vime Documentation

vime is an LLM post-training framework for RL scaling, providing two core capabilities:

• High-Performance Training: Supports efficient training in various modes by connecting Megatron with vLLM;

• Flexible Data Generation: Enables arbitrary training data generation workflows through custom data generation interfaces and server-based engines.

vime is built on slime, the RL framework behind GLM-4.7, GLM-4.6 and GLM-4.5. vime keeps slime’s training stack and data-generation design while using vLLM as the default rollout backend, and inherits broad model support from slime, including:

• Qwen3 series (Qwen3Next, Qwen3MoE, Qwen3), Qwen2.5 series;

• DeepSeek V3 series (DeepSeek V3, V3.1, DeepSeek R1);

• Llama 3.

Get Started

• Quick Start
• Usage Guide
• Customization Guide
• FAQ

Dense

• Qwen3-4B with 8xH100

MoE

• Qwen3-30B-A3B with 8xH100

Advanced Features

• Speculative Decoding
• Reproducibility
• Fault Tolerance
• PD Disaggregation
• vLLM Config: Advanced Engine Deployment
• Megatron Config: Role-Based Training Overrides
• Supporting Model Architectures Beyond Megatron-LM

Other Usage

• Fully-Async Rollout Example
• Multi-Agent RL

Developer Guide

• CI (Continuous Integration)
• Debugging
• Trace Viewer
• Profiling