Navigating These Docs
These docs are organized by workflow: start with Weaver basics, then move into the API surface, and finally connect NexRL and NexAU for agentic model training.
Documentation Structure
Getting Started
For first-time Weaver users:
- What is Weaver?: Weaver's purpose, abstractions, and use cases.
- Core Concepts: how Organizations, Projects, Sessions, Training Runs, and Checkpoints fit together.
- Installation: install the Python SDK, configure an API key, and run a first training example.
Weaver API
For writing training scripts or integrating an existing framework:
- Training and Sampling: Session metadata, sync/async clients, metrics, training loops, sampling, and logprobs.
- Loss Functions: cross entropy, importance sampling, PPO/GRPO-related losses, and custom loss paths.
- Saving and Loading: sampler exports, durable checkpoints, optimizer restore, and TTL.
- Model Catalog: runtime model discovery, training modes, context limits, pricing, and performance tiers.
Console and Access
For collaboration, Project setup, and cost management:
- Organizations, Teams & Projects: member roles, sharing, archiving, and selecting an Organization and Project in the SDK.
- Usage, Billing & Quota: token usage, balance, cost, and quota requests.
Agentic Model Training
For training models that use tools, run multi-step tasks, or learn from agent traces:
- Finetune Models: use NexRL to orchestrate rollout, trajectory pools, rewards, and Weaver training.
- Build & Finetune Agents: define NexAU agents and connect their trajectories to NexRL/Weaver.
Suggested Paths
If you want the smallest working example:
If you are doing RL or agent training:
If you are integrating experiment management, pay special attention to:
- Checkpoint save strategy and TTL.
- The default 1-hour TTL for sampler weight exports.
- How
performance_tieraffects throughput and cost. - Async
OperationHandleresults fromwait=False.