Ecosyste.ms: Issues
An open API service for providing issue and pull request metadata for open source projects.
GitHub / nyu-mll/jiant-v1-legacy issues and pull requests
#50 - [CLOSED] Fix/simplify regression code.
Issue -
State: closed - Opened by jeswan over 4 years ago
#50 - [CLOSED] Fix/simplify regression code.
Issue -
State: closed - Opened by jeswan over 4 years ago
#49 - [CLOSED] Stream data from disk instead of loading entirely to RAM
Issue -
State: closed - Opened by jeswan over 4 years ago
- 6 comments
#49 - [CLOSED] Stream data from disk instead of loading entirely to RAM
Issue -
State: closed - Opened by jeswan over 4 years ago
- 6 comments
#48 - [CLOSED] De-dupe path scripts
Issue -
State: closed - Opened by jeswan over 4 years ago
#48 - [CLOSED] De-dupe path scripts
Issue -
State: closed - Opened by jeswan over 4 years ago
#47 - [CLOSED] STS-B classifier is broken
Issue -
State: closed - Opened by jeswan over 4 years ago
#47 - [CLOSED] STS-B classifier is broken
Issue -
State: closed - Opened by jeswan over 4 years ago
#46 - [CLOSED] Update readme for new configs
Issue -
State: closed - Opened by jeswan over 4 years ago
#46 - [CLOSED] Update readme for new configs
Issue -
State: closed - Opened by jeswan over 4 years ago
#45 - [CLOSED] Config
Issue -
State: closed - Opened by jeswan over 4 years ago
#45 - [CLOSED] Config
Issue -
State: closed - Opened by jeswan over 4 years ago
#44 - [CLOSED] Best checkpoint logic changes.
Issue -
State: closed - Opened by jeswan over 4 years ago
#44 - [CLOSED] Best checkpoint logic changes.
Issue -
State: closed - Opened by jeswan over 4 years ago
#43 - [CLOSED] Configuration files using pyhocon
Issue -
State: closed - Opened by jeswan over 4 years ago
#42 - [CLOSED] Config file system to replace multitude of flags
Issue -
State: closed - Opened by jeswan over 4 years ago
- 5 comments
#42 - [CLOSED] Config file system to replace multitude of flags
Issue -
State: closed - Opened by jeswan over 4 years ago
- 5 comments
#41 - [CLOSED] Add MT pretraining task
Issue -
State: closed - Opened by jeswan over 4 years ago
#41 - [CLOSED] Add MT pretraining task
Issue -
State: closed - Opened by jeswan over 4 years ago
#40 - [CLOSED] include options for specifying classifier to use for probing tasks
Issue -
State: closed - Opened by jeswan over 4 years ago
- 9 comments
#40 - [CLOSED] include options for specifying classifier to use for probing tasks
Issue -
State: closed - Opened by jeswan over 4 years ago
- 9 comments
#39 - [CLOSED] Probing infra
Issue -
State: closed - Opened by jeswan over 4 years ago
#39 - [CLOSED] Probing infra
Issue -
State: closed - Opened by jeswan over 4 years ago
#38 - [CLOSED] Training for multiple tasks in the eval phase is broken.
Issue -
State: closed - Opened by jeswan over 4 years ago
- 5 comments
#38 - [CLOSED] Training for multiple tasks in the eval phase is broken.
Issue -
State: closed - Opened by jeswan over 4 years ago
- 5 comments
#37 - [CLOSED] Not possible to add a new eval task when loading an old pretrained model.
Issue -
State: closed - Opened by jeswan over 4 years ago
- 1 comment
#37 - [CLOSED] Not possible to add a new eval task when loading an old pretrained model.
Issue -
State: closed - Opened by jeswan over 4 years ago
- 1 comment
#36 - [CLOSED] Separate checkpoint names for main training phase and eval phase
Issue -
State: closed - Opened by jeswan over 4 years ago
#36 - [CLOSED] Separate checkpoint names for main training phase and eval phase
Issue -
State: closed - Opened by jeswan over 4 years ago
#35 - [CLOSED] JOCI pretraining task
Issue -
State: closed - Opened by jeswan over 4 years ago
- 1 comment
#34 - [CLOSED] refactor main so all blocks of code are opt-in
Issue -
State: closed - Opened by jeswan over 4 years ago
- 1 comment
#34 - [CLOSED] refactor main so all blocks of code are opt-in
Issue -
State: closed - Opened by jeswan over 4 years ago
- 1 comment
#33 - [CLOSED] Train a GLUE MTL run for use in analysis fine-tuning.
Issue -
State: closed - Opened by jeswan over 4 years ago
- 2 comments
#33 - [CLOSED] Train a GLUE MTL run for use in analysis fine-tuning.
Issue -
State: closed - Opened by jeswan over 4 years ago
- 2 comments
#32 - [CLOSED] Double check that early stopping strategy is sane for single-task GLUE training
Issue -
State: closed - Opened by jeswan over 4 years ago
- 1 comment
#32 - [CLOSED] Double check that early stopping strategy is sane for single-task GLUE training
Issue -
State: closed - Opened by jeswan over 4 years ago
- 1 comment
#31 - [CLOSED] Dis sent
Issue -
State: closed - Opened by jeswan over 4 years ago
#31 - [CLOSED] Dis sent
Issue -
State: closed - Opened by jeswan over 4 years ago
#30 - [CLOSED] Use ELMo character handling in all experiments, set as default
Issue -
State: closed - Opened by jeswan over 4 years ago
- 7 comments
#30 - [CLOSED] Use ELMo character handling in all experiments, set as default
Issue -
State: closed - Opened by jeswan over 4 years ago
- 7 comments
#29 - [CLOSED] Add a projection layer before max pooling
Issue -
State: closed - Opened by jeswan over 4 years ago
- 2 comments
#29 - [CLOSED] Add a projection layer before max pooling
Issue -
State: closed - Opened by jeswan over 4 years ago
- 2 comments
#28 - [CLOSED] Make checkpointing during target-task training less wonky.
Issue -
State: closed - Opened by jeswan over 4 years ago
- 1 comment
#28 - [CLOSED] Make checkpointing during target-task training less wonky.
Issue -
State: closed - Opened by jeswan over 4 years ago
- 1 comment
#27 - [CLOSED] Flag to control whether or not loaded/pretrained model is re-trained …
Issue -
State: closed - Opened by jeswan over 4 years ago
#27 - [CLOSED] Flag to control whether or not loaded/pretrained model is re-trained …
Issue -
State: closed - Opened by jeswan over 4 years ago
#26 - [CLOSED] set up command line invocation for probing with NLI-style data
Issue -
State: closed - Opened by jeswan over 4 years ago
- 5 comments
#26 - [CLOSED] set up command line invocation for probing with NLI-style data
Issue -
State: closed - Opened by jeswan over 4 years ago
- 5 comments
#25 - [CLOSED] pretrained cnns, grounded tasks plus cnn encoder
Issue -
State: closed - Opened by jeswan over 4 years ago
#24 - [CLOSED] Don't automatically delete checkpoints, even when it might be useful.
Issue -
State: closed - Opened by jeswan over 4 years ago
#23 - [CLOSED] Absolute paths & scripts for GCP
Issue -
State: closed - Opened by jeswan over 4 years ago
- 1 comment
#22 - [CLOSED] bidirectional rnn language model
Issue -
State: closed - Opened by jeswan over 4 years ago
- 10 comments
#21 - [CLOSED] find sensible transformer params
Issue -
State: closed - Opened by jeswan over 4 years ago
- 12 comments
#20 - [CLOSED] implement pre-pooling linear projection
Issue -
State: closed - Opened by jeswan over 4 years ago
Labels: low-priority
#19 - [CLOSED] training eval-only tasks will erase model checkpoints create during training
Issue -
State: closed - Opened by jeswan over 4 years ago
- 4 comments
#18 - [CLOSED] separate training parameters for train tasks and eval tasks
Issue -
State: closed - Opened by jeswan over 4 years ago
- 7 comments
#17 - [CLOSED] modify AllenNLP RNN decoder to fit in framework (sequence generation tasks)
Issue -
State: closed - Opened by jeswan over 4 years ago
- 3 comments
#16 - [CLOSED] writeup options and trainer setup in README
Issue -
State: closed - Opened by jeswan over 4 years ago
#15 - [CLOSED] Learn linear combinations of core LSTM weights
Issue -
State: closed - Opened by jeswan over 4 years ago
- 4 comments
#14 - [CLOSED] Switch to AllenNLP ELMoTokenEmbedder
Issue -
State: closed - Opened by jeswan over 4 years ago
- 1 comment
#13 - [CLOSED] Fix paths & allow for read-only data_dir
Issue -
State: closed - Opened by jeswan over 4 years ago
- 1 comment
#12 - [CLOSED] Crashing when trying to use only char embs in demo.sh
Issue -
State: closed - Opened by jeswan over 4 years ago
- 1 comment
#11 - [CLOSED] Flag-guard import pdb
Issue -
State: closed - Opened by jeswan over 4 years ago
- 1 comment
#10 - [CLOSED] Masked transformer
Issue -
State: closed - Opened by jeswan over 4 years ago
- 1 comment
#9 - [CLOSED] Clarify `epoch`
Issue -
State: closed - Opened by jeswan over 4 years ago
- 1 comment
#8 - [CLOSED] Delete checkpoints when reinitializing.
Issue -
State: closed - Opened by jeswan over 4 years ago
#7 - [CLOSED] Save optimizer and LR scheduler state
Issue -
State: closed - Opened by jeswan over 4 years ago
- 1 comment
#6 - [CLOSED] Can now rebuild vocabulary before loading model.
Issue -
State: closed - Opened by jeswan over 4 years ago
#5 - [CLOSED] Delete checkpoints when starting training from scratch
Issue -
State: closed - Opened by jeswan over 4 years ago
- 2 comments
#4 - [CLOSED] Don't load the test set unless you need it.
Issue -
State: closed - Opened by jeswan over 4 years ago
- 1 comment
#3 - [CLOSED] fastText embeddings
Issue -
State: closed - Opened by jeswan over 4 years ago
- 1 comment
#2 - [CLOSED] PyTorch clip warning
Issue -
State: closed - Opened by jeswan over 4 years ago
#1 - [CLOSED] AllenNLP vocabulary warning.
Issue -
State: closed - Opened by jeswan over 4 years ago
- 1 comment