Could you please share the approximate resource requirements for finetuning with AgentRL with Qwen 2.5 7B or Qwen 2.5 3B?
Specifically:
How many GPUs are recommended?
How much memory?
How much storage space should I plan for?
Thanks for your awesome work! 🙏