Training the baseline (without VAC, SMKD)

VIPL-SLP / VAC_CSLR

Visual Alignment Constraint for Continuous Sign Language Recognition. ( ICCV 2021)

https://openaccess.thecvf.com/content/ICCV2021/html/Min_Visual_Alignment_Constraint_for_Continuous_Sign_Language_Recognition_ICCV_2021_paper.html

Apache License 2.0

116 stars 19 forks source link

Training the baseline (without VAC, SMKD) #44

Closed atonyo11 closed 7 months ago

atonyo11 commented 7 months ago

The VAC result is corresponding to the setting ofloss_weights: SeqCTC: 1.0, ConvCTC: 1.0. In addition to that, the VAC+SMKD adopt the setting of model_args: share_classifier: True, weight_norm: True.

With the default setting, is it training for the baseline?

ycmin95 commented 7 months ago

Yes, we adopt the network trained with single CTC loss as the baseline.

atonyo11 commented 7 months ago

Thanks!