Age | Commit message (Collapse) | Author | Files | Lines | |
---|---|---|---|---|---|
2016-05-16 | Add cuDNN v5 support, drop cuDNN v3 support | Felix Abecassis | 1 | -2/+21 | |
cuDNN v4 is still supported. | |||||
2015-12-01 | dismantle layer headers | Evan Shelhamer | 1 | -1/+1 | |
No more monolithic includes: split layers into their own headers for modular inclusion and build. | |||||
2015-10-21 | Clean redundant/unnecessary headers | Kang Kim | 1 | -2/+0 | |
2015-03-24 | replace cuDNN alphas and betas with coefficient values | Evan Shelhamer | 1 | -12/+4 | |
Give cuDNN {0, 1} constants for controlling accumulation through the alpha and beta coefficients. | |||||
2015-03-24 | switch to cuDNN R2 | Simon Layton | 1 | -5/+19 | |
2014-10-10 | fix instantiation | Yangqing Jia | 1 | -1/+1 | |
2014-09-19 | fix types of (Layer)SetUp, Reshape, Forward, and Backward calls | Jonathan L Long | 1 | -5/+5 | |
Using the type vector<Blob<Dtype*>* for outputs allows modification of the vector itself, while it is only okay to modify the blobs pointed to by the elements of the vector. Switching the types to const vector<Blob<Dtype>*>& makes them more correct. | |||||
2014-09-07 | CUDNN_CHECK | Evan Shelhamer | 1 | -8/+4 | |
2014-09-07 | strategize cuDNN activations: ReLU, Sigmoid, TanH | Evan Shelhamer | 1 | -0/+55 | |