decagon-pytorch

IF YOU WOULD LIKE TO GET AN ACCOUNT, please write an email to s dot adaszewski at gmail dot com. User accounts are meant only to report issues and/or generate pull requests. This is a purpose-specific Git hosting for ADARED projects. Thank you for your understanding!

Historik

Stanislaw Adaszewski 2eff854467 Start working on FastLoop.		3 år sedan
..
unused	Remove loss as no longer necessary.	3 år sedan
__init__.py	Work on icosagon.trainprep.	3 år sedan
batch.py	Introduce FlatPredictions, flatten_predictions(), BatchIndices, gather_batch_indices().	3 år sedan
bulkdec.py	Fix a typo.	3 år sedan
compile.py	Performance tests.	3 år sedan
convlayer.py	Add BatchedData support to DecodeLayer.	3 år sedan
convolve.py	Remove an extra newline.	3 år sedan
data.py	Clean up _equal().	3 år sedan
databatch.py	Add BatchedData support to DecodeLayer.	3 år sedan
declayer.py	Add BatchedData support to DecodeLayer.	3 år sedan
decode.py	Use torch.nn.Parameter(List) in decode.	3 år sedan
dropout.py	Remove unnecessary .detach() call in dropout_dense().	3 år sedan
fastconv.py	Reimplement FastConvLayer.	3 år sedan
fastloop.py	Start working on FastLoop.	3 år sedan
fastmodel.py	Add test_fast_model_01().	3 år sedan
input.py	Fix for non-leaf .grad warning.	3 år sedan
model.py	Add TrainLoop.	3 år sedan
normalize.py	Fixes for bool sparse tensors.	3 år sedan
sampling.py	Fixes for the GPU.	3 år sedan
trainloop.py	Add some debug output for profiling, the bottleneck is in DecodeLayer but also comes generally from computing always all nodes.	3 år sedan
trainprep.py	Fixes for the GPU.	3 år sedan
weights.py	Bring more elements to icosagon.	3 år sedan