Skip to content

Move aux stats file to be part of soft loss#120

Open
gabrieletijunaityte wants to merge 174 commits into
developfrom
train/alignment-improvements
Open

Move aux stats file to be part of soft loss#120
gabrieletijunaityte wants to merge 174 commits into
developfrom
train/alignment-improvements

Conversation

@gabrieletijunaityte

@gabrieletijunaityte gabrieletijunaityte commented Jun 19, 2026 •
edited
Loading

Copy link
Copy Markdown
Contributor

What does this PR do?

Move aux stats file to be part of soft loss
Depends on PR #118

Before submitting

  • Did you make sure title is self-explanatory and the description concisely explains the PR?
  • Did you make sure your PR does only one thing, instead of bundling different changes together?
  • Did you list all the breaking changes introduced by this pull request?
  • Did you test your PR locally with pytest command?

gabrieletijunaityte and others added 30 commits May 27, 2026 13:26
@gabrieletijunaityte
Add validation run of the best ckpt
a03e98c
@gabrieletijunaityte
Merge branch 'develop' of github.com:WUR-AI/aether into train/alignment
34d35df
@gabrieletijunaityte
Add avr embedding modalities
1428e92
@gabrieletijunaityte
Make search in power of 2, plus add max batch size
e6c141e
@gabrieletijunaityte
Add average embedding modalities
8e035d8
@gabrieletijunaityte
Add no gradient computation option if model is not trained
695bd3c
@gabrieletijunaityte
Allow identity encoder for the average embeddings
54bb273
@gabrieletijunaityte
Compute average embeddings into csv
fea48d5
@gabrieletijunaityte
Adapt identity encoder for average embeddings
0d1c4cd
@gabrieletijunaityte
Avoid gradient computation where not needed
90ebd95
@gabrieletijunaityte
Inference -> no_grad
545c3ce
@vdplasthijs @gabrieletijunaityte
Download GEE data in parallel
4fa833d
@vdplasthijs @gabrieletijunaityte
Sample unlabelled aux data for UK s2bms
1e0974e
@gabrieletijunaityte
Add ability to load unlabelled data
ca46643 
# Conflicts:
#	src/data/butterfly_dataset.py
@vdplasthijs @gabrieletijunaityte
adding unlabbelled s2bms data, bypassing hook
af4f689
@vdplasthijs @gabrieletijunaityte
Load unlabelled and original s2bms together
1249511
@vdplasthijs @gabrieletijunaityte
adding unlabbelled splits, bypassing hook
14ed922
@vdplasthijs @gabrieletijunaityte
adding unlabbelled s2bms data splits, bypassing hook
a5b35b3
@gabrieletijunaityte
Add avr embedding modalities
bee388c 
# Conflicts:
#	src/data/base_dataset.py
@gabrieletijunaityte
Add setup flag
991c604
@gabrieletijunaityte
Re do unlabelled split
9052b2f
@gabrieletijunaityte
Add size print statements
e991dc9
@gabrieletijunaityte
Make format hook changes
e726b1c
@gabrieletijunaityte
Remove full model name printing
325f0ee
@gabrieletijunaityte
Hook security
a67f56a
@gabrieletijunaityte
Fix logging
7d80c9b
@gabrieletijunaityte
Fix: getting size
808d876
@gabrieletijunaityte
Hook fixes
e349b5a
@gabrieletijunaityte
Save average embeddings to csv
633ff31
@gabrieletijunaityte
emb_ reading for csv of avr embeds
57cc2cb
gabrieletijunaityte and others added 28 commits June 17, 2026 08:31
@gabrieletijunaityte
make geo and text adapters not logged in hparams
936f403
@gabrieletijunaityte
Inspect text encoders
8b905b8
@gabrieletijunaityte
Fix LLM downloading, flatten captions for efficient forward pass
b256c38
@gabrieletijunaityte
Add aux stats calculation
2f94ecc
@gabrieletijunaityte
Create soft contrastive loss
0f10083
@gabrieletijunaityte
Implement loss pass
106d5bc
@gabrieletijunaityte
Dimension miss match warning
af0cee3
@vdplasthijs @gabrieletijunaityte
Add ability to load unlabelled data
7b6c009 
# Conflicts:
#	configs/data/butterfly_coords_text_unlabelled.yaml
#	src/data/base_datamodule.py
#	src/data/base_dataset.py
#	src/data/butterfly_dataset.py
@gabrieletijunaityte
New datasplits
286a975
@gabrieletijunaityte
Split after getting missing data, in the setup() call
92e7d2e
@gabrieletijunaityte
Add ability to load unlabelled data
b9081fd 
# Conflicts:
#	src/data/butterfly_dataset.py
@gabrieletijunaityte
Re do unlabelled split
88e2b5e
@gabrieletijunaityte
Add size print statements
9f5fe04
@gabrieletijunaityte
New data splits and their explanations
ea5fc18
@gabrieletijunaityte
fix the splits
b51f91c
@gabrieletijunaityte
Merge remote-tracking branch 'origin/train/alignment' into train/alig…
8596566 
…nment

# Conflicts:
#	notebooks/09-GT-aef-tessera-datasplit-filtering.ipynb
@gabrieletijunaityte
Update stats based on train split only
1b323f0
@gabrieletijunaityte
Fix error of missing wandb logger
eb32ee0
@gabrieletijunaityte
Fix argument order mix-up
8360805
@gabrieletijunaityte
Add aux stats computation notebook
a030e60
@gabrieletijunaityte
Add average embedding computation notebook
120455b
@gabrieletijunaityte
Fix config for the tests
0610a49
@gabrieletijunaityte
Fix tests
ac9b5df
@gabrieletijunaityte
Correct explanations
f7f352d
@gabrieletijunaityte
Remove duplicate
9549ddf
@gabrieletijunaityte
Automatic experiment name generation
8ff3ce1
@gabrieletijunaityte
Automatic experiment name generation for prediction models
40f1a18
@gabrieletijunaityte
Move aux col statistics from datamodule to loss function
0738579
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

@vdplasthijs vdplasthijs Awaiting requested review from vdplasthijs

At least 1 approving review is required to merge this pull request.

Assignees

No one assigned

Labels

None yet

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

2 participants

@gabrieletijunaityte @vdplasthijs