Swift DOCUMENTATION

Get Started

• Installation and Usage
  • Wheel Package Installation
  • Source Code Installation
  • Notebook Environment
  • Build Documentation
• Interface Training and Inference
• Push Model
  • Pushing Model During Training
  • Pushing Model During Export
• Basic Usage
  • Using in Training
  • Using in Inference
• Interface List
  • Swift Class Static Interfaces
  • SwiftModel Interface
• Res-Tuning Component
  • [NeurIPS 2023] Res-Tuning: A Flexible and Efficient Tuning Paradigm via Unbinding Tuner from Backbone
• 🔥SCEdit
• Compatibility with Peft

Instruction

• Instructions
  • 🐔Features
• LLM Fine-tuning Documentation
  • Table of Contents
  • Environment Preparation
  • Fine-Tuning
  • DPO
  • ORPO
  • Merge LoRA
  • Quantization
  • Inference
  • Web-UI
• LLM Inference Documentation
  • Table of Contents
  • Environment Preparation
  • Inference
  • Web-UI
• LLM Evaluation Documentation
  • Table of Contents
  • Introduction
  • Environment Setup
  • Evaluation
  • Custom Evaluation Sets
• LLM Quantization and Export Documentation
  • Table of Contents
  • Environment Preparation
  • Original Model
  • Fine-tuned Model
  • QLoRA
  • Pushing Models
• LLM Experiment Documentation
  • Table of Contents
  • Environment Setup
  • Prepare Experiment Configuration
  • Run Experiments
  • Collect Experiment Results
• Command Line Arguments
  • Table of Contents
  • sft Parameters
  • PT Parameters
  • RLHF Parameters
  • infer merge-lora Parameters
  • export Parameters
  • eval parameters
  • app-ui Parameters
  • deploy Parameters
  • web-ui Parameters
• Supported models and datasets
  • Table of Contents
  • Models
  • Datasets
• Customization and Extension
  • Table of Contents
  • Custom Dataset
  • Custom Models
  • Custom Dialogue Templates
• Frequently Asked Questions in LLM & VLM Training, Inference, Deployment, and Evaluation
  • Training
  • Inference
  • Deployment
  • Evaluation

LLM Training and Inference

• LLM Documentation
  • 📚Tutorials!
  • ⭐️Best Practices!
  • 🐔References！
  • 🍀Multi-Modal Best Practices!
• Human Preference Alignment Training Documentation
  • Table of Contents
  • Environment Setup
  • Dataset
  • DPO
  • RM
  • PPO
  • KTO
  • CPO
  • ORPO
  • SimPO
• OLLaMA Export Documentation
  • Contents
  • Environment Setup
  • Export
  • Points to Note
• VLLM Inference Acceleration and Deployment
  • Table of Contents
  • Environment Preparation
  • Inference Acceleration
  • Web-UI Acceleration
  • Deployment
  • Multiple LoRA Deployments
  • VLLM & LoRA
• LmDeploy Inference Acceleration and Deployment
  • Table of Contents
  • Environment Preparation
  • Inference Acceleration
  • Deployment
  • Multimodal
• Megatron Training Documentation
  • Table of Contents
  • Environment-Preparation
  • SFT-Example
  • Multi-Node Pre-Training Example
  • Mapping between MegatronArguments and SftArguments
• Best Practices for Self-Cognition Fine-Tuning
  • Table of Contents
  • Environment Setup
  • Inference Before Fine-Tuning
  • Fine-Tuning
  • Inference After Fine-Tuning
  • Web-UI
• Agent Fine-tuning Best Practices
  • Table of Contents
  • Environment Setup
  • Data Preparation
  • Fine-tuning
  • Inference
  • Usage-with-Modelscope_Agent
  • Summary
• Agent Deployment Best Practice
  • Table of Contents
  • Environment Setup
  • Tools Field
  • Deployment
• Qwen1.5 Full Process Best Practices
  • Table of Contents
  • Environment Preparation
  • Qwen1.5-7B-Chat
  • Qwen1.5-72B-Chat
• NPU Best Practice
  • Table of Contents
  • Environment Preparation
  • Fine-tuning
  • Inference
• Hands-on Training and Inference with Grok 300B
  • Table of Contents
  • Environment Setup
  • Finetuning
• LLM Human Alignment Training Documentation
  • Table of Contents
  • Environment Preparation
  • Human Alignment Training
• Best Practices for ORPO Algorithm
  • Environment Preparation
  • Dataset
  • Training
  • Inference
• Best Practices for SimPO Algorithm
  • Environment Preparation
  • Dataset
  • Training
  • Inference
• HuggingFace Eco-compatibility
• Benchmark
  • Table of Contents
  • Parameter Settings
  • Quantization
  • Model Type & Max Length
  • Batch Size
  • Use Flash Attn & Gradient Checkpointing
  • LoRA Rank & LoRA Target Modules
  • Gradient Accumulation Steps
  • Tuners
  • unsloth
  • Export
  • AWQ
  • AQLM
  • Sequence Parallel

Multi-Modal LLM Training and Inference

• Multi-Modal Documentation
  • 📚 Tutorial
  • Multi-Modal Best Practice
• Human Preference Alignment Training Documentation
  • Table of Contents
  • Environment Setup
  • Dataset
  • DPO
  • RM
  • CPO
  • ORPO
  • SimPO
• LmDeploy Inference Acceleration and Deployment
  • Table of Contents
  • Environment Preparation
  • Inference Acceleration
  • Deployment
• vLLM Inference Acceleration Documentation
  • Table of Contents
  • Environment Setup
  • Inference Acceleration
  • Deployment
• Mutlimoda LLM Deployment
  • Table of Contents
  • Environment Setup
  • qwen-vl-chat
  • yi-vl-6b-chat
  • minicpm-v-v2_5-chat
  • Audio and Video Modalities
• Qwen-VL Best Practice
  • Table of Contents
  • Environment Setup
  • Inference
  • Fine-tuning
  • Inference after Fine-tuning
• Qwen2-VL Best Practice
  • Table of Contents
  • Environment Setup
  • Inference
  • Fine-tuning
• Qwen-Audio Best Practice
  • Table of Contents
  • Environment Setup
  • Inference
  • Fine-tuning
  • Inference After Fine-tuning
• Llava Best Practice
  • Table of Contents
  • Environment Setup
  • Inference
  • Fine-tuning
  • Inference after Fine-tuning
• Llava Video Best Practice
  • Table of Contents
  • Environment Setup
  • Inference
  • Fine-tuning
  • Inference after Fine-tuning
• InternVL Best Practice
  • Table of Contents
  • Environment Setup
  • Inference
  • Fine-tuning
  • Inference after Fine-tuning
• Deepseek-VL Best Practice
  • Table of Contents
  • Environment Preparation
  • Inference
  • Fine-tuning
  • Inference After Fine-tuning
• Internlm-Xcomposer2 Best Practice
  • Table of Contents
  • Environment Preparation
  • Inference
  • Fine-tuning
  • Inference After Fine-tuning
• Phi3-Vision Best Practice
  • 目录
  • Environment Setup
  • Inference
  • Fine-tuning
  • Inference After Fine-tuning
• Yi-VL Best Practice
  • Table of Contents
  • Environment Setup
  • Inference
  • Fine-tuning
  • Inference After Fine-tuning
• Florence Best Practice
  • Table of Contents
  • Environment Setup
  • Inference after Fine-tuning
• CogVLM Best Practice
  • Table of Contents
  • Environment Setup
  • Inference
  • Fine-tuning
  • Inference After Fine-tuning
• CogVLM2 Best Practice
  • Table of Contents
  • Environment Setup
  • Inference
  • Fine-tuning
  • Inference After Fine-tuning
• GLM4V Best Practice
  • Table of Contents
  • Environment Setup
  • Inference
  • Fine-tuning
  • Inference After Fine-tuning
• CogVLM2 Video Best Practice
  • Table of Contents
  • Environment Setup
  • Inference
  • Fine-tuning
  • Inference After Fine-tuning
• MiniCPM-V Best Practice
  • Table of Contents
  • Environment Setup
  • Inference
  • Fine-tuning
  • Inference After Fine-tuning

API Doc

• Hub
  • HubApi
  • check_local_model_is_latest
  • push_to_hub
  • push_to_hub_async
  • snapshot_download
  • model_file_download
• Trainer
  • Seq2SeqTrainer
  • Trainer
• Tuner
  • AdapterConfig
  • SwiftModel
  • Swift
  • LoRAConfig
  • PromptConfig
  • ResTuningConfig
  • SideConfig
  • SwiftConfig
  • SwiftOutput

Indices and tables

• Index

• Module Index

• Search Page