Update readme to detail the lerobot dataset format #275
Conversation
There was a problem hiding this comment.
Thanks so much @thomwolf . I have comments but my disposition is approved / non-blocking.
| Here are the important details and internal structure organization of a typical `LeRobotDataset` instantiated with `dataset = LeRobotDataset("lerobot/aloha_static_coffee")`. The exact features will change from dataset to dataset but not the main aspects: | ||
|
|
||
| ``` | ||
| dataset attributes: |
There was a problem hiding this comment.
I wish there was a way of flagging this snippet so that there's a CI mechanism to draw our attention to it in case we change anything in the datasets module ... or something, anything to make sure it doesn't go stale. I wonder if someone at HF has solved this problem...
There was a problem hiding this comment.
Yes indeed, we have that in transformers for docstring examples. It's quite a lot of work to setup and maintain so I would advise to wait that the library has really built a large community of users before diving in this.
|
Thanks @alexander-soare and @aliberts Added all the suggestions. Feel free to merge when you feel like. |
There was a problem hiding this comment.
Thank you @thomwolf! 🙏
All good to me, I'll let @alexander-soare merge if it's good for him too
What this does
This Pull Request detail how the LeRobot Dataset format is structured and created.
It also adds a mention of the often overlooked
DATA_DIRenvironment variable.It's extracted from #246
How it was tested
It's only a readme update thus no tests.
How to checkout & try? (for the reviewer)
Read the changes.
This change is