0.4.0

Merge pull request #1 from NielsRogge/add_demo

Host weights on HuggingFace hub

0.3.6

add more freezing unfreezing functions for finetuning

0.3.5

give convnext blocks a kernel size of 3, based on empiricism - also m…

0.3.4

allow for customizing resolution of dna sequence being trained on (en…

0.3.2

make sure seq indices can undergo reverse complement augmentation too

0.3.1

allow for reverse complement augmentation from FastaInterval

0.3.0

release enformer models trained on 191k basepairs

0.2.17

add convnext blocks, for ablating the contribution of enformer from t…

0.2.16

truly make gdown optional

0.2.15

simplify pearson corr coef