clay-lab transductions issues

clay-lab / transductions

A PyTorch framework for creating, running, and reproducing experiments on seq2seq models.

6 stars 3 forks source link

issues

Newest

Newest Most commented Recently updated Oldest Least commented Least recently updated

Ignore <pad> when computing cross-entropy loss and sequence accuracy.

#65 bdusell opened 2 years ago
0
Padding token is not ignored during training and evaluation

#64 bdusell opened 2 years ago
0
Conda gets stuck on "Solving environment" phase

#63 bdusell opened 2 years ago
0
Sort imports

#62 jopetty closed 2 years ago
0
Add docker support and pre-commit hooks

#61 jopetty closed 2 years ago
0
Add per-batch counting for validation tests

#60 jopetty opened 3 years ago
0
Add option to not use early stopping

#59 jopetty opened 3 years ago
0
Why are reported losses in TQDM monotonically increasing

#58 jopetty opened 3 years ago
0
BERT

#57 jopetty opened 3 years ago
2
Get multirun working

#56 jopetty closed 3 years ago
0
Delete all processed data when overwriting dataset

#55 jopetty closed 3 years ago
1
Change dirname of models to be a parameter specified in config

#54 jopetty closed 3 years ago
1
Support multiple transformation tokens

#53 jopetty opened 3 years ago
0
Docker image

#52 jopetty closed 2 years ago
1
Move away from legacy torch text code

#51 jopetty opened 3 years ago
0
Make teacher-forcing ratio configurable in yaml

#50 jopetty closed 3 years ago
1
Different embedding and hidden sizes don't work for encoders and decoders

#49 jopetty closed 3 years ago
3
Rethink property naming conventions

#48 jopetty opened 3 years ago
0
Should it be stoi['pad'] or stoi['<pad>']

#47 jopetty closed 3 years ago
1
Nondeterminism

#46 jopetty opened 3 years ago
0
Transformer

#45 jopetty closed 3 years ago
0
Merge in changes to master

#44 jopetty closed 3 years ago
0
Why aren't hydra configs getting type'd right?

#43 jopetty closed 3 years ago
2
Hydra

#42 jopetty closed 3 years ago
0
Restructure YAML configs

#41 jopetty closed 3 years ago
2
Add schema for configurations

#40 jopetty opened 3 years ago
1
Add dependency list for easy installation

#39 jopetty closed 3 years ago
2
Add checkpoint logic

#38 jopetty closed 3 years ago
1
Add attention

#37 jopetty closed 3 years ago
3
Add teacher forcing

#36 jopetty closed 3 years ago
1
Add capability to perform multiple runs, possibly with a parameter sweep

#35 jopetty opened 3 years ago
3
Add hook for peeking at arbitrary properties of the model mid-pass

#34 jopetty opened 3 years ago
0
Support moving the Transform field between Source and Target

#33 jopetty closed 3 years ago
1
Allow multiple layers in units

#32 jopetty closed 3 years ago
2
Add metrics

#31 jopetty opened 3 years ago
4
Fix TransductionDataset string representation

#30 jopetty closed 3 years ago
1
Make the decoders do something

#29 jopetty closed 3 years ago
1
Why are the embedding dimensions different between the old and new versions?

#28 jopetty closed 3 years ago
2
Separate TGT field for training and test data

#27 NoahAmsel opened 4 years ago
0
Add parameter for max depth

#26 NoahAmsel opened 4 years ago
0
Add cuda magic to GRUTridentDecoder

#25 NoahAmsel opened 4 years ago
0
Remove old models from models.py

#24 NoahAmsel closed 4 years ago
0
More flexibility in main -- mix and match tree input sequence with sequential output sequence

#23 NoahAmsel closed 4 years ago
1
Simplify file structure

#22 NoahAmsel closed 4 years ago
0
Make different arithmetic task data files broken down by length

#21 NoahAmsel opened 4 years ago
1
Update functionality of Tree decoders

#20 NoahAmsel opened 4 years ago
3
Fix Single token accuracy meter

#19 NoahAmsel opened 4 years ago
0
Does test code fail if prediction is longer than target?

#18 NoahAmsel closed 4 years ago
2
Can we mere in the repl/test code now?

#17 jopetty closed 4 years ago
0
Using annotation tag

#16 NoahAmsel opened 4 years ago
2