Closed khatchad closed 6 months ago
Sorry, may need to make some changes here.
I think I'm hitting an unrelated bug here.
I made it a bit more robust regarding the "for each" statement.
Thanks, @msridhar. While this works, as I try to add more API, I realize that I've treated this particular API procedurally. It should return a Dataset
object, which subsequent API calls can be invoked on. I'll most likely change some of this in the next PR to accommodate the Dataset
object being returned.
Thanks, @msridhar. While this works, as I try to add more API, I realize that I've treated this particular API procedurally. It should return a
Dataset
object, which subsequent API calls can be invoked on. I'll most likely change some of this in the next PR to accommodate theDataset
object being returned.
I've done that here: https://github.com/ponder-lab/ML/commit/c0bc8b25ba8e51403a5dd87f4edf08ed08bac7f7.
Now, I think we can add more methods to: https://github.com/ponder-lab/ML/commit/c0bc8b25ba8e51403a5dd87f4edf08ed08bac7f7#diff-679e457ba33ce5658ad9444fd41a8839e363bac4c0a268b27a86bbaad0c22056R340-R345.
Dataset
s are used in chains that apply transformations, and I think we can add the transforming methods there.
The tf2_test_dataset.py
test works, but that commit is now failing on the newly added test which applies a shuffle()
transformation: https://github.com/ponder-lab/ML/commit/1fc1d3df9ef4d50f0fc4e5aae54063307adc5329. I'll work on that now.
Basic fix for #89; more APIs to come.