castor

castor copied to clipboard

I want to set mpcnn on stsbenchmark

noooob-coder

mengf821

Improve KimCNN results

1

One and a half years later, I'm finally getting better results on KimCNN using the original hyperparameters in the paper. There are a few discrepancies with the PyTorch and Castor...

daemon

In preparation for the NAACL2019 camera ready and future work, @achyudh is currently working on moving the document classification models into [Hedwig](https://github.com/castorini/hedwig).

daemon

Create a simple, reproducible, out-of-the-box snapshot for doc classification paper

1

The document classification paper deserves a _clean_ snapshot for reproducibility and extensibility. Readers should be able to click the link in the paper, run a few commands in the README,...

daemon

Different results for different batch sizes when evaluating trained models

2

Hi, First of all, thanks for making your great code and models available. I am currently trying out two of your models (MP-CNN and VDPWI) and noticed that when evaluating...

AxelMueller

Reduce code redundancy

1

A few issues we should address: 1. Redundant [argument parsers](https://github.com/castorini/Castor/blob/master/vdpwi/__main__.py#L25) throughout the codebase. I think moving to a [hierarchical JSON config system](https://github.com/daemon/argconf/tree/master/examples) makes sense, where we have a single global...

daemon

in training sm_cnn, ValueError: could not convert string to float: '<pad>'

6

$ python train.py --mode static --gpu 1 Note: You are using GPU for training Dataset TREC Mode static VOCAB num 13 LABEL.target_class: 13 LABELS: ['', '2', '0', '7', '3', '1',...

liudonglei

The ConvRNN implementation doesn't use anything in `common.*`, not to mention that `getData.sh` doesn't work -- and it's bad practice, since we agreed to use `Castor-data` only. Related: #102

daemon

Fix insane memory usage when loading datasets

2

@achyudhk reports that CharCNN on some dataset uses 63GB of RAM (Hydra and Dragon both have 64GB). I think a solution would be some mechanism for moving data between disk...

daemon