I want to ask if anyone enabled cudnn with yolo-windows?
I am seeing a flag in make file for CUDNN
but there is nothing like CUDNN from the src files,
what is the flag name that I should add in the preprocessor definitions to enable cudnn in visual studio??

Darknet compiles and runs correctly under window, but the yolo's result is not rational (see attached image). Is this a known issue? thanks.

When run on the default Yolo.cfg and default yolo.weights I get the following scores. Are there any preferred pre trained models for this, or anything to change besides .cfg and .weight files?

car: 16247449.00
bicycle: 134012440.00
sheep: 353145248.00
sheep: 224504528.00
sheep: 87352904.00
tvmonitor: 55276032.00
boat: 39831636.00
sheep: 341857536.00
sofa: 33466872.00
sofa: 340076672.00
person: 25631650.00
tvmonitor: 227381920.00
diningtable: 120787944.00
diningtable: 399327072.00
horse: 47826764.00
horse: 374694464.00
bus: 105163360.00
sofa: 202150400.00
pottedplant: 140311280.00
pottedplant: 92972976.00
bus: 105661352.00
bus: 303332608.00
cat: 593129856.00
cat: 401858752.00
bird: 537685760.00
chair: 41400060.00
car: 341706272.00
motorbike: 228624352.00
bird: 78465672.00
train: 11579180.00
bus: 10792380.00
bicycle: 215910960.00
bottle: 30197206.00
boat: 39359748.00
bicycle: 307236928.00
aeroplane: 351922304.00
pottedplant: 213473072.00
bird: 482928832.00
bird: 200470672.00
train: 519941056.00
horse: 21093454.00
car: 138302544.00
bicycle: 711727360.00
chair: 310756416.00
tvmonitor: 150253104.00
bird: 195140400.00
cow: 179580144.00
bottle: 316159264.00
bottle: 71264216.00
train: 18080070.00
diningtable: 103147320.00
sheep: 133795880.00
aeroplane: 82944616.00
sheep: 170625632.00
tvmonitor: 787143040.00
bicycle: 250857296.00
aeroplane: 93312384.00
motorbike: 303807424.00
train: 187714800.00
pottedplant: 135719648.00
pottedplant: 108505568.00
bicycle: 22330226.00
chair: 468975584.00
aeroplane: 292650720.00
diningtable: 136172768.00
cow: 352235712.00
cow: 438434208.00
sofa: 250186192.00
sofa: 197378272.00
train: 148903552.00
motorbike: 168899088.00
dog: 64905336.00
person: 74196096.00
car: 251464400.00
bus: 38416608.00
bus: 30357044.00
bus: 309075040.00
diningtable: 329295584.00
car: 80410272.00
car: 274951360.00
chair: 294920896.00
chair: 261512944.00
bicycle: 262204384.00
bicycle: 83653792.00
bicycle: 249232592.00
aeroplane: 435345920.00
car: 263475568.00
sheep: 146227728.00
sheep: 43476220.00
cat: 412864544.00
cat: 214374272.00
cat: 95979816.00
bus: 187131168.00
bus: 184859504.00

How to fine tune your model?
I don't have sufficient data to retrain your model from scratch.
I want to fine tune your model on my data which has only two classes ?

for example if there are two dogs in the image and I trained on one of them in all images that exist in training set,
is the other dogs in the training set that I didn't label and train on them will affect on the process and will cause to consider them part of background?

it seems so, because after 3000 batches it didn't detect anything?

错误 95 error MSB3721: 命令“"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v7.5\bin\nvcc.exe" -gencode=arch=compute_20,code="sm_20,compute_20" --use-local-env --cl-version 2013 -ccbin "E:\Program Files (x86)\Microsoft Visual Studio 12.0\VC\bin\x86_amd64" -I......\3rdparty\include -I"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v7.5\include" -I"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v7.5\include" -G --keep-dir x64\Debug -maxrregcount=0 --machine 64 --compile -cudart static -g -DGPU -D_CRT_SECURE_NO_WARNINGS -DWIN32 -D_DEBUG -D_CONSOLE -D_LIB -D_UNICODE -DUNICODE -Xcompiler "/EHsc /W3 /nologo /Od /Zi /RTC1 /MDd " -o x64\Debug\activation_kernels.cu.obj "D:\compute vision\yolo-windows-master\src\activation_kernels.cu"”已退出，返回代码为 2。 C:\Program Files (x86)\MSBuild\Microsoft.Cpp\v4.0\V120\BuildCustomizations\CUDA 7.5.targets 604 9 darknet