something wrong when I built lenet5
Opened this issue · 0 comments
when I built this network by :
./nvdla_compiler --prototxt Lenet5/lenet.prototxt --caffemodel Lenet5/lenet_iter_10000.caffemodel -o . --profile fast-math --cprecision int8 --configtarget nv_small --quantizationMode per-filter --batch 1 --informat nhwc
, something went wrong. here is the log:
creating new wisdom context...
opening wisdom context...
parsing caffe network...
libnvdla<3> mark prob
Marking total 1 outputs
initialize all tensors with const scaling factors of 127...
attaching parsed network to the wisdom...
compiling profile "fast-math"... config "nv_small"...
libnvdla<2> Overriding input multibatch size from 64 to 1
libnvdla<2> Prototxt #chnls (C = 1) != Profile #chnls for input (NVDLA_IMG_A8B8G8R8: C = 4). Preferring #chnls from Profile for compiling.
closing wisdom context...
where can I change the channels of the sw?