add SIFT Flow FCNs
These nets are jointly trained for segmentation of semantic and geometric classes since this dataset includes annotations for both. - FCN-32s SIFT Flow - FCN-16s SIFT Flow - FCN-8s SIFT Flow TODO: fix semantic class evaluation for this dataset, which requires special care since there are missing classes in the test set.
Showing
data/sift-flow/README.md
0 → 100644
data/sift-flow/classes.txt
0 → 100644
data/sift-flow/test.txt
0 → 100644
data/sift-flow/trainval.txt
0 → 100644
此差异已折叠。
siftflow-fcn16s/caffemodel-url
0 → 100644
siftflow-fcn16s/net.py
0 → 100644
siftflow-fcn16s/solve.py
0 → 100644
siftflow-fcn16s/solver.prototxt
0 → 100644
siftflow-fcn16s/test.prototxt
0 → 100644
siftflow-fcn16s/trainval.prototxt
0 → 100644
siftflow-fcn32s/caffemodel-url
0 → 100644
siftflow-fcn32s/net.py
0 → 100644
siftflow-fcn32s/solve.py
0 → 100644
siftflow-fcn32s/solver.prototxt
0 → 100644
siftflow-fcn32s/test.prototxt
0 → 100644
此差异已折叠。
siftflow-fcn32s/trainval.prototxt
0 → 100644
此差异已折叠。
siftflow-fcn8s/caffemodel-url
0 → 100644
siftflow-fcn8s/net.py
0 → 100644
siftflow-fcn8s/solve.py
0 → 100644
siftflow-fcn8s/solver.prototxt
0 → 100644
siftflow-fcn8s/test.prototxt
0 → 100644
此差异已折叠。
siftflow-fcn8s/trainval.prototxt
0 → 100644
此差异已折叠。
siftflow_layers.py
0 → 100644
想要评论请 注册 或 登录