Tensorflow object detection using squeezedet (deployment)

Our project is based in squeezedet which can be found here ([login to view URL]) We are able to train the model and get fairly good results. Basically, we now want to bring this model to deployment (inference). This is the help we need. 1. To help us identify if the model is suitable for deployment. For example, we see the input to first layer is batch, we want to do object detection in real-time. 2. Take the checkpoints files and write a script to freeze the graph. 3. Write a script to optimize the graph for inference. 4. Write a script to convert to TFLite format. 5. Write a simple script to load the model in TFLite format and load an image to test object detection.