小哥哥您好，很抱歉打扰到您，我在window10（无gpu ）上用vs code 运行这个项目时出现了两个问题。
1：

2：我在使用MyresNet34网络的时候出现了           

AttributeError: ‘MyresNet34‘ object has no attribute ‘load’错误
可是我在config.py里面确确实实修改了配置，如下图：

这个问题困扰了我好多天，如果您愿意的话请给我一点指点，感激不尽！！！我的微信是18186473512，谢谢你

myData, test= False
Traceback (most recent call last):
File "main.py", line 322, in
fire.Fire()
File "/anaconda3/lib/python3.7/site-packages/fire/core.py", line 127, in Fire
component_trace = _Fire(component, args, context, name)
File "/anaconda3/lib/python3.7/site-packages/fire/core.py", line 366, in _Fire
component, remaining_args)
File "/anaconda3/lib/python3.7/site-packages/fire/core.py", line 542, in _CallCallable
result = fn(*varargs, **kwargs)
File "main.py", line 100, in train
data_source='none')
File "/Users/b0074981/Desktop/facelivenesspython/2/data/require_dataset.py", line 39, in init
fopen = open(filedir,'r')
FileNotFoundError: [Errno 2] No such file or directory: '/home/cv/zjh/aich/dataset/raw/client_train_raw.txt'

Please help

When I used VS code on windows10 running python main.py train running python main.py train, it occurred the following errors:

I would appreciate it if you could reply to me ,This is my email : [email protected]

Dear sir @JinghuiZhou, thanks for a great repository and many network architecture implementations. But i have an issue that i think the code need to do a little bit more configuration. When i was training Resnet18, Resnet30, VGG,... everything is ok. But when i switch to MultiscaleresNet18, i got a bug below

RuntimeError: Expected 4-dimensional input for 4-dimensional weight [64, 3, 7, 7], but got 5-dimensional input of size [2, 1, 3, 224, 224] instead

The error happens in the line summary(model, (3, 224, 224)) and the input of the model in main.py.
I think the code need a small change when training with MultiscaleresNet. But i couldn't find where to fix it. Can you show me how? Thanks a lot!!!

Need to talk, how can we contact you?

This is a great project, but I think if you can make the model smaller, that would be even better for embedded system...

Please me how to train.

python3 main.py train
Traceback (most recent call last):
File "main.py", line 5, in
from data import myData
ModuleNotFoundError: No module named 'data'

when i run "python main.py train ", I encountered “axes don't match array” probelm, what is the reason? Thank you!

@JinghuiZhou @JaryHuang 测试数据之外，用手机拍的结果很不好，10个epoch太少了吗，是不是需要训练更多的epoch，求解？？

requirements.txt now blank，wish add pkgs with versions ， Tks a lot

when choosing a different network like MyseNet1_0 for training loss becomes nan.
Also when i try to modified the resnet18 architecture training loss becomes nan.
please see the screenshot below:
(https://user-images.githubusercontent.com/47495143/66979488-56fd9f00-f0cb-11e9-974b-9c96cc18bb21.png)

i am not able to identify data. please help. Need your help here.

I saw your showcropface_train.jpg and showcropface_val.jpg. I found that most of the spoofing images contained the finger of the person holding the picture. So, the model might learn the finger instead of the difference between a real image and a spoofing image.

你好，方便告诉一下邮箱吗

File "/home/lyn/awesome_face_antispoofing/utils/Visualizer.py", line 67, in plot_many_stack
update=None if x == 0 else 'append'
File "/root/anaconda3/envs/py35-torch-liveness/lib/python3.5/site-packages/visdom/init.py", line 335, in wrapped_f
return f(*args, **kwargs)
File "/root/anaconda3/envs/py35-torch-liveness/lib/python3.5/site-packages/visdom/init.py", line 1367, in line
update=update, name=name)
File "/root/anaconda3/envs/py35-torch-liveness/lib/python3.5/site-packages/visdom/init.py", line 335, in wrapped_f
return f(*args, **kwargs)
File "/root/anaconda3/envs/py35-torch-liveness/lib/python3.5/site-packages/visdom/init.py", line 1292, in scatter
return self._send(data_to_send, endpoint=endpoint)
File "/root/anaconda3/envs/py35-torch-liveness/lib/python3.5/site-packages/visdom/init.py", line 548, in _send
data=json.dumps(msg),
File "/root/anaconda3/envs/py35-torch-liveness/lib/python3.5/json/init.py", line 230, in dumps
return _default_encoder.encode(obj)
File "/root/anaconda3/envs/py35-torch-liveness/lib/python3.5/json/encoder.py", line 198, in encode
chunks = self.iterencode(o, _one_shot=True)
File "/root/anaconda3/envs/py35-torch-liveness/lib/python3.5/json/encoder.py", line 256, in iterencode
return _iterencode(o, 0)
File "/root/anaconda3/envs/py35-torch-liveness/lib/python3.5/json/encoder.py", line 179, in default
raise TypeError(repr(o) + " is not JSON serializable")
TypeError: tensor(0.9837, device='cuda:0', dtype=torch.float64) is not JSON serializable

Hey,
I would like to test your network, yet I don't have the resources to train it myself from scratch. Could you provide a google drive / pan baidu link if possible?
Regards

Did you use the pretrained models and remove the last two layers and add Convolution layer and fully connected layer, and then fintuned?
Did i understand your project correctly?

Thank you :)

I noticed that you have uploaded some pretrained models. Is there any inference script for test these models? thx.

你好，请问一下可以放个百度网盘的链接吗，谷歌的下载总是出错

Hi, thanks for providing your interesting work.
I have an issue when implementing your code on VGG-11 with the NUAA dataset. I got
EER = 0.048105, TPR(1%) = 0.876410, TPR(.5%) = 0.831453, AUC = 0.98986704,
which huge difference from yours publish result.
Can u explain it?

Thanks

RuntimeError: Cannot initialize CUDA without ATen_cuda library. PyTorch splits its backend into two shared libraries: a CPU library and a CUDA library; this error has occurred because you are trying to use some CUDA functionality, but the CUDA library has not been loaded by the dynamic linker for some reason. The CUDA library MUST be loaded, EVEN IF you don't directly use any symbols from the CUDA library! One common culprit is a lack of -Wl,--no-as-needed in your link arguments; many dynamic linkers will delete dynamic library dependencies if you don't depend on any of their symbols. You can check if this has occurred by using ldd on your binary to see if there is a dependency on *_cuda.so library.

About main.py:line 229
In this part:
`vis.plot_many_stack({'train_loss': train_loss.value()[0],
'val_loss': val_loss.value()[0]}, win_name="Loss")

vis.plot_many_stack({'train_acc': train_acc.value()[0],
'val_acc': val_acc.value()[0]}, win_name='Acc')`

When using GPU training，var train_acc and val_acc should use
.cpu() to load in memory，So I change source code like this:
if opt.use_gpu: val_acc.add(v_accuracy.cpu()) else: val_acc.add(v_accuracy)
And
if opt.use_gpu: train_acc.add(epoch_acc.cpu()) else: train_acc.add(epoch_acc)

model_path = r'../data/model/71.pth'

model = getattr(models, 'MyresNet34')().eval()
if torch.cuda.is_available():
model = model.cuda()

model.load(model_path)

else:

model.load(model_path, map='cpu')

transformer = transforms.Compose([transforms.ToTensor()])#, transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225])

def val(img):
img = maxcrop(img)
img = transformer(img)
if torch.cuda.is_available():
img = img.cuda()
img = img.view(1, *img.size())
with torch.no_grad():
val_input = Variable(img)

model.eval()
score = model(val_input)

outputs = torch.softmax(score, dim=-1)
preds = outputs[0].to('cpu').detach().numpy()

return preds

if name == 'main':
path = r'../data/img/val/y'
filelist = os.listdir(path)
for file in filelist:
image = os.path.join(path, file)
img = Image.open(image).convert('RGB')
res = val(img)
print(res)

First ,thanks for your great work,it really helps me a lot.
The progress of train is OK. While try to put the result into visdom and visualize it,here is a problem:
TypeError: object of type 'Tensor' is not JSON serializable.
My Pytorch is 1.0.0, and some questions on StackOverflow tell me to change Pytorch 1.0.0 to 0.3.1, but when I tried that, it showed more problems. so give it up.
Can you give me some advice about that? thanks in advance.

Can you please let me know what path or value i need to put here-

env = model # visdom 环境
train_root = '' # 训练集存放路径
val_root = '' # 验证集存放路径
test_root = 'data/AgriculturalDisease_testA/' # 测试集存放路径

some sample required please.