search

JavierGalvan9 / V1_GLIF_model

TensorFlow implementation of the new version from the V1 column from the Allen Institute
0 stars 4 forks source link

Fixing core masking + change in OSI loss function #14

Closed shixnya closed 7 months ago

shixnya commented 7 months ago
  1. The target firing rate distribution was fixed.
    • Previously, it was inflated at the lowest value. In an extreme case (where there are few target cells), it caused the FR to be zero for half of the population. (the figures below illustrates the fix)
  2. The OSI loss is modified to include an offset that is proportional to the firing rates. (
    • With the previous implementation, the OSI loss is minimized at zero spikes (and fight against the rate-loss). With the new implementation, pure increase in firing rates is not penalized.
  3. The neuron IDs are filtered before making the target firing rates. (it needs to be done before sampling the target firing rates)
  4. The voltage loss also uses core_masks
  5. Fixed the incorrect sampling of the target firing rates.
shixnya commented 7 months ago

Also, the NEST data is updated with the new FilterNet inputs.