search

wzzheng / HDML

Implementation of Hardness-Aware Deep Metric Learning (CVPR 2019 Oral) in Tensorflow.
151 stars 28 forks source link

Inferences #2

Closed thaiat closed 5 years ago

thaiat commented 5 years ago

Hi, and thks for this repo Could you provide some guidances on running inferences when the model is trained ?

wzzheng commented 5 years ago

Hi, you can load the trained model by restoring the checkpoint (set FLAGS.load_formalVal = True), and then compute embedding_z with your data (something like sess.run([embedding_z])). You can use the embeddings to do your own tasks.

We will try to add inference codes for easy use.

batrlatom commented 5 years ago

@wzzheng Is there any easy inference code as you mentioned earlier?