In pseudo_mos.py, the data_type vector is currently initialized as a one-hot vector with index 1 set to 0. According to the surrounding comment, index 1 should be enabled to include the mentioned default "sarulab" dataset.
|
# we use general version dataset label: sarulab (1) |
|
data_type[1] = 0 |
Suggested change:
In pseudo_mos.py, the data_type vector is currently initialized as a one-hot vector with index 1 set to 0. According to the surrounding comment, index 1 should be enabled to include the mentioned default "sarulab" dataset.
versa/versa/utterance_metrics/pseudo_mos.py
Lines 153 to 154 in a2378f5
Suggested change: