Has anyone incorporated the ability to save a graphdef protobuf along with each checkpoint in the inception-v3 model?
Lively discussions at https://github.com/tensorflow/tensorflow/issues/616 provided some solutions to export graph protobufs, but I cannot get any of them working using the inception training model (inception_train.py).
I am trying to implement graph.util.convery_variables_to_constants, but I am failing to capture trainable variables, as I always get assertion errors: [variable] is not in graph. I am using output names from tensorboard, which may be incorrect procedure.
Any solutions to this issue would definitely be of interest to the general community, not only for learning but deployment.
Thanks!
Related
I'm kinda new to TensorFlow and Keras, so please excuse any accidental stupidity, but I have an issue. I've been trying to load in models from the TensorFlow Detection Zoo, but haven't had much success.
I can't figure out how to read these saved_model folders (they contain a saved_model.pb file, and an assets and variables folder), so that they're accepted by Keras. Nor can I figure out a way to convert these models so that they may be loaded in. I've tried converting the SavedModel to ONNX, and then convert the ONNX-model to Keras, but that didn't work. Trying to load the original model as a saved_model, and then trying to to save this loaded model in another format gave me no success either.
Since you are new to Tensorflow (and I guess deep learning) I would suggest you stick with the API because the detection zoo models best interface with the object detection API. If you have already downloaded the model, you just need to export it using the exporter_main_v2.py script. This article explains it very well link.
I trained a model using Tensorflow object detection API for detecting swimming pools using satellite images. I used 'faster_rcnn_inception_v2_coco_2018_01_28' model for training. I generated a frozen inference graph (.pb). I want to evaluate the precision and recall of the model. Can someone tell me how I can do that, preferably without using pycocotools as I was facing some issues with that. Any suggestions are welcome :)
From the Object Detection API you can run "eval.py" from "models/research/object_detection/legacy/".
Your have to define an evaluation metric in your config file (see the supported evaluation protocols)
For example:
eval_config: {metrics_set: "coco_detection_metrics"}
The Pascal VOC e.g. then gives you the mean Average Precsion (mAP)
I am a beginner in NN APIs and TensorFlow.
I am trying to save my trained model in protobuff format (.pb), there are many blogs explaining how to save the model as protobuff. One thing I did not understand is what is the importance of freezing the graph before saving it as protobuff? I read that freezing coverts variable to constants, does that mean the model is not trainable anymore?
What else will freezing do on models?
What is that model loses after freezing?
Can anyone please explain or give some pointers on details of freezing?
This is only a partial answer to your question.
A freezed graph is easily optimizable. When doing inference (forward propagation) for instance you can fuse some of the layers together. This you can't do with a graph separated between variables and operations (a not frozen graph). Why would you want to fuse layers together? There are multiple reasons. Going hardware specific: it might be easier to compute a number of operations together in a group of tensors, specific to the structure of your cpu or gpu. TensorRT is a graph optimizer for instance that works starting from a frozen graph (here more info on graph optimizations done by tensorRT: https://devblogs.nvidia.com/tensorrt-integration-speeds-tensorflow-inference/ ). This software does graph optimizations as well as hardware specific ones.
As far as I understand you can unfreeze a graph. I have only worked optimizing them, so I haven't use this feature. But here there is code for it: https://gist.github.com/tokestermw/795cc1fd6d0c9069b20204cbd133e36b
Here is another question that might be useful:
TensorFlow: Is there a way to convert a frozen graph into a checkpoint model?
It has not yet been answered though.
Freezing the model means producing a singular file containing information about the graph and checkpoint variables, but saving these hyperparameters as constants within the graph structure. This eliminates additional information saved in the checkpoint files such as the gradients at each point, which are included so that the model can be reloaded and training continued from where you left off. As this is not needed when serving a model purely for inference they are discarded in freezing. A frozen model is a file of the Google .pb file type.
I've successfully trained the inception v3 model on custom 200 classes from scratch. Now I have ckpt files in my output dir. How to use those models to run inference?
Preferably, load the model on GPU and pass images whenever I want while the model persists on GPU. Using TensorFlow serving is not an option for me.
Note: I've tried to freeze these models but failed to correctly put output_nodes while freezing. Used ImagenetV3/Predictions/Softmax but couldn't use it with feed_dict as I couldn't get required tensors from freezed model.
There is poor documentation on TF site & repo on this inference part.
It sounds like you're on the right track, you don't really do anything different at inference time as you do at training time except that you don't ask it to compute the optimizer at inference time, and by not doing so, no weights are ever updated.
The save and restore guide in tensorflow documentation explains how to restore a model from checkpoint:
https://www.tensorflow.org/programmers_guide/saved_model
You have two options when restoring a model, either you build the OPS again from code (usually a build_graph() method) then load the variables in from the checkpoint, I use this method most commonly. Or you can load the graph definition & variables in from the checkpoint if the graph definition was saved with the checkpoint.
Once you've loaded the graph you'll create a session and ask the graph to compute just the output. The tensor ImagenetV3/Predictions/Softmax looks right to me (I'm not immediately familiar with the particular model you're working with). You will need to pass in the appropriate inputs, your images, and possibly whatever parameters the graph requires, sometimes an is_train boolean is needed, and other such details.
Since you aren't asking tensorflow to compute the optimizer operation no weights will be updated. There's really no difference between training and inference other than what operations you request the graph to compute.
Tensorflow will use the GPU by default just as it did with training, so all of that is pretty much handled behind the scenes for you.
I have used the TensorFlow object detection API to train the SSD Inception model from scratch. The evaluation script shows that the model has learned something and now I want to use the model.
I have looked at the object detection ipynb that can feed single images to a trained model. However, this is for SSD with MobileNet. I have used the following line (after loading the meta graph) to print the tensor names of the TensorFlow model I trained.
print([str(op.name) for op in tf.get_default_graph().get_operations()] )
But it does not contain the same input or output tensor names as in the ipynb. I have also searched through the code, but many functions point toward each other and it is difficult to find what I am looking for.
How can I find the tensor names I need? Or is there another method I do not know about?
To use the graph, you need to freeze/export it, using this provided script. The resulting .pb file will contain the nodes you need. I don't know why it's organized like that, but it is.