Benjamin Minixhofer

Multithreading across the batch dimension should be quite easy to implement (ref [wasm-bindgen rayon example](https://rustwasm.github.io/docs/wasm-bindgen/examples/raytrace.html)) without any need for changes in tract ([like this](https://github.com/snipsco/tract/issues/326#issuecomment-661149080)). Would only need a new option...

As mentioned in https://github.com/bminixhofer/tractjs/blob/master/README.md#what-about-size not inlining the Worker and WASM would reduce size. With the recent changes in the build system this should be much easier than before. At the...

`q.site.upload` is slow

9

#### Wave SDK Version, OS 0.17.0, Linux #### Actual behavior We are dealing with some large files (predictions, model weights etc.) with a couple hundred MB, which we want to...

How is accuracy on OPUS-100 computed?

1

Hi! Thanks for this library. Since there is no notion of documents in the OPUS-100 dataset it is not clear to me how accuracy is computed. I tried a naive...

Details on training of published models

1

Hi! I really enjoyed your paper and the code. Do you have some more details regarding the published Italian and Dutch models? Specifically: - for how many steps did you...

Increase accuracy by weighting probs with softmax

2

Hi! I just came across this library. Great idea, and great implementation! This PR improves the score on GermaNet compounds from 95.20% to 95.91% on my machine. This is done...

Questionable result in Gamblers Problem Solution

11

I encountered an interesting issue in the [Gamblers Problem notebook ](https://github.com/dennybritz/reinforcement-learning/blob/master/DP/Gamblers%20Problem%20Solution.ipynb). The graph of capital vs policy looks like this with `p_h = 0.4`:  In the book (on page...