Fix Docs #368
Fix Docs #368
Conversation
ganler left a comment
ganler left a comment There was a problem hiding this comment.
I would approve the fix for now. But I think we still need a lot effort to make the training library easily installed by 'pip'.
This won't be hard b.c. we only handle some python files here.
In the best case, we should let users to install by 'pip install hyperpose[train]'.
As for the gpu requirements, you can simply advise the users to install cuda dependency via 'conda'.
| The hypeprose python training library handles the whole pipelines for developing the pose estimation system, including training, evaluating and testing. Its goal is to produce a **.npz** file that contains the well-trained model weights. | ||
| | ||
| For the training platform, the enviroment configuration above is engough. However, most inference engine only accept .pb format or .onnx format model, such as [TensorRT](https://docs.nvidia.com/deeplearning/tensorrt/install-guide/index.html). Thus, one need to convert the trained model loaded with .npz file weight to .pb format or .onnx format for further deployment, which need extra configuration below:<br> | ||
| For the training platform, the enviroment configuration above is engough. However, most inference engine only accept .pb format or .onnx format model, such as [TensorRT](https://docs.nvidia.com/deeplearning/tensorrt/install-guide/index.html). |
There was a problem hiding this comment.
Would mentioning our inference lib make more sense (instead of tensorrt)?
| We already provide a script with cli to facilitate conversion, which located at [export_pb.py](https://github.com/tensorlayer/hyperpose/blob/master/export_pb.py). What we need here is only **tensorflow** library that we already installed. | ||
| | ||
| * (II)Convert to .onnx format:<br> | ||
| We already provide a script with cli to facilitate conversion, which located at [export_pb.py](https://github.com/tensorlayer/hyperpose/blob/master/export_pb.py). What we need here is only *tensorflow* library that we already installed. |
There was a problem hiding this comment.
In future version, I think we'd better put everything related to python training lib into '/hyperpose' folder. And if it is some utility scripts, let's put it in '/script'. Because putting all those .py files on the top-level is very confusing.
For example, I made some useful scripts in the '/script' folder. Also you want to verify them if you make any changes.
Fix train installation documents format error