Wl coref chains #1309
Merged
Wl coref chains #1309
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
AngledLuffa
force-pushed
the
wl_coref_chains
branch
17 times, most recently
from
7e9ca73
to
6277d92
Compare
AngledLuffa
force-pushed
the
wl_coref_chains
branch
6 times, most recently
from
c25ce00
to
a9f51b4
Compare
AngledLuffa
force-pushed
the
wl_coref_chains
branch
2 times, most recently
from
7eec7f3
to
1be662e
Compare
AngledLuffa
force-pushed
the
wl_coref_chains
branch
2 times, most recently
from
4387221
to
77110c6
Compare
…rmer, non-peft model. We will need to check that the pipeline loads in a non-transformer world
…of the coref model
…berta no finetuning, and roberta lora finetuning. Add toml as a dependency to allow for the config usage
AngledLuffa
force-pushed
the
wl_coref_chains
branch
from
7684ac7
to
3f546d9
Compare
Loads the Coref model from a version downloaded via HuggingFace Uses a rough draft of namedtuples for coref chains, connects the chains to the document Could possibly make the chains have the text of the mentions. Adds backlinks from words to the coref chains Assigns the model to the proper device when building the coref processor Hide the CorefModel import in coref_processor to avoid importing transformers when not available
… cuda isn't available
… peft & transformers imports
Don't add a default yet to the default packages, as it is quite expensive Instead, add an optional coref model for EN to the resources Specifying "tokenize,coref" when creating a pipeline will find the optional EN model
… be used elsewhere
…use CCONJ instead of CC (upos)
…ber of steps. Helps but doesn't completely fix the problem of the models going to 0 F1
…y allow for overlapping entities
…hment into an object rather than a namedtuple, since the default encoder doesn't appear to have a way to override the decoding of a namedtuple. Only output a couple items from the coref chain as part of the json output Add a potential conllu format as well Uses unit- for singletons, start-, end-, middle- otherwise Add an index to the coref chains
…g data) if only doing eval mode. Also, turn off the noisy output by default
AngledLuffa
force-pushed
the
wl_coref_chains
branch
from
7234618
to
203f922
Compare
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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.
WIP: add English coref to the pipeline