Jianhui Chen
Jianhui Chen
> Hmm reach is a pretty easy task. Are you computing your success metric correctly. There is a guide on how to do so in the paper. Yes, reach-v2 is...
Just curious. How is the performance on atrai ram? I suffered from very poor performance using ppo compared to RGB state.
> @alexxchen ALE Ram state is useful for object-centric learning work @renos Thanks for your information! I hit a blind spot in my knowledge!
OK， It's the max_eval, I use an extra loop to control the optimization, so it is set to 1, which will lead to the problem of length of queue to...
I found the following code in vllm/inputs/registry.py ``` def process_input(self, model_config: "ModelConfig", inputs: ProcessorInputs) -> ProcessorInputs: """ Apply an input processor to an instance of model inputs. The model is...
How about supporting multi-modal RAG?