issues
search
Oufattole
/
meds-torch
MIT License
15
stars
1
forks
source link
issues
Newest
Newest
Most commented
Recently updated
Oldest
Least commented
Least recently updated
We should consider making nheads a configurable parameter in a future commit. Had to reduce 8 to 2 so that it was divisible by test data token embedding dimension.
#137
zackeberger
opened
3 days ago
0
Add e2e intermediate fusion and early fusion supervised tests
#136
Oufattole
closed
2 days ago
2
Adds the multimodal pytorch dataset class
#135
Oufattole
closed
3 days ago
2
Multimodal Synthetic ECG data
#134
Oufattole
closed
3 days ago
2
Enhance benchmark and throughput speed
#133
mmcdermott
opened
6 days ago
0
Text code encoder
#132
teyaberg
opened
1 week ago
2
I'm not sure these both should be `side="right"`
#131
mmcdermott
opened
1 week ago
0
Throw an error when the custom_normalization.py script is run in parallel, as it require being run with only 1 worker.
#130
Oufattole
opened
2 weeks ago
2
Massive Memory Usage when initializing the dataloader for large tasks
#129
Oufattole
opened
2 weeks ago
0
Enable key-value caching when there is a mask
#128
Oufattole
opened
2 weeks ago
0
Zero shot
#127
Oufattole
opened
2 weeks ago
2
Refactor PytorchDataset class and other dataset classes to a `Window` dataclass for batches instead of a dictionary
#126
Oufattole
opened
2 weeks ago
0
Enable users to define their own dataset class
#125
payalchandak
opened
2 weeks ago
1
Define a Window in documentation somewhere
#124
Oufattole
opened
2 weeks ago
0
Reorganize DataModule and Dataset Configuration Structure
#123
Oufattole
opened
2 weeks ago
0
removed MIMICIV_TUTORIAL. updated README for MIMICIV_INDUCTIVE_EXPERI…
#122
kingrc15
opened
2 weeks ago
2
Multimodal EHR Data Processing Enhancements
#121
Oufattole
closed
2 weeks ago
2
Tutorial Clarification
#120
payalchandak
closed
2 weeks ago
1
Code-to-Text Integration
#119
Oufattole
opened
3 weeks ago
0
Multimodal Modeling Methods to Support
#118
Oufattole
opened
3 weeks ago
1
Refactor multiwindow_dataset to support single time window data loading and improve documentation
#117
Oufattole
opened
3 weeks ago
0
Multimodal custom meds-transforms
#116
Oufattole
opened
3 weeks ago
0
Tutorial Improvement
#115
Oufattole
opened
3 weeks ago
0
API Update To Support Noise Augmentation and Masking Intermediate Stages
#114
Oufattole
opened
3 weeks ago
0
MEDS death code/ Refer to MEDS-DEV tutorial
#113
rvandewater
opened
3 weeks ago
1
Allow Predict Script to Include Custom Prefixed Keys in Output and Batch Indexes
#112
Oufattole
opened
3 weeks ago
0
The `Module` ABC class is confusing and should be re-done as a decorator or Mixin
#111
mmcdermott
opened
3 weeks ago
0
Add a prediction script that generates meds-evaluation compliant parquets.
#110
Oufattole
closed
2 weeks ago
2
Add a predict script that generates embeddings
#109
Oufattole
closed
2 weeks ago
0
Forecast api update
#108
Oufattole
closed
2 weeks ago
3
Update the Triplet autoregressive model to perform conditional sampling of values conditioned on the code
#107
Oufattole
opened
4 weeks ago
0
Unconstrained Generative evaluation
#106
Oufattole
opened
4 weeks ago
0
Zero-shot evaluation
#105
Oufattole
opened
4 weeks ago
0
Add API support for generation that allow generating data after a prediction time and caching both the predictions and the real data.
#104
Oufattole
closed
2 weeks ago
1
Updated Slow Tests and Tuning Scripts, Cleaned Some pytorch_dataset Documentation, Improved EIC Quantile Binning
#103
Oufattole
closed
3 weeks ago
4
Update to use the latest JNRT features.
#102
mmcdermott
closed
1 month ago
2
The resolver for the max sequence length is not needed as it is now handled in the dataset class.
#101
mmcdermott
closed
1 month ago
1
Static data is not being used in cases where it is not prepended.
#100
mmcdermott
opened
1 month ago
0
The EOS max seq length adjustment should happen in the dataset, not in the config, to avoid violating the user's inputs.
#99
mmcdermott
closed
1 month ago
0
Removed unused event-stream remnant function.
#98
mmcdermott
closed
1 month ago
1
Reduces to a single collate function with prepending and postpending additions standardized.
#97
mmcdermott
closed
1 month ago
1
Delete the unused `dynamic_only_collate`
#96
mmcdermott
closed
1 month ago
0
Is the max seq len constraint not respected if postpend EOS or prepend static is enabled?
#95
mmcdermott
closed
1 month ago
1
All collate functions should be the same and static data pre-pending should happen in getitem.
#94
mmcdermott
closed
1 month ago
6
Add the source directory to the testpaths so that doctests run by default.
#93
mmcdermott
closed
1 month ago
1
Simplify the dataset index for now to just have a subject ID and an end event index.
#92
mmcdermott
opened
1 month ago
1
Doctests are not being run by github workflows
#91
mmcdermott
closed
1 month ago
3
Removes the unused `out_labels` block.
#90
mmcdermott
closed
1 month ago
2
Out labels is never used and should be deleted.
#89
mmcdermott
closed
1 month ago
0
Simplify the task code by leveraging the reliability of the meds label schema.
#88
mmcdermott
closed
1 month ago
1
Next