jax

Environment info

• transformers version: 4.11.2
• Platform: Linux-5.4.0-66-generic-x86_64-with-glibc2.10
• Python version: 3.8.8
• PyTorch version (GPU?): 1.9.1+cu102 (True)
• Tensorflow version (GPU?): not installed (NA)
• F

Jax is missing support for scipy.special.erfcx (https://docs.scipy.org/doc/scipy/reference/generated/scipy.special.erfcx.html). Please add this function.

Pyro's HMC and NUTS implementations are feature-complete and well-tested, but they are quite slow in models like the one in our Bayesian regression tutorial that operate on small tensors for reasons that are largely beyond our control (mostly having to do with the design and implementation of torch.autograd), which is unfortunate because these

Is there a way to train a bidirectional RNN (like LSTM or GRU) on trax nowadays?

TruncatedDistribution has both low and high. Why do TruncatedNormal and TruncatedCauchy only have low?

Dear Brax team,

Since Brax is fully differentiable, I thought it'd be possible to use it like DiffTaichi or GradSim for system identification (e.g. determining the mass of an object from a trajectory and known force) but I couldn't find any example for this.
Do you happen to have any demo or tips for this?

From the top of my head I would do something like this:
Let's say the task is to es

We should add binary_sparsemax_loss, the binary classification counterpart of multiclass_sparsemax_loss, see section 4.4 of https://arxiv.org/abs/1901.02324. This loss is also known modified Huber loss. The associated mapping is the projection on the unit cube so let's add projection_unit_cube, which is easy to implement using projection_box. Once done, they need to be added to the documen

Summary

@dantrim has shown very nicely with his hamming-codec project that Typer provides a nice Click based CLI interface with typing support to minimize the amount of code that needs to be written. I now also see that Typer nicely motivates shell completions for the user

$ docker run --rm -ti pytho