v1.4.0
What's new
Changed ⚠️
- Updated default layer norm epsilon for OLMo models from
1e-5to1e-6to match latest model. - Renamed
FSLDataLoadertoNumpyFSLDataLoader. - Renamed
VSLDataLoadertoNumpyVSLDataLoader. - The trainer now takes a
data_loader: DataLoaderBaseinstead of adataset: NumpyDatasetBase.
Commits
55343dd fix loading training state dict
b921299 Allow unknown number of batches with data loaders
87f1e89 fix restarts for custom data loader
767c550 Add example of custom data loader
6237f7d Trainer now takes a data loader instead of a dataset (#59)
f6fc369 update default LN eps to match latest OLMo model (#58)
db522d1 allow loading via pickling
7d26589 make VSL curr config more flexible