agent-lightning
agent-lightning copied to clipboard
microsoft
The absolute trainer to light up AI agents.
When there is a VLLM serving error, such as out of the length of the max length of a tiny LLM, the output is empty, then the trace backward to...
This PR addresses an issue with **stale task requeueing** that could cause **duplicate rollouts** to be received and stored. #### Problem - The server requeues tasks when they time out...
Hi. I wrote a client script based on the Agno framework by following the RAG example. I defined a custom reward function and added @reward, the return value of training_rollout_async...
I build an agent by Golang. Is possible to train the LLM that this agent use?
This PR enhances the Text2SQL evaluation capabilities in Agent Lightning by providing comprehensive benchmark results and detailed metrics that were previously unavailable. ## What's Added ### 📊 Detailed Execution Accuracy...
The AgentOps instrumentation was causing agent crashes when LLM servers returned HTTP 503 (Service Unavailable) errors. This resulted in agents returning `answer: None` instead of computing actual results. ## Problem...
When I run the calc_agent example, the client side is able to generate rollouts and rewards, but on the server side it always shows empty response — the Triplet objects...
I’m encountering an authentication issue when running the sql_agent.py script in the examples/spider directory. The error message is: AgentOps: Authentication failed: Authentication failed: 504 Has anyone else encountered this issue...
The ARPO algorithm effectively improves the performance of multi-round tool inference agents and solves the problems of insufficient exploration and lack of generalization ability of existing sample-level RL methods in...
