Introducing Weaver
Weaver is a training API designed for researchers and developers working with LLMs. As an ecosystem-oriented product, it seamlessly integrates with agent frameworks and reinforcement learning training frameworks, enabling flexible fine-tuning of agentic models. Weaver lets you focus on what matters – your data, algorithms and agents – while handling the complexity of distributed training infrastructure.
Weaver is inspired by Thinking Machine Labs' Tinker.
Training API
Weaver provides a Python API for fine-tuning large language models with efficient training and flexible deployment capabilities.
- Training:
forward_backward()andoptim_step()for gradient computation and model updates, with support for various loss functions - Inference:
sample()andcompute_logprobs()for text generation with configurable parameters and probability computation - Weight Management: Simple APIs for model persistence and checkpoint management
Agent-Native Ecosystem
NexRL
NexRL is an RL training framework seamlessly work with Weaver backend for large-scale post-training. It features modularized components for building custom RL pipelines, providing maximum flexibility and extensibility while maintaining clean abstractions and ease of use.
NexAU
NexAU is a general-purpose agent framework for building intelligent agents with tool capabilities. It provides a modular tool system, a flexible agent architecture, and seamless integration with various LLM providers. It also supports seamless tracing for both standalone usage and reinforcement learning.
版权所有 © 上海奇绩智峰智能科技有限公司 | 沪ICP备2025144628号-2