Hi,
In the API: image_embeddings.inference.run_inference(tfrecords_folder, output_folder, batch_size=1000), can we have a model parameter such that we can use the same API to get embeddings from any model we choose.
The default model can be EfficientNetB0 and we can modify it to compare between different versions.
(Label: Enhancement)
Hi,
In the API: image_embeddings.inference.run_inference(tfrecords_folder, output_folder, batch_size=1000), can we have a model parameter such that we can use the same API to get embeddings from any model we choose.
The default model can be EfficientNetB0 and we can modify it to compare between different versions.
(Label: Enhancement)