Dou
> 先把代码中的model文件夹名字改成models，然后加载时用下面代码 > comb = torch.load(path) > fsrnet = comb['model'] > 这样写就没问题了 > 你试试能不能跑起来 改名models后,那个错误就过去了,但是出现了新的错. 我是参照你的train.py中加载预训练模型的部分,这样写的: weights = torch.load(path) print(weights['model']) pretrained_dict = weights['model'].state_dict() model_dict = fsrnet.state_dict() pretrained_dict = {k: v...
非常感谢, 我可以改一下网络的定义来加载预训练网络
我下载好模型，运行代码没有什么错误，但是无论输入什么prompt，生成的都是上边这张图。请问有什么原因吗
另外，可以再分享一下 微信群二维码吗，原来的过期了
Sorry! I have another error with the latest version in CPU_ONLY, "fatal error C1083: : “..\..\src\caffe\layers\fcn_data_layer.cpp”: No such file or directory". And I wonder what source files are vital and...
@happynear thank you! The libcaffe and caffe have been built successful. But when I build the pycaffe , there is an error pointing to " DEFINE_BOOST_GET_POINTER(caffe::NCCL)" in "\python\caffe\_caffe.cpp" . I...