Gym
Gym copied to clipboard
NVIDIA-NeMo
Build RL environments for LLM training
This PR enables running Gym on Aviary environments. The two main concepts: - `AviaryResourcesServer`: maps to an Aviary `TaskDataset`: spawns and manages multiple environments - Unlike other `ResourcesServer`s, it doesn't...
This implements the `ng_status` command to list all running servers on the system and ping for health check.
need to set uv pip install python flag in colab environments when launching servers usage: `ng_run "+config_paths=[...]" +uv_pip_set_python=true ` defaults to false For https://github.com/NVIDIA-NeMo/Gym/issues/370 Needed for notebook here: https://docs.unsloth.ai/models/nemotron-3#reinforcement-learning--nemo-gym
**Use cases, pain points, and background** Why should we do this? Why is this needed or wanted? **Description**: What should we do? **Design**: What files should be touched? What logic...
**Use cases, pain points, and background** **Description**: **Design**: **Out of scope**: **Acceptance Criteria**: - [ ] All training environments must be trainable easily with an instruct and thinking model -...
**Use cases, pain points, and background** Openhands is well support in SkyRL as an environment. Will there be an Openhands environment in NemoGym? Looking to train a coding agent on...
**Use cases, pain points, and background** Why should we do this? Why is this needed or wanted? **Description**: What should we do? **Design**: What files should be touched? What logic...
**Use cases, pain points, and background** Why should we do this? Why is this needed or wanted? **Description**: What should we do? **Design**: What files should be touched? What logic...
**Use cases, pain points, and background** Why should we do this? Why is this needed or wanted? **Description**: Just use OpenAI Agents SDK which uses OpenAI Responses schema which is...