edgetpu-ssdlite-mobiledet-retrain

edgetpu-ssdlite-mobiledet-retrain copied to clipboard

hello can someone help me? i got error when i try to trainig the data Nan loss during training. how can i solve this?

4212011042-fajarrahmatullah

Below is the compiler error: Edge TPU Compiler version 16.0.384591198 Started a compilation timeout timer of 180 seconds. ERROR: :317 scale_diff / output_scale

KushalHM

Lost 40% accuracy after converting to tflite model

2

Hi Nam Vu, first thank you much for that nice instruction and the amazing possability through Colab :) I used your colab notebook to train my own coustom model. Only...

rudikx6

learning_rate probblem

2

hi the project is cool but for lower loss i suggest to use a "cosine total steps" similar to the total steps, for example if you are doing 30k steps...

lucaSartore

Why it runs on GPU

3

Hi, what is the "secret" to let the model train on GPU? The tutorial from google runs only on CPU...? ( https://colab.research.google.com/github/google-coral/tutorials/blob/master/retrain_detection_qat_tf1.ipynb ) Thank you!

hannes56a

hi @Namburger thanks for your reseach on mobiledet especially on coral. i have 9000 training images of single class. but unfortunately, getting low accuracy with this pipeline, 1. should i...

parameswaraRao-13

Tensorflow 2.0 Support

1

Is there an easy way to convert this example to support tensorflow 2.0?

teraSurfer