PyVision-Image-RL: 10K-100K RL Trajectories for Vision Agents

PyVision-Image-RL-Data provides between 10,000 and 100,000 reinforcement learning trajectories for training agentic vision models, released by Agents-X in February 2026. The data supports the PyVision-RL framework, which focuses on stabilizing multimodal model interactions during complex vision-language tasks.

Use Cases

• Training multimodal models for agentic vision tasks using RL trajectories
• Evaluating model stability in image-to-text interaction loops
• Developing open-weight vision-language agents within the PyVision-RL framework

Strengths

• Contains 10,000 to 100,000 RL training records
• Released under Apache 2.0 license for open-weight development
• Directly linked to Arxiv 2602.20739 research paper

Limitations

• Lack of explicit column documentation in the metadata
• Highly specialized for agentic RL tasks, limiting utility for general supervised learning

Provenance

Source

Agents-X, 'PyVision-RL: Forging Open Agentic Vision Models via RL'

Freshness

Last updated February 2026.