esvhd

HDBSCAN on GPU?

11

Hi guys, Will there be any plan to develop a GPU enabled version of this amazing algorithm/package? Unfortunately I have to admit that I'm new to GPU so won't be...

Detail on WeightDrop class `_setup()` cuDNN RNN weight compacting issue & `register_parameter()`

7

Hi there, cc @Smerity Thanks for sharing the code first of all. I've been diving into the details and would really appreciate if you could share some insight into `WeightDrop`...

Hi, I noticed that when a `FileDocument` object is created, its `is_error` field is set to `True` if the document has content. Pls see line below. https://github.com/LexPredict/openedgar/blob/1d1b8bc8faa3c59e05d9883e53039b6328b7d831/lexpredict_openedgar/openedgar/tasks.py#L89 Would someone be...

Package requirement conflicts with lexnlp 0.1.8

1

Hi, I encountered at least two python package version conflicts when trying to set this up. Based on code here, lexnlp needs 0.1.8 https://github.com/LexPredict/openedgar/blob/1d1b8bc8faa3c59e05d9883e53039b6328b7d831/lexpredict_openedgar/requirements/full.txt#L138 This is conflicting with `pandas` and...

Hi @vinhkhuc , Would you be interested in a PR for the cuda version of these 4 models? I made some simple modifications to make them work on GPUs, while...

Why is dividing by e**(1/4) for both keys and queries more memory efficient?

4

Hi, Would you mind explaining why the follow code is more memory efficient than just dividing one of them by `sqrt(e)`? https://github.com/pbloem/former/blob/7b12ae6025cb5797c7ce95750c9777eb2130ba25/former/modules.py#L48-L52 Thank you.

Or if this is not possible is there a way to get the dividend history so one can perform the necessary adjustment? Also, are q/i/x/p the only parameter keywords available?...