Yuge Zhang
Yuge Zhang
1. I can successfully use open preview to the side without an issue. Please confirm that extension is loaded and initialized properly. 2. Please provide more details (what models you...
Are you returning rewards from `@rollout` method or using `emit_reward`? If you are using `emit_reward`, it actually creates an opentelemetry spans and you need to make sure that your tracer...
put a print here: https://github.com/Lincyaw/agent-lightning/blob/8a4491f821f1ee234561ab0b2f5c52beec23e738/agentlightning/runner/agent.py#L258 Do you see a span generated here?
If you are seeing a span here, and the span did get sent into the store. And you are not seeing any rewards at VERL. there could be too reasons:...
not really. agentops instrumented the generally openai provider, which is general. Missing instrumentation of agent framework will only affect discriminating multiple agents. There is however a bug fix for openai...
> Switched to the latest main, the issue remains > > ``` > (TaskRunner pid=2885675) Warning: Rollout ro-34788af75c53 contains empty response: [Triplet(prompt={'token_ids': []}, response={'token_ids': []}, reward=None, metadata={'response_id': 'chatcmpl-5e22b526f0324d0698df20396f5552b2', 'agent_name': '*'}),...
What are the log output you are getting from running this script?
I think the problem is with the LlmProxy. LlmProxy registers a global tracer provider that will interfere with other tracer providers registered in runner. You must run LlmProxy in another...
partially resolved by #293.