swift web-ui

Training

Training Scripts

You can refer to the following scripts to customize your own training script.

• full: qwen1half-7b-chat (A100), qwen-7b-chat (2*A100)
• full+ddp+zero2: qwen-7b-chat (4*A100)
• full+ddp+zero3: qwen-14b-chat (4*A100)
• lora: chatglm3-6b (3090), baichuan2-13b-chat (2*3090), yi-34b-chat (A100), qwen-72b-chat (2*A100)
• lora+ddp: chatglm3-6b (2*3090)
• lora+ddp+zero3: qwen-14b-chat (4*3090), qwen-72b-chat (4*A100)
• qlora(gptq-int4): qwen-7b-chat-int4 (3090)
• qlora(gptq-int8): qwen1half-7b-chat-int8 (3090)
• qlora(bnb-int4): qwen-7b-chat (3090)

Supported Training Processes

Training Process	Training Method
Pretraining	Text Generation
Fine-tuning	Single-turn/Multi-turn Agent Training/Self-cognition Multi-modal Vision/Multi-modal Speech
Human Alignment	DPO ORPO SimPO CPO KTO
Text-to-Image	DreamBooth, etc.
Text-to-Video	-

Single GPU Training

Start single GPU fine-tuning with the following command:

LoRA:

# Experimental Environment: A100
# GPU Memory Requirement: 20GB
# Runtime: 3.1 hours
CUDA_VISIBLE_DEVICES=0 \
swift sft \
    --model_type qwen1half-7b-chat \
    --dataset blossom-math-zh \
    --num_train_epochs 5 \
    --sft_type lora \
    --output_dir output \
    --eval_steps 200 \

Full-parameter:

# Experimental Environment: A100
# GPU Memory Requirement: 80GB
# Runtime: 2.5 hours
CUDA_VISIBLE_DEVICES=0 \
swift sft \
    --model_type qwen1half-7b-chat \
    --dataset blossom-math-zh \
    --num_train_epochs 5 \
    --sft_type full \
    --output_dir output \
    --eval_steps 500 \

Model Parallel Training

# Experimental Environment: 2 * A100
# GPU Memory Requirement: 10GB + 13GB
# Runtime: 3.4 hours
CUDA_VISIBLE_DEVICES=0,1 \
swift sft \
    --model_type qwen1half-7b-chat \
    --dataset blossom-math-zh \
    --num_train_epochs 5 \
    --sft_type lora \
    --output_dir output \

Data Parallel Training

# Experimental Environment: 4 * A100
# GPU Memory Requirement: 4 * 30GB
# Runtime: 0.8 hours
NPROC_PER_NODE=4 \
CUDA_VISIBLE_DEVICES=0,1,2,3 \
swift sft \
    --model_type qwen1half-7b-chat \
    --dataset blossom-math-zh \
    --num_train_epochs 5 \
    --sft_type lora \
    --output_dir output \

Combining Model Parallelism and Data Parallelism:

# Experimental Environment: 4 * A100
# GPU Memory Requirement: 2*14GB + 2*18GB
# Runtime: 1.7 hours
NPROC_PER_NODE=2 \
CUDA_VISIBLE_DEVICES=0,1,2,3 \
swift sft \
    --model_type qwen1half-7b-chat \
    --dataset blossom-math-zh \
    --num_train_epochs 5 \
    --sft_type lora \
    --output_dir output \

Deepspeed Training

Deepspeed supports training of quantized GPTQ and AWQ models.

ZeRO2:

# Experimental Environment: 4 * A100
# GPU Memory Requirement: 4 * 21GB
# Runtime: 0.9 hours
NPROC_PER_NODE=4 \
CUDA_VISIBLE_DEVICES=0,1,2,3 \
swift sft \
    --model_type qwen1half-7b-chat \
    --dataset blossom-math-zh \
    --num_train_epochs 5 \
    --sft_type lora \
    --output_dir output \
    --deepspeed default-zero2 \

ZeRO3:

# Experimental Environment: 4 * A100
# GPU Memory Requirement: 4 * 19GB
# Runtime: 3.2 hours
NPROC_PER_NODE=4 \
CUDA_VISIBLE_DEVICES=0,1,2,3 \
swift sft \
    --model_type qwen1half-7b-chat \
    --dataset blossom-math-zh \
    --num_train_epochs 5 \
    --sft_type lora \
    --output_dir output \
    --deepspeed default-zero3 \

ZeRO3-Offload:

# Experimental Environment: 4 * A100
# GPU Memory Requirement: 4 * 12GB
# Runtime: 60 hours
NPROC_PER_NODE=4 \
CUDA_VISIBLE_DEVICES=0,1,2,3 \
swift sft \
    --model_id_or_path AI-ModelScope/WizardLM-2-8x22B \
    --dataset blossom-math-zh \
    --num_train_epochs 5 \
    --sft_type lora \
    --output_dir output \
    --deepspeed zero3-offload \

Multi-node Multi-GPU

# If the disk is not shared, please additionally specify `--save_on_each_node true` in the shell scripts on each machine.
# node0
CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7 \
NNODES=2 \
NODE_RANK=0 \
MASTER_ADDR=127.0.0.1 \
NPROC_PER_NODE=8 \
swift sft \
    --model_type qwen1half-32b-chat \
    --sft_type full \
    --dataset blossom-math-zh \
    --output_dir output \
    --deepspeed default-zero3 \

# node1
CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7 \
NNODES=2 \
NODE_RANK=1 \
MASTER_ADDR=xxx.xxx.xxx.xxx \
NPROC_PER_NODE=8 \
swift sft \
    --model_type qwen1half-32b-chat \
    --sft_type full \
    --dataset blossom-math-zh \
    --output_dir output \
    --deepspeed default-zero3 \

AliYun-DLC multi-node training

In DLC product, WORLD_SIZE is the node number, RANK is the node index, this is different from the definition of torchrun.

NNODES=$WORLD_SIZE \
NODE_RANK=$RANK \
swift sft \
    --model_type qwen1half-32b-chat \
    --sft_type full \
    --dataset blossom-math-zh \
    --output_dir output \
    --deepspeed default-zero3

Pretraining

# Experimental Environment: 4 * A100
# GPU Memory Requirement: 4 * 30GB
# Runtime: 0.8 hours
NPROC_PER_NODE=4 \
CUDA_VISIBLE_DEVICES=0,1,2,3 \
swift pt \
    --model_type qwen1half-7b \
    --dataset chinese-c4#100000 \
    --num_train_epochs 1 \
    --sft_type full \
    --deepspeed default-zero3 \
    --output_dir output \
    --lazy_tokenize true

RLHF

# We support rlhf_type dpo/cpo/simpo/orpo/kto
CUDA_VISIBLE_DEVICES=0 \
swift rlhf \
    --rlhf_type dpo \
    --model_type qwen1half-7b-chat \
    --dataset shareai-llama3-dpo-zh-en-emoji \
    --num_train_epochs 5 \
    --sft_type lora \
    --output_dir output \

Inference

Original model:

CUDA_VISIBLE_DEVICES=0 swift infer --model_type qwen1half-7b-chat
# use VLLM
CUDA_VISIBLE_DEVICES=0 swift infer --model_type qwen1half-7b-chat \
    --infer_backend vllm --max_model_len 8192

LoRA fine-tuned:

CUDA_VISIBLE_DEVICES=0 swift infer --ckpt_dir xxx/checkpoint-xxx --load_dataset_config true
# use VLLM
CUDA_VISIBLE_DEVICES=0 swift infer \
    --ckpt_dir xxx/checkpoint-xxx --load_dataset_config true \
    --merge_lora true --infer_backend vllm --max_model_len 8192

Evaluation

Original model:

CUDA_VISIBLE_DEVICES=0 swift eval --model_type qwen1half-7b-chat \
    --eval_dataset ARC_c --infer_backend vllm

LoRA fine-tuned:

CUDA_VISIBLE_DEVICES=0 swift eval --ckpt_dir xxx/checkpoint-xxx \
    --eval_dataset ARC_c --infer_backend vllm \
    --merge_lora true \

Quantization

Original model:

CUDA_VISIBLE_DEVICES=0 swift export --model_type qwen1half-7b-chat \
    --quant_bits 4 --quant_method awq

LoRA fine-tuned:

CUDA_VISIBLE_DEVICES=0 swift export \
    --ckpt_dir xxx/checkpoint-xxx --load_dataset_config true \
    --quant_method awq --quant_bits 4 \
    --merge_lora true \

Deployment

The client uses the OpenAI API for invocation, for details refer to the LLM deployment documentation.

Original model:

CUDA_VISIBLE_DEVICES=0 swift deploy --model_type qwen1half-7b-chat
# 使用VLLM加速
CUDA_VISIBLE_DEVICES=0 swift deploy --model_type qwen1half-7b-chat \
    --infer_backend vllm --max_model_len 8192

LoRA fine-tuned:

CUDA_VISIBLE_DEVICES=0 swift deploy --ckpt_dir xxx/checkpoint-xxx
# 使用VLLM加速
CUDA_VISIBLE_DEVICES=0 swift deploy \
    --ckpt_dir xxx/checkpoint-xxx --merge_lora true \
    --infer_backend vllm --max_model_len 8192

Supported Models

The complete list of supported models and datasets can be found at Supported Models and Datasets List.

LLMs

Model Type	Model Introduction	Language	Model Size	Model Type
Qwen Qwen1.5 Qwen2 Qwen2.5	Tongyi Qwen series models	Chinese English	0.5B-110B including quantized versions	base model chat model MoE model code model
ChatGLM2 ChatGLM3 Codegeex2 GLM4 Codegeex4	Zhipu ChatGLM series models	Chinese English	6B-9B	base model chat model code model long text model
Baichuan Baichuan2	Baichuan 1 and Baichuan 2	Chinese English	7B-13B including quantized versions	base model chat model
Yuan2	Langchao Yuan series models	Chinese English	2B-102B	instruct model
XVerse	XVerse series models	Chinese English	7B-65B	base model chat model long text model MoE model
LLaMA2	LLaMA2 series models	English	7B-70B including quantized versions	base model chat model
LLaMA3 LLaMA3.1	LLaMA3 series models	English	8B-70B including quantized versions	base model chat model
Mistral Mixtral	Mistral series models	English	7B-22B	base model instruct model MoE model
Yi Yi1.5 Yi-Coder	01AI's YI series models	Chinese English	1.5B-34B including quantized	base model chat model long text model
InternLM InternLM2 InternLM2-Math InternLM2.5	Pujiang AI Lab InternLM series models	Chinese English	1.8B-20B	base model chat model math model
DeepSeek DeepSeek-MoE DeepSeek-Coder DeepSeek-Math DeepSeek-V2 DeepSeek-Coder-V2	DeepSeek series models	Chinese English	1.3B-236B	base model chat model MoE model code model math model
MAMBA	MAMBA temporal convolution model	English	130M-2.8B	base model
Gemma Gemma2	Google Gemma series models	English	2B-27B	base model instruct model
MiniCPM MiniCPM3	OpenBmB MiniCPM series models	Chinese English	2B-3B	chat model MoE model
OpenBuddy	OpenBuddy series models	Chinese English	7B-70B	base model chat model
Orion	OrionStar AI series models	Chinese English	14B	base model chat model
BlueLM	VIVO BlueLM large model	Chinese English	7B	base model chat model
Ziya2	Fengshenbang series models	Chinese English	13B	base model chat model
Skywork	Skywork series models	Chinese English	13B	base model chat model
Zephyr	Zephyr series models based on Mistral	English	7B	chat model
PolyLM	Tongyi Lab self-developed PolyLM series models	Multilingual	13B	base model
SeqGPT	Tongyi Lab self-developed text understanding model for information extraction and text classification	Chinese	560M	semantic understanding model
SUS	Southern University of Science and Technology model fine-tuned on YI	Chinese English	34B	chat model
Tongyi-Finance	Tongyi finance series models	Chinese English	14B	base model chat model financial model
CodeFuse-CodeLLaMA CodeFuse-Codegeex2 CodeFuse-Qwen	Ant CodeFuse series models	Chinese English	6B-34B	chat model code model
phi2/phi3	Microsoft's PHI series models	English	3B/4B	base model instruct model code model
Grok	X-ai	English	300B	base model
TeleChat	Tele-AI	Chinese English	7B-12B	chat model
dbrx	databricks	English	132B	base model chat model
mengzi3	Langboat	Chinese English	13B	base model
c4ai-command-r	c4ai	Multilingual	35B-104B	chat model
WizardLM2	WizardLM2 series models	English	7B-8x22B including quantized versions	chat model MoE model
Atom	Atom	Chinese	7B	base model chat model
Chinese-LLaMA-Alpaca-2	Chinese-LLaMA-Alpaca-2	Chinese	1.3B-13B	base model chat model long text model
Chinese-LLaMA-Alpaca-3	Chinese-LLaMA-Alpaca-3	Chinese	8B	base model chat model
ModelScope-Agent	ModelScope Agent series models	Chinese	7B-14B	agent model
Numina	AI-MO	English	7B	Math

MLLMs

Model Type	Model Introduction	Language	Model Size	Model Type
Qwen-VL Qwen2-VL	Tongyi Qwen vision model	Chinese English	2B-72B including quantized versions	base model chat model
Qwen-Audio Qwen2-Audio	Tongyi Qwen speech model	Chinese English	7B	base model chat model
YI-VL	01AI's YI series vision models	Chinese English	6B-34B	chat model
XComposer2 XComposer2.5	Pujiang AI Lab InternLM vision model	Chinese English	7B	chat model
DeepSeek-VL	DeepSeek series vision models	Chinese English	1.3B-7B	chat model
MiniCPM-V MiniCPM-V-2 MiniCPM-V-2.5 MiniCPM-V-2.6	OpenBmB MiniCPM vision model	Chinese English	3B-9B	chat model
CogVLM CogAgent CogVLM2 CogVLM2-Video GLM4V	Zhipu ChatGLM visual QA and Agent model	Chinese English	9B-19B	chat model
Llava-HF	Llava-HF series models	English	0.5B-110B	chat model
Llava1.5 Llava1.6	Llava series models	English	7B-34B	chat model
Llava-Next Llava-Next-Video	Llava-Next series models	Chinese English	7B-110B	chat model
mPLUG-Owl2 mPLUG-Owl2.1 mPLUG-Owl3	mPLUG-Owl series models	English	11B	chat model
InternVL Mini-InternVL InternVL2	InternVL	Chinese English	1B-40B including quantized version	chat model
Llava-llama3	xtuner	English	8B	chat model
Phi3-Vision	Microsoft	English	4B	chat model
PaliGemma	Google	English	3B	chat model
Florence	Microsoft	English	0.23B-0.77B	chat model
Idefics3	HuggingFaceM4	English	8B	chat model

Diffusion Models

Model Type	Model Introduction	Language	Model Type
AnimateDiff	AnimateDiff animation model	English	text-to-video
SD1.5/SD2.0/SDXL	StabilityAI series diffusion models	English	text-to-image

Supported Open Source Datasets

Dataset Type	Training Task	Documentation
General	Fine-tuning	🔥ruozhiba, 🔥ms-bench, 🔥alpaca-en(gpt4), 🔥alpaca-zh(gpt4), multi-alpaca, instinwild, cot-en, cot-zh, firefly-zh, instruct-en, gpt4all-en, sharegpt, tulu-v2-sft-mixture, wikipedia-zh, open-orca, sharegpt-gpt4, deepctrl-sft, coig-cqia.
Agent	Fine-tuning	🔥ms-agent, 🔥ms-agent-for-agentfabric, ms-agent-multirole, 🔥toolbench-for-alpha-umi, damo-agent-zh, damo-agent-zh-mini, agent-instruct-all-en.
General	Human Alignment	hh-rlhf, 🔥hh-rlhf-cn, stack-exchange-paired.
Code	Fine-tuning	code-alpaca-en, 🔥leetcode-python-en, 🔥codefuse-python-en, 🔥codefuse-evol-instruction-zh.
Medical	Fine-tuning	medical-en, medical-zh, 🔥disc-med-sft-zh.
Legal	Fine-tuning	lawyer-llama-zh, tigerbot-law-zh, 🔥disc-law-sft-zh.
Math	Fine-tuning	🔥blossom-math-zh, school-math-zh, open-platypus-en.
SQL	Fine-tuning	text2sql-en, 🔥sql-create-context-en.
Text Generation	Fine-tuning	🔥advertise-gen-zh, 🔥dureader-robust-zh.
Classification	Fine-tuning	cmnli-zh, 🔥jd-sentiment-zh, 🔥hc3-zh, 🔥hc3-en.
Quantization Assist	Quantization	pileval.
Other	Fine-tuning	finance-en, poetry-zh, webnovel-zh, generated-chat-zh, cls-fudan-news-zh, ner-jave-zh.
Vision	Fine-tuning	coco-en, 🔥coco-en-mini, coco-en-2, coco-en-2-mini, capcha-images.
Audio	Fine-tuning	aishell1-zh, 🔥aishell1-zh-mini.

Supported Technologies

Technology Name
🔥LoRA: LORA: LOW-RANK ADAPTATION OF LARGE LANGUAGE MODELS
🔥LoRA+: LoRA+: Efficient Low Rank Adaptation of Large Models
🔥GaLore:GaLore: Memory-Efficient LLM Training by Gradient Low-Rank Projection
🔥LISA: LISA: Layerwise Importance Sampling for Memory-Efficient Large Language Model Fine-Tuning
🔥UnSloth: https://github.com/unslothai/unsloth
🔥LLaMA PRO: LLAMA PRO: Progressive LLaMA with Block Expansion
🔥SCEdit: SCEdit: Efficient and Controllable Image Diffusion Generation via Skip Connection Editing < arXiv \
🔥NEFTune: Noisy Embeddings Improve Instruction Finetuning
LongLoRA: Efficient Fine-tuning of Long-Context Large Language Models
Adapter: Parameter-Efficient Transfer Learning for NLP
Vision Prompt Tuning: Visual Prompt Tuning
Side: Side-Tuning: A Baseline for Network Adaptation via Additive Side Networks
Res-Tuning: Res-Tuning: A Flexible and Efficient Tuning Paradigm via Unbinding Tuner from Backbone < arXiv \
Tuners provided by PEFT, such as IA3, AdaLoRA, etc.

Supported Hardware

Hardware Environment	Notes
CPU
RTX 20/30/40 series, etc.	After 30 series, BF16 and FlashAttn can be used
Computing cards T4/V100, etc.	BF16 and FlashAttn not supported
Computing cards A10/A100, etc.	Support BF16 and FlashAttn
Huawei Ascend NPU

Environment variables

• DATASET_ENABLE_CACHE: Enable cache when preprocess dataset, you can use 1/True or 0/False, default False
• WEBUI_SHARE: Share your web-ui, you can use 1/True or 0/False, default False
• SWIFT_UI_LANG: web-ui language, you can use en or zh, default zh
• WEBUI_SERVER: web-ui host ip，0.0.0.0 for all routes，127.0.0.1 for local network only. Default 127.0.0.1
• WEBUI_PORT: web-ui port
• USE_HF: Use huggingface endpoint or ModelScope endpoint to download models and datasets. you can use 1/True or 0/False, default False
• FORCE_REDOWNLOAD: Force to re-download the dataset

Other variables like CUDA_VISIBLE_DEVICES are also supported, which are not listed here.

📚 Classroom

Tutorial Name
Introduction to Deep Learning
Large Model Basics
Prompt Engineering
Transformer Architecture Introduction
Training Technique Selection
Data Preprocessing
Quantization
Training
Inference
Deployment
Evaluation

🏛 License

This framework is licensed under the Apache License (Version 2.0). For models and datasets, please refer to the original resource page and follow the corresponding License.

📎 Citation

@misc{zhao2024swiftascalablelightweightinfrastructure,
      title={SWIFT:A Scalable lightWeight Infrastructure for Fine-Tuning},
      author={Yuze Zhao and Jintao Huang and Jinghan Hu and Xingjun Wang and Yunlin Mao and Daoze Zhang and Zeyinzi Jiang and Zhikai Wu and Baole Ai and Ang Wang and Wenmeng Zhou and Yingda Chen},
      year={2024},
      eprint={2408.05517},
      archivePrefix={arXiv},
      primaryClass={cs.CL},
      url={https://arxiv.org/abs/2408.05517},
}

Star History