Sören Brunk

Thanks @parmentelat for the explanations. For anyone running into this until we have a proper fix, here's the workaround I used for publishing my slides on binder: i.e. in `postBuild`...

@betatim this seems to be the code generating the link. https://github.com/damianavila/RISE/blob/e7f0e183f148f89926b004aa2fa36278da8f9a05/rise/static/main.js#L487-L499

Sorry @hmf missed that somehow. I'd suggest we start with the simplest option, adding `hasBias(): Boolean` to `Module`. Since enabling/disabling bias is often a constructor parameter, I think it is...

> Would a "pico" example from Karpathy's ["Let's build GPT: from scratch, in code, spelled out"](https://www.youtube.com/watch?v=kCc8FmEb1nY) be of any interest? > > Later we could also "port" his nano GPT...

Thanks for your inputs @emartinezs44. > It may be a ver very very large task. The huggingface library has become hugh. I´m pretty sure that the most used models in...

@hmf You're right `Embedding` is an example where the input type might be different from the output, so we can't inherit from `TensorModule`. Note that @davoclavo has also added `Embedding`...

Sounds good to me 👍

> I see that we have 2 options, a function in the Loss package (does not exist yet, only binary version available) and the `torch.nn.loss.CrossEntropyLoss` version. The storch examples use...

> I will also be available to continue work on the pending PR to get it merged, in case I can help in any way @sbrunk @davoclavo feel free to...

> Best of luck on the Scala Days talk! Hopefully it will be streamed/recorded, I'd love to watch it :D Thanks! I'm sure it will be recorded and put on...