train.py following ''has no field named "fmodel"

[swordzxz]

Hi. I try to train.py with cmd: python3 train.py train --config_path=/root/second.pytorch/second/configs/car.config --model_dir=/root/model/second_model

but encountered following error message

Traceback (most recent call last): File "train.py", line 660, in fire.Fire() File "/root/anaconda3/envs/py34/lib/python3.6/site-packages/fire/core.py", line 127, in Fire component_trace = _Fire(component, args, context, name) File "/root/anaconda3/envs/py34/lib/python3.6/site-packages/fire/core.py", line 366, in _Fire component, remaining_args) File "/root/anaconda3/envs/py34/lib/python3.6/site-packages/fire/core.py", line 542, in _CallCallable result = fn(*varargs, **kwargs) File "train.py", line 107, in train text_format.Merge(proto_str, config) File "/root/anaconda3/envs/py34/lib/python3.6/site-packages/google/protobuf/text_format.py", line 536, in Merge descriptor_pool=descriptor_pool) File "/root/anaconda3/envs/py34/lib/python3.6/site-packages/google/protobuf/text_format.py", line 590, in MergeLines return parser.MergeLines(lines, message) File "/root/anaconda3/envs/py34/lib/python3.6/site-packages/google/protobuf/text_format.py", line 623, in MergeLines self._ParseOrMerge(lines, message) File "/root/anaconda3/envs/py34/lib/python3.6/site-packages/google/protobuf/text_format.py", line 638, in _ParseOrMerge self._MergeField(tokenizer, message) File "/root/anaconda3/envs/py34/lib/python3.6/site-packages/google/protobuf/text_format.py", line 730, in _MergeField (message_descriptor.full_name, name)) google.protobuf.text_format.ParseError: 1:1 : Message type "second.protos.TrainEvalPipelineConfig" has no field named "fmodel".

[traveller59]

this looks like a simple typo problem. check the first line of your config, it should be model{

[swordzxz]

thank you very much！ the problem had been solved。

and I meet another question，when I try kitti viewer web,I could open the browser with the frontend url（xxx.xxx.xxx.xx:8000），but when I click the ‘load’，it show ：

error：load kitti info fail，please check your backend!

what should I change ?

[swordzxz]

@traveller59

[traveller59]

@swordzxz could you please provide the error messages by press F12 when the error shows?

[swordzxz]

I am running on a server cluster, the server ip is 220.243.137.40, the localhost is 127.0.0.1, I tried 0.0.0.0:16666 220.243.137.40:16666 127.0.0.1:16666 Will not work

@traveller59

[swordzxz]

[traveller59]

this looks like the server doesn't open that port. firstly please change the ip address in backend.py to 127.0.0.1:16666. then check that server port is available.

[swordzxz]

Thanks for your reply, I re-viewed on the local server and can load the data. But can't ‘buildnet’, it's error. I think the path is wrong, so I want to ask how the root path, info path and det path，checkpoint path，config path in the kitti view web should be set. @traveller59

[traveller59]

see this example.

[swordzxz]

How can I visualize the loss curve of the training process？ @traveller59

[smallzm]

非常感谢！ 问题已解决。

我遇到另一个问题，当我尝试使用kitti viewer web时，我可以使用前端网址（xxx.xxx.xxx.xx:8000）打开浏览器，但是当我点击“加载”时，它显示：

错误：加载kitti信息失败，请检查您的后台！

我应该改变什么？

I have same problem. google.protobuf.text_format.ParseError: 3:5 : Message type "second.protos.VoxelNet" has no field named "num_class". How to slove it ? thanks