Aurélien Geron
Aurélien Geron
Thanks for your kind message, I really appreciate it! Regarding your question, I understand how this can be confusing: the arrow's "x3" label is meant to indicate that the output...
Hi @liganega , My apologies for the very late reply. You make an excellent point, indeed we should not use the targets for clustering, it would be a leakage. But...
Thanks a lot, I merged the PR. 👍
I'm getting different results from `Blip2ForConditionalGeneration` depending on the Transformers library version. Showing the usual photo of two cats sleeping on a couch, I get the following results: * ✅...
Here's a [gist colab notebook](https://colab.research.google.com/gist/ageron/8a3c87d9af26f0517fcf813af1ab567d/blip-2-issue.ipynb) that reproduces the issue.
FYI, I just tried to apply the #38510 but it didn't fix the issue.
Thanks for your question. Yes, this can happen, ML models don't always get things right. Moreover, the `SGDClassifier` model uses randomness during training (by randomly initializing the weights, and by...
I'm glad I could help! 👍
Thanks for your feedback! Have you looked at the repository for the 3rd edition? It now uses Gymnasium instead of Gym, and it should work fine. https://github.com/ageron/handson-ml3
Yes, that makes sense. However, my code only calls this function with lists of F64s, so I feel like it's still a bug, wouldn't you agree?