Added batch iteration into input and target variables (#285) #288
Conversation
torchtext/data/batch.py
Outdated
| @@ -58,6 +62,20 @@ def __str__(self): | |||
| def __len__(self): | |||
| return self.batch_size | |||
|
|
|||
| def _get_vars(self, target): | |||
This comment was marked as off-topic.
This comment was marked as off-topic.
Sorry, something went wrong.
torchtext/data/batch.py
Outdated
| def _get_vars(self, target): | ||
| fields = self.target_fields if target else self.input_fields | ||
| vars = tuple(getattr(self, f) for f in fields) | ||
| if len(vars) == 0: |
This comment was marked as off-topic.
This comment was marked as off-topic.
Sorry, something went wrong.
|
Hi, thanks for the review! I changed the code to reflect my intent more clearly, I think this should resolve the above problems, but if anything can be improved please tell me :) |
|
Awesome, thanks, everything seems good now! |
|
Hey everyone, I ran into the same issue as @raheelqader in issue 422, with the "is_target" attribute being referenced for a "RawField" instance. I'd like to suggest we add an "is_target" attribute to "RawField" and set it to False by default. I'd like to contribute, but this is my first time. Should I just go ahead and do a PR once I've made the fix? Thanks! Looking forward to contributing. |
|
Hey, sure, feel free to submit a PR! |
Now, the batch can be iterated into input and target variables like
where
x1tox3are inputs to a downstream model andyis the label data for this batch.