TFLite model maker 自定义 object 检测器训练使用 tfrecord

Question

I am trying to train a custom object detector using tflite model maker ( https://www.tensorflow.org/lite/tutorials/model_maker_object_detection ). I want to deploy trained tflite model to coral edgeTPU. I want to use tensorflow tfrecord (multiple) as input for training a model like object detection API. I tried with tflite_model_maker.object_detector.DataLoader( tfrecord_file_patten, size, label_map, annotations_json_file=None ) but I am not able to work around it. I have following questions.

• Is it possible to tfrecord for training like mentioned above?
• Is it also possible to pass multiple CSV files for training?

Answer 1

For multiple CSV files, you could probably just append one file to the other. Then you'd just have to pass one csv file.

As for passing a tfrecord instead, this should be possible. I'm also attempting to do this, so if I get it working I'll update my post. Looking at the source , it seems from_cache is the function internally used. Following that structure, should be able to create a DataLoader object similarly:

DataLoader(tfrecord_file_patten, meta_data['size'],
                      meta_data['label_map'], ann_json_file)

Just taking a quick look at from_csv , I noticed that it returns a list called data , which seems to append potentially multiple DataLoader objects. So if creating the DataLoader didn't work, perhaps try just wrapping it in a list, like tfrecord_dl = [DataLoader(tfrecord_file_patten, meta_data['size'], meta_data['label_map'], ann_json_file)] .

I haven't tried it yet though so these are just my initial thoughts. If you share what you tried we can take a closer look at what actually went wrong.