Blobs are N-D arrays (for N not necessarily equals 4) #1970
Merged
Commits
Show all changes
35 commits
Select commit
Hold shift + click to select a range
1434e87
Blobs are ND arrays (for N not necessarily equals 4).
jeffdonahue 559ff3a
InnerProductLayer weights are 2D; biases are 1D
jeffdonahue a0fa2a9
LossLayer output is 0D (scalar)
jeffdonahue 29581e6
InnerProductLayer can multiply along any axis
jeffdonahue c4e9ec4
TestBlob: test that legacy BlobProtos are correctly handled by ShapeE…
jeffdonahue 94179cc
ConvLayer biases are 1D
jeffdonahue 5407f82
Add BlobShape message; use for Net input shapes
jeffdonahue 119a1c6
add offset, {data,diff}_at nd blob accessors
jeffdonahue cf9fdda
Fix sparse GaussianFiller for new IPLayer weight axes
jeffdonahue bf73cb1
EltwiseLayer need not assume old 4D dim names
jeffdonahue 8afdcd0
ConcatLayer: generalized Blob axes
jeffdonahue 1b97c06
FlattenLayer: generalized Blob axes
jeffdonahue 704e524
TestConcatLayer: fix style errors
jeffdonahue d52e9a8
TestConcatLayer: add forward/gradient tests for concatenation along num
jeffdonahue b868916
SliceLayer: generalized Blob axes
jeffdonahue fb9caee
common_layers.hpp: remove unused "Blob col_bob_"
jeffdonahue d8c6aeb
AccuracyLayer output is 0D (scalar)
jeffdonahue 6b8a765
AccuracyLayer generalized to N instance axes
jeffdonahue 8e96445
Test{Net,Solver} fixes for AccuracyLayer generalization
jeffdonahue abec302
SoftmaxLayer: generalized Blob axes
jeffdonahue e2bc9f9
SplitLayer: change Reshape(n,h,c,w) to ReshapeLike(...)
jeffdonahue e56377d
DataLayer outputs 1D labels
jeffdonahue e6468e9
HDF5DataLayer shapes output according to HDF5 shape
jeffdonahue 7c8725b
MemoryDataLayer outputs 1D labels
jeffdonahue c87a136
ImageDataLayer outputs 1D labels
jeffdonahue 9505001
WindowDataLayer outputs 1D labels
jeffdonahue fcbb933
EuclideanLossLayer: generalized Blob axes
jeffdonahue 7462c84
DummyDataLayer outputs blobs of arbitrary shape
jeffdonahue 269dafa
PyBlobs support generalized axes
jeffdonahue 69fc1f6
Add CHECK_EQ(4, ...)s to "vision layers" to enforce that the
jeffdonahue 71df6f9
Add option not to reshape to Blob::FromProto; use when loading Blobs
jeffdonahue 94d93da
SoftmaxLossLayer generalized like SoftmaxLayer
jeffdonahue 60c288b
CuDNNSoftmaxLayer: generalized Blob axes
jeffdonahue aa242aa
[pycaffe] expose Blob.reshape as *args function
longjon 8c79d65
[pytest] use non-4d blobs in test_python_layer
longjon
Jump to file or symbol
Failed to load files and symbols.
Oops, something went wrong.