Frédéric Branchaud-Charron

Sounds good! My calendar is available on Calendly: https://calendly.com/baal-org/30min?month=2023-05 If these don't work for you let me know and we will figure something out :) We can chat on Slack...

Fixed in #270 thank you @reeshipaul for the PR!

Hello! I'm not super familiar with s2t, but it should be similar to text generation which we support to some extent. Do you use huggingface or another library? I can...

Cool! So as expected it's pretty similar to text generation. I recorded a Loom for you and there is also a gist. [Loom](https://www.loom.com/share/b12dce4d9f6d4a68b078186cb529ab5e?sid=3dae342a-2b1e-4f96-af50-a62c34197f50) [Gist](https://gist.github.com/Dref360/89d101fb296ee12ef24d1b923fa02d0b) Now that the uncertainty estimation seems...

It is simple to switch between MC-Dropout and Deterministic inference with Baal. But theoretically, we tend to use the mean prediction (we call this Bayesian Model Average, but we just...

Thank you for submitting this issue! I'll take a look more deeply over the weekend, I don't have a super good idea how to fix this right now unfortunately.

Hello, Thank you for submitting the issue. I should be able to take a look over the weekend. I'm a bit puzzled because we simply call [self.get_eval_dataloader(dataset)](https://github.com/baal-org/baal/blob/master/baal/transformers_trainer_wrapper.py#L66C9-L66C55) which is managed...

I was thinking about it and maybe it's because of the stacking we perform. For our HF implementation, we always perform MC-Dropout in a single pass meaning that batch_size=2 will...

I just opened #281 which should allow you to run your experiment. If you can install Baal from source from this branch, you could update your code with: ``` trainer...

Hello, I'm not sure I understand. The return value of `ds.load_faiss_index` is None as expected. I see that loading an Index on a dataset that doesn't have an `embedding` column...