Zhuoheng Li

https://github.com/stepjam/RLBench/blob/46cb9d97003961428f491dd7cc75433f6ab351dd/rlbench/gym/rlbench_env.py#L50-L63 Although the original RLBench environment includes depth images and point clouds in the observation, the gym version does not provide such information. They are important if researchers want to...

Great work! Any plan to train a smaller version, e.g. around 3B?

4

Hello, It's a really great work which contributes a lot to the community! Do you have any plan to train a smaller version of large world model (e.g., 1~3B), which...

[Feature Request] Support multimodal LLM, e.g., llava

2

Hello, Would you like to support mllm like llava? ```[tasklist] ### Tasks ```

Questions about the technical report

2

Hello, It's a great work! And there are several questions: 1. In the technical report you mentioned > We find that LoRA empirically leads to better performance than fully tuning...

Hello @siddk How many TPU hours are needed to train this model (including image reconstruction)?

[Bug] ImportError: DLL load failed while importing _turbomind: 找不到指定的模块。

21

### Checklist - [ ] 1. I have searched related issues but cannot get the expected help. - [ ] 2. The bug has not been fixed in the latest...

Do you support finetuning with LoRA?

2

Hello, Do you support finetuning with LoRA? Or I can directly use your training code for finetuning (but I am afraid that the model will forget too much knowledge in...

Hello @cheng-chi , In your colab notebook I find that you mention the EMA model can "accelerate training and improves stability": However, in the training loop, you may only update...

Can we use mmh3 instead of pyhash? It's still difficult to install pyhash in Google Colab. My script: ``` # Switch to python3.8 !apt-get update -y !apt-get install python3.8 python3.8-distutils...