g-simmons 289G_NLP_project_FQ2020 issues

g-simmons / 289G_NLP_project_FQ2020

0 stars 1 forks source link

issues

Newest

Newest Most commented Recently updated Oldest Least commented Least recently updated

fix testing

#61 g-simmons closed 3 years ago
0
added test_epoch_end to get layer-wise loss.

#60 fangzhouli opened 3 years ago
0
fix BERT

#59 g-simmons closed 3 years ago
0
Use multiple averaging strategies in val and test to calculate metrics

#58 g-simmons closed 3 years ago
1
added early stoppoing

#57 fangzhouli opened 3 years ago
0
add nll_loss_weight as a variable parameter

#56 g-simmons closed 3 years ago
0
Refactor bert

#55 g-simmons closed 3 years ago
0
Parametrize training script

#54 g-simmons closed 3 years ago
0
reduce word embedding dim back to 256

#53 g-simmons closed 3 years ago
0
Bert stopping

#52 msarmadsaeed closed 3 years ago
2
added a new random baseline; randomness is based on the true label's …

#51 sajia28 closed 3 years ago
0
debugged token splitter

#50 fangzhouli closed 3 years ago
0
test on the test split after completing training

#49 g-simmons closed 3 years ago
0
Tmp

#48 fangzhouli closed 3 years ago
0
updated logging

#47 g-simmons closed 3 years ago
0
updated logging

#46 g-simmons closed 3 years ago
0
map steps to epochs, log to wandb

#45 g-simmons closed 3 years ago
1
Fixes to BERT

#44 g-simmons closed 3 years ago
0
[HIGH] log precision and recall to wandb

#43 g-simmons closed 3 years ago
0
[MED] Reimplement prediction masking based on previous predictions

#42 g-simmons opened 3 years ago
0
[LOW] Error propagation analysis - implement Guided training

#41 g-simmons opened 3 years ago
0
[LOW] Distribution of (actual_positives / number of candidates) per sample

#40 g-simmons opened 3 years ago
0
Updated logging to wandb

#39 g-simmons closed 3 years ago
0
[HIGH] all-negative F1 baseline undef, what is a reasonable baseline?

#38 g-simmons opened 3 years ago
0
[HIGH] Fix validation naive accuracy

#37 g-simmons closed 3 years ago
0
[LOW] Improvements to model checkpointing

#36 g-simmons opened 3 years ago
0
[LOW] Stop BERT fine-tuning after a couple epochs

#35 g-simmons closed 3 years ago
1
[HIGHEST] Check/fix BERT encodings so they align with entity_spans

#34 g-simmons closed 3 years ago
1
[HIGH] Evaluate performance metrics on heldout set after training is complete and store to wandb

#33 g-simmons closed 3 years ago
1
Evaluate same metrics as INN paper at validation step

#32 g-simmons closed 3 years ago
0
Save model files to wandb

#31 g-simmons closed 3 years ago
0
efficient training for DAG-LSTM

#30 g-simmons closed 3 years ago
0
BERT incorporated

#29 g-simmons closed 3 years ago
0
weights & biases logging

#28 g-simmons closed 3 years ago
1
Fang gpu refined

#27 g-simmons closed 3 years ago
0
log to wandb

#26 g-simmons closed 3 years ago
0
Sammy gpu tensorboard

#25 sajia28 closed 3 years ago
0
Parallelize data processing, update training script

#24 g-simmons closed 3 years ago
0
Revert "Revert "updated data format, model training""

#23 sajia28 closed 3 years ago
0
Revert "updated data format, model training"

#22 sajia28 closed 3 years ago
0
Sammy batch learning

#21 sajia28 closed 3 years ago
0
Run Code and Check Accuracy/Error

#20 sajia28 closed 3 years ago
0
Parallelize Code

#19 sajia28 closed 3 years ago
0
Move Code to Google Cloud

#18 sajia28 closed 3 years ago
0
DAG-LSTM cell state exploding in a single forward pass, clipping may be limiting gradient flow?

#17 g-simmons closed 3 years ago
1
updated data format, model training

#16 g-simmons closed 3 years ago
0
Create a Plot of the Loss Over Time

#15 sajia28 closed 3 years ago
0
[DRAFT] model training

#14 g-simmons closed 3 years ago
0
Perform a Code Review

#13 sajia28 closed 3 years ago
0
Put the Training Examples into Batches

#12 sajia28 closed 3 years ago
0