EternaBrain
EternaBrain copied to clipboard
Deep learning to solve RNA design puzzles
Bumps [tensorflow](https://github.com/tensorflow/tensorflow) from 1.15.5 to 2.11.1. Release notes Sourced from tensorflow's releases. TensorFlow 2.11.1 Release 2.11.1 Note: TensorFlow 2.10 was the last TensorFlow release that supported GPU on native-Windows. Starting...
## CNN Use BEAR notation as additional structural feature ## SAP For comparisons, use BEAR instead of dot-bracket notation
- Needs to work for moves with multiple base changes in one move - Figure out a way to encode number of moves needed to complete the puzzle
- On 1 GPU, limited to 10 convolutional layers - When parallelized across multiple GPUs, can add more layers as more memory available
Currently `structure_and_energy_at_current_time` works only with 1 puzzle ID. Would reduce the number of pickles and the amount of time unpickling when training.
- Python running out of memory when clustering movesets (not on reduced dimension data) - 8GB - Too many features for clustering algorithm - 6892344, 6892345, 7254756, 7254759