search

ubclaunchpad / minutes

:telescope: Speaker diarization via transfer learning
https://medium.com/ubc-launch-pad-software-engineering-blog/speaker-diarisation-using-transfer-learning-47ca1a1226f4
27 stars 5 forks source link

Improve Base Model Naming #124

Open chadlagore opened 6 years ago

chadlagore commented 6 years ago

Our current base model (cnn) could be named better. Which makes me think, perhaps we could use a naming convention. Something like: 

<model-type>-<number-of-base-layers>-<number-of-transfer-layers>

So this model would become cnn-1-2.

iKevinY commented 6 years ago

Seems like a sensible naming convention to me!

grig-guz commented 6 years ago

How about we do <model-type>-<number-of-base-layers>-<number-of-transfer-layers>-<user-id>? Seems like quite a few people want to add their own models.

chadlagore commented 6 years ago

It would be nice to capture the model architecture uniquely somehow in the name—perhaps Keras provides such a checksum? I haven't been able to find it in their API 😞