Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Hugging Face Model Training #45

Merged
merged 66 commits into from
Jun 21, 2023
Merged

Hugging Face Model Training #45

merged 66 commits into from
Jun 21, 2023

Conversation

tieandrews
Copy link
Collaborator

This PR contains the proposed final code for running the model training using Huggingface. The primary parts in it are:

  • the run_roberta_base_training.sh that can be used to recreate the model results in the final report
  • updated model evaluation code including tests
  • documentation for how to run the training under the readme
  • initial documentation on setting up LabelStudio

@tieandrews
feat: added entity label file loading function
281cb57
@tieandrews
feat: script to coordinate training process
79fd36f
@tieandrews
feat: bash script to run training and set params
f289c36
@tieandrews
feat: script to generate hf formatted data
ca9c5fa
@tieandrews
docs: updated ner_training target name
dd3696f
@tieandrews
docs: model training readme
7f4ac68
@tieandrews
feat: automated utils to evaluate trained models
c64e778
@tieandrews
feat: bash script to run evaluation on checkpoint
7361e91
@tieandrews
bug: added _init_ to src to make module
26b20ce
@tieandrews
bug: token based evaluation same as entity based
8d69aa6 
Lists passed to funcion were modifying data outside the
function. This was causing the evaluation to be the same
for both token and entity based evaluation.
@tieandrews
feat: added new return objects
2cb379a
@tieandrews
docs: updated docs to match new input format
eff149e
@tieandrews
feat: improved error checking and logging
df022a7
@tieandrews
feat: cleaned up imports and constants
234aa3d
@tieandrews
feat: last imports cleaned
a5348f8
@tieandrews
feat: added custom metrics to mlflow logging
0944ea3
@tieandrews
feat: model training bash script
9cf2140
@tieandrews
feat: basic notebook to use on colab to train
1a18d96
@tieandrews
docs: initial commit
2357286
@tieandrews
docs: results and models directory setup
9ea5abe
@tieandrews
feat: removed unused label2id object
9192ad8
@tieandrews
feat: separated folder location setting
0153d73
@tieandrews
bug: setup individual copied objects for pred/true
8e31500
@tieandrews
bug: added defined max token length
1ac2bbc
@tieandrews
bug: added quotes for paths with spaces in colab
89be2df
@tieandrews
bug: fixed labelled file location with spaces
4534ba2
@tieandrews
Merge remote-tracking branch 'origin/dev' into 22-fine-tune-allenai-s…
f2bf1ac 
…pecter2-model-for-ner
@tieandrews
feat: rename and update standard params
4cbfaaf
@tieandrews
docs: switch default data folder
5ca03dd
@tieandrews
docs: clean up and make better defaults
e7faccb
@tieandrews
bug: moved docopt to main
31c2fc8
@tieandrews
tests: initial basic tests
c135b09
@tieandrews
feat: colab train notebook autoreload used
92df812
@tieandrews
docs: delete model card of specter2
da53411
@tieandrews
tests: added tests for calculate/plot methods
67088c4
@tieandrews
feat: added stride to generating overlap windows
ca10b58
@tieandrews
feat: added early stopping with patience 5
37fefaf
@tieandrews
feat: added warmup and seed
5357087
@tieandrews
bug: added load_best_model_at_end for early stop
9521d9c
@tieandrews
bug: added control for when mlflow diasbled
a9572c6
@tieandrews
feat: updated preprocessing stride length
babdb8b
@tieandrews
feat: updated grouped_entities to aggregationstrat
50ac915
@tieandrews
feat: baseline test results
af0e2d4
@tieandrews
docs: roberta finetune v3 results
bc669ec
@tieandrews
feat: bert-multilanguage results
386c058
@tieandrews
feat: specter2 results
b84e7ae
@tieandrews
docs: removed results folder
7215a39
@tieandrews
docs: HF model training readme update
129fac0
@tieandrews
docs: update colab notebook
1c617e2
@tieandrews
feat: final roberta model training script
743780b
@tieandrews
feat: automated baseline evaluation script
81bc738
@tieandrews
docs: labelsetudio setup guide
8d99f29
@tieandrews tieandrews added the NER Tasks related to NER tasks label Jun 21, 2023
@tieandrews tieandrews self-assigned this Jun 21, 2023
shaunhutch
shaunhutch approved these changes Jun 21, 2023
View reviewed changes
Copy link
Collaborator

@shaunhutch shaunhutch left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Looks great! Awesome work

@shaunhutch shaunhutch merged commit 6d43182 into dev Jun 21, 2023
1 check passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers

@shaunhutch shaunhutch shaunhutch approved these changes

@brabbit61 brabbit61 Awaiting requested review from brabbit61

Assignees

@tieandrews tieandrews

Labels
NER Tasks related to NER tasks
Projects
Neotoma's Metadata Extractor
Status: Done
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.
2 participants
@tieandrews @shaunhutch