Dev v0.6.0 #167

tmills · 2023-08-03T20:26:13Z

No description provided.

…file

Move versioning into cnlpt.__version__

…gging+classification.

…ings that came up in code review.

…sue-138-datasets Issue 138 datasets

merge docs and github actions changes into dev-v0.6.0

… no tasks at cli.

…ut time.

… testing.

Realign baseline models with new task flow

explosion/spaCy#12659

Add support for loading local encoders for fine-tuning

… write out during training.

…Language/cnlp_transformers into dev-v0.6.0

Drop support for Python 3.7

Had to add extra conditional for non-training branch. Adjusted offset for layer parameter so that -1 means the last layer.

…erAttn Grab chunk attentions so that we can visualize hierarchical attention.

… class.

* Update CONTRIBUTING.md with base branch info * Added dev dockerfiles --------- Co-authored-by: Angus L'Herrou <[email protected]>

…tion

Add requirements for transform_prot

…ve_load_fix save config.json in training; actually loads best model before prediction

tmills and others added 30 commits April 18, 2023 14:52

Raise exception if the user actually passes in multiple datasets.

3bb110c

Add label space to model config.

f1b3bef

move versioning into cnlpt.__version__

8ca6131

Bump to 0.6.0

7339455

Fix version bump instructions in CONTRIBUTING.md

ae0ffad

Remove src/cnlpt/VERSION from MANIFEST.in and delete resulting empty …

c0020c7

…file

Merge pull request #139 from angus-lherrou/version_attr

4d27d28

Move versioning into cnlpt.__version__

Add backwards compatible config reading to temporal rest api.

aad3834

WIP for multi-task/multi-dataset training. Still doesn't work with ta…

bbafa2f

…gging+classification.

WIP: works to do tagging with classification.

c7d95ae

WIP but everything is working except hierarchical model.

594f97c

Update HierarchicalModel to use moduledict instead of arrays.

ee0fdce

Add some missing function comments and fix some logging messages - th…

2ccd6ce

…ings that came up in code review.

Merge pull request #142 from Machine-Learning-for-Medical-Language/is…

51f568f

…sue-138-datasets Issue 138 datasets

Add version number to output models.

8a64f19

Merge pull request #144 from Machine-Learning-for-Medical-Language/main

a5881e3

merge docs and github actions changes into dev-v0.6.0

Add method for getting official task list in case where user supplies…

e1f6710

… no tasks at cli.

Get ordered tsak names from command line to ensure same order at outp…

df9ca02

…ut time.

realign baseline models with new task flow

0317511

Fix for python 3.7 syntax

adfa66d

Fixed other spot where I was using >3.7 syntax.

95ae854

Add cleanup code to delete all the pytorch_model.bin generated during…

b1e3393

… testing.

Merge pull request #150 from angus-lherrou/149-num-labels-list

3aa9743

Realign baseline models with new task flow

Allow for fine-tuning pre-trianed hierarchical models.

4d7ad5b

Add pointer to new negation model for v0.6.0.

0e501c5

Update temporal rest to get task ordering from model config file.

100cb52

Update temporal test for thyme2 model.

2613560

Upgrade pydantic to ~=1.10.8

f61bb67

explosion/spaCy#12659

make all CnlpConfig arguments keyword-only

5f59c06

add new functions is_cnlpt_model and encoder_inferred

c45a32e

tmills and others added 29 commits July 18, 2023 13:50

Update all rest apis to use get_dataset without task argument.

164b0b2

Merge branch 'dev-v0.6.0' into non-hub-encoders

1fe01d3

Merge pull request #152 from angus-lherrou/non-hub-encoders

2694b4a

Add support for loading local encoders for fine-tuning

Add cnn rest. puts some demands on cnn models that we don't currently…

e063fcf

… write out during training.

Merge branch 'dev-v0.6.0' of github.com:Machine-Learning-for-Medical-…

fd27f63

…Language/cnlp_transformers into dev-v0.6.0

Fix issue with not using the best model for writing predictions.

8023ac6

Require python 3.8+ for transform_prot.py

3dc8b09

Drop support for Python 3.7

7a87580

Merge pull request #161 from angus-lherrou/drop-37

91bd23b

Drop support for Python 3.7

Merge branch 'dev-v0.6.0' into scispacy

50dc061

Dropped support for Python 3.7: remove version gating

21d365a

Add platform info to transform_prot error message

ed6be90

Revise platform note and add platform note to chemprot README

fddbe5c

Grab chunk attentions so that we can visualize hierarchical attention.

2285947

Had to add extra conditional for non-training branch. Adjusted offset for layer parameter so that -1 means the last layer.

Merge pull request #162 from Machine-Learning-for-Medical-Language/hi…

23a6aa3

…erAttn Grab chunk attentions so that we can visualize hierarchical attention.

Fix to Hierarchical model to pass back chunk attentions in a new data…

633e759

… class.

Added dev dockerfiles (#163)

3fc1ea0

* Update CONTRIBUTING.md with base branch info * Added dev dockerfiles --------- Co-authored-by: Angus L'Herrou <[email protected]>

Remove labels argument from get_dataset

b234c2f

save config.json in training; actually loads best model before predic…

988d5fc

…tion

Remove docker files from dev for this release.

c9fb3df

Remove dev docker from compose file.

e0ae473

Update to add commands for new rest endpoints.

4756f7e

New api endpoints for "current" and "termExists" classifiers.

cfdbd61

Fix to add function to call for new rest endpoints.

35f7a66

Merge pull request #153 from angus-lherrou/scispacy

7c74766

Add requirements for transform_prot

re-add cnn_rest.py

877e5ce

Only write out config file specially for cnn and lstm models.

7b69e58

Merge pull request #164 from Machine-Learning-for-Medical-Language/sa…

cfa8959

…ve_load_fix save config.json in training; actually loads best model before prediction

gate transform_prot to Linux

6d9e11d

tmills merged commit db3612b into main Aug 3, 2023
9 checks passed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Dev v0.6.0 #167

Dev v0.6.0 #167

tmills commented Aug 3, 2023

Dev v0.6.0 #167

Dev v0.6.0 #167

Conversation

tmills commented Aug 3, 2023