xiezheng-cs / mtcnn_pytorch Goto Github PK
View Code? Open in Web Editor NEWMTCNN_Pytorch Codes Adapted from https://github.com/wujiyang/MTCNN_TRAIN
MTCNN_Pytorch Codes Adapted from https://github.com/wujiyang/MTCNN_TRAIN
差评!
Hi, @xiezheng-cs
Thanks for your great work, do you have the result on FDDB?
`
Train Epoch: 1 [1473536/1474127 (100%)] Loss: 0.183586 Accuracy: 0.935780
Train Epoch: 1 [227441/1474127 (100%)] Loss: 0.229168 Accuracy: 0.927273
Traceback (most recent call last):
File "C:\ProgramData\Anaconda3\lib\site-packages\google\protobuf\internal\python_message.py", line 526, in init
setattr(self, field_name, field_value)
File "C:\ProgramData\Anaconda3\lib\site-packages\google\protobuf\internal\python_message.py", line 674, in setter
field_setter(self, new_value)
File "C:\ProgramData\Anaconda3\lib\site-packages\google\protobuf\internal\python_message.py", line 662, in field_setter
new_value = type_checker.CheckValue(new_value)
File "C:\ProgramData\Anaconda3\lib\site-packages\google\protobuf\internal\type_checkers.py", line 109, in CheckValue
raise TypeError(message)
TypeError: tensor(0.1918, device='cuda:0', grad_fn=) has type <class 'torch.Tensor'>, but expected one of: ((<class 'numbers.Real'>,),)
During handling of the above exception, another exception occurred:
Traceback (most recent call last):
File "training/pnet/train.py", line 56, in
cls_loss_, box_offset_loss, total_loss, accuracy = trainer.train(epoch)
File "D:\ubuntu\Project\mtcnn_ms\mtcnn_pytorch\training\pnet\trainer.py", line 89, in train
self.logger.scalar_summary(tag, value, self.run_count)
File "D:\ubuntu\Project\mtcnn_ms\mtcnn_pytorch\tools\logger.py", line 22, in scalar_summary
summary = tf.Summary(value=[tf.Summary.Value(tag=tag, simple_value=value)])
File "C:\ProgramData\Anaconda3\lib\site-packages\google\protobuf\internal\python_message.py", line 528, in init
_ReraiseTypeErrorWithFieldName(message_descriptor.name, field_name)
File "C:\ProgramData\Anaconda3\lib\site-packages\google\protobuf\internal\python_message.py", line 448, in _ReraiseTypeErrorWithFieldName
six.reraise(type(exc), exc, sys.exc_info()[2])
File "C:\ProgramData\Anaconda3\lib\site-packages\six.py", line 692, in reraise
raise value.with_traceback(tb)
File "C:\ProgramData\Anaconda3\lib\site-packages\google\protobuf\internal\python_message.py", line 526, in init
setattr(self, field_name, field_value)
File "C:\ProgramData\Anaconda3\lib\site-packages\google\protobuf\internal\python_message.py", line 674, in setter
field_setter(self, new_value)
File "C:\ProgramData\Anaconda3\lib\site-packages\google\protobuf\internal\python_message.py", line 662, in field_setter
new_value = type_checker.CheckValue(new_value)
File "C:\ProgramData\Anaconda3\lib\site-packages\google\protobuf\internal\type_checkers.py", line 109, in CheckValue
raise TypeError(message)
`
请教一下,训练报这样的错误,是由tf 引起的吗,可以屏蔽tf模块吗,还有
1.pnet需要训练多少轮,大概loss与准确达到多少可以停止了,
2.gpu为啥才占300多M,您训练一个epoch大概需要多久呢,为啥我一个epoch需要两个小时以上有问题吗
感谢!
A declarative, efficient, and flexible JavaScript library for building user interfaces.
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google ❤️ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.