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!
Nevar pievienot vairāk kā 25 tēmas Tēmai ir jāsākas ar burtu vai ciparu, tā var saturēt domu zīmes ('-') un var būt līdz 35 simboliem gara.
Stanislaw Adaszewski 2eff854467 Start working on FastLoop. pirms 3 gadiem
..
unused Remove loss as no longer necessary. pirms 3 gadiem
__init__.py Work on icosagon.trainprep. pirms 3 gadiem
batch.py Introduce FlatPredictions, flatten_predictions(), BatchIndices, gather_batch_indices(). pirms 3 gadiem
bulkdec.py Fix a typo. pirms 3 gadiem
compile.py Performance tests. pirms 3 gadiem
convlayer.py Add BatchedData support to DecodeLayer. pirms 3 gadiem
convolve.py Remove an extra newline. pirms 3 gadiem
data.py Clean up _equal(). pirms 3 gadiem
databatch.py Add BatchedData support to DecodeLayer. pirms 3 gadiem
declayer.py Add BatchedData support to DecodeLayer. pirms 3 gadiem
decode.py Use torch.nn.Parameter(List) in decode. pirms 3 gadiem
dropout.py Remove unnecessary .detach() call in dropout_dense(). pirms 3 gadiem
fastconv.py Reimplement FastConvLayer. pirms 3 gadiem
fastloop.py Start working on FastLoop. pirms 3 gadiem
fastmodel.py Add test_fast_model_01(). pirms 3 gadiem
input.py Fix for non-leaf .grad warning. pirms 3 gadiem
model.py Add TrainLoop. pirms 3 gadiem
normalize.py Fixes for bool sparse tensors. pirms 3 gadiem
sampling.py Fixes for the GPU. pirms 3 gadiem
trainloop.py Add some debug output for profiling, the bottleneck is in DecodeLayer but also comes generally from computing always all nodes. pirms 3 gadiem
trainprep.py Fixes for the GPU. pirms 3 gadiem
weights.py Bring more elements to icosagon. pirms 3 gadiem