deep_qa

deep_qa copied to clipboard

evaluate_model() gives parsing error

1

I am trying to evaluate a model using evaluate_model() `evaluate_model("some.json") ` but it gives a ParseException. some.json- ``` { "model_class": "BidirectionalAttentionFlow", "model_serialization_prefix": "/models/bidaf", "encoder": { "word": { "type": "cnn", "ngram_filter_sizes":...

mohitbadwal

Use virtualenv instead of conda

6

We use conda to create a consistent environment for Python. Is there a reason we use conda over virtualenv? I was reading the [AI2 Python Guide](https://github.com/allenai/wiki/wiki/Python) and it recommends using...

schmmd

This is similar to tuple instance.

nikett

Tests and fixes to new data api

2

DeNeutoy

Currently, if you want to load a model trained on one dataset and continue training, you can kind of do that, but it will end up overwriting the model that...

matt-gardner

Documentation for doing model parallelism on multiple GPUs

4

With dropping theano support, it should be easy to make our models use multiple GPUs, not just with batch parallelism, and to put some parts of the model on the...

matt-gardner

Unpin Sphinx version

1

See #362. We had to pin the sphinx version because of a conflict with how we're using python's logging functionality. We need to fix this somehow.

matt-gardner

[WIP] possible way to get READMEs into docs

6

Here is a possible way to get the .md files in the docs - not ideal to have to convert them and it also means we can't use relative github...

DeNeutoy

This would cut down pre-processing time, at the expense of having to make sure you're using the right vocabulary files and such. It would probably also make some of the...

matt-gardner

Avoid instantiating huge tensors as input to similarity functions

4

I'm not sure how this would work, really, but it takes a whole lot of memory to do it like we do it, tiling everything and then doing elementwise multiplication....

matt-gardner