Linear Chain CRF with SGD method #11
Open
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Hi,
I tried to implement linear chain CRF with SGD method based on these two paper:
For performance, I tested it on CoNLL 2000 shared task and got:
I think the biggest problem in my implementation is the training time. I will try to work on parallel training or improve the cython code.
Maybe you can review it and give me some feedback.
Any suggestions are welcomed.
Thanks!