issues
search
eth-easl
/
modyn
Modyn is a research-platform for training ML models on growing datasets.
MIT License
22
stars
3
forks
source link
issues
Newest
Newest
Most commented
Recently updated
Oldest
Least commented
Least recently updated
Add option to clean up database and temp files after pipeline
#577
MaxiBoether
opened
4 days ago
0
Update README
#576
MaxiBoether
opened
6 days ago
3
feat: Add `SIGMOD` plots
#575
robinholzi
closed
6 days ago
2
Make embeddings optional for drift
#574
MaxiBoether
opened
1 week ago
0
Remove necessity of pipeline patching for CGLM
#573
MaxiBoether
opened
1 week ago
0
Feed a copy of model config in `RHO-LOSS` Twin Model
#572
XianzheMa
closed
1 week ago
2
Allow evaluation executor to return evaluation results mixed with success and failure
#571
XianzheMa
closed
1 week ago
2
fix: Adjust tool for new eval format
#570
robinholzi
closed
1 week ago
1
feat: Add `SIGMOD` plots
#569
robinholzi
closed
1 week ago
2
feat: Add tool for patching old logfiles
#568
robinholzi
closed
1 week ago
7
Port changes from sigmod branch: drain the metric queue
#567
XianzheMa
closed
1 week ago
1
fix: Adjust aggregation tool for batch eval format
#566
robinholzi
closed
1 week ago
2
fix: Adjust analytics tool for batch eval format
#565
robinholzi
closed
1 week ago
2
feat: Add more powerful windowing strategies, warumup and dynamic thresholds
#564
robinholzi
opened
2 weeks ago
1
Enable finetuning on `RHO-LOSS` selection strategy when `Twin` strategy is used to produce the irreducible loss model
#563
XianzheMa
closed
1 week ago
5
fix: Adjust currently trained model to be defined for dataset end
#562
robinholzi
closed
2 weeks ago
0
fix: Fix minor issue in alibi detect config
#561
robinholzi
closed
2 weeks ago
3
Support threshold-based triggering for alibi-detect MMD
#560
MaxiBoether
closed
2 weeks ago
2
feat!: Batch evaluations in evaluator
#559
robinholzi
closed
2 weeks ago
7
Add rho training log to the pipeline file
#558
XianzheMa
closed
2 weeks ago
3
Windowing for data drift detection
#557
MaxiBoether
closed
2 weeks ago
3
Windowing for drift detection
#556
MaxiBoether
closed
2 weeks ago
0
feat: Create `DriftDetector` abstraction and introduce alibi-detect
#555
robinholzi
closed
2 weeks ago
3
Batch evaluation intervals into a single request and a single evaluation process
#554
XianzheMa
closed
2 weeks ago
8
Fix small issues in aggregation script
#553
MaxiBoether
closed
2 weeks ago
2
Twin RHO Model Step 2: split the training set and train the twin model
#552
XianzheMa
closed
2 weeks ago
3
feat: Create `DriftDetector` abstraction and introduce alibi-detect
#551
robinholzi
closed
2 weeks ago
3
Parameterize `drop_last`
#550
XianzheMa
closed
3 weeks ago
3
Add `cuda-cudart-dev` and `cuda-profiler-api` to environment
#549
MaxiBoether
closed
3 weeks ago
2
feat: Add unique sequence number for aggregation into supervisor
#548
robinholzi
closed
3 weeks ago
2
Twin RHO Model Step 1: create the Twin RHO Model
#547
XianzheMa
closed
2 weeks ago
2
feat: Refine triggering abstraction
#546
robinholzi
closed
3 weeks ago
2
fix: Allow to ignore certain assertion errors
#545
robinholzi
closed
3 weeks ago
2
Add `ratio_max` for presampling strategies
#544
MaxiBoether
closed
4 weeks ago
2
Weighted Optimization
#543
MaxiBoether
opened
4 weeks ago
1
BtS accumulation with non-multiples
#542
MaxiBoether
opened
4 weeks ago
0
Implement `ratio_max` scaling for downsamplers
#541
MaxiBoether
closed
4 weeks ago
2
Adding another way of calculating gradients in downsamplers
#540
XianzheMa
closed
3 weeks ago
3
Record training loss
#539
XianzheMa
closed
3 weeks ago
2
Missing commit SHA when running in Docker
#538
MaxiBoether
opened
1 month ago
0
feat: Log commit sha
#537
robinholzi
closed
1 month ago
2
Speeding up evaluations
#536
MaxiBoether
closed
2 weeks ago
0
fix: Adjust evaluation business logic
#535
robinholzi
closed
1 month ago
5
feat: More sophisticated evaluation logic
#534
robinholzi
closed
4 weeks ago
3
Fix batch number
#533
XianzheMa
closed
4 weeks ago
2
fix: Fix evaluation logic
#532
robinholzi
closed
1 month ago
2
Fix unfixed naming inconsistencies
#531
XianzheMa
closed
1 month ago
3
Fail explicitly in storage when asking for duplicate keys
#530
MaxiBoether
closed
1 month ago
2
Fix replacement option in loss and gradnorm downsampling
#529
MaxiBoether
closed
1 month ago
3
fix: Adjust active model definition
#528
robinholzi
closed
1 month ago
1
Next