search

Howe-Young / asc_triplet

triplet loss on Acoustic Scene Classification-PyTorch
MIT License
9 stars 6 forks source link

TODO #1

Open Howe-Young opened 5 years ago

Howe-Young commented 5 years ago
  1. Batch All
  2. Batch All(not equal zero)
  3. Batch Hard
  4. network architecture
  5. soft margin
  6. distance(cosine distance)
  7. device B, C
hackerekcah commented 5 years ago

network arch

consider residual block as in deep speaker, but with less channels

hackerekcah commented 5 years ago

mean pooling over time axis instead of last frame of LSTM