Is the decoder of the AE2 freezed?

[leonardozcm]

ASFM-Net/train.py

Line 69 in 8d1c282

train_var_list = [var for var in tf.trainable_variables() if 'AE1' not in var.name]

hi, i'm not very familiar with tf, i wonder where did you freeze the decoder of AE2? i think in your paper your decoder of AE2 has the fixed parameteres, which are keeping the same with AE1's.

ASFM-Net/train.py

Line 191 in 8d1c282

parser.add_argument('--num_input_points', type=int, default=4096)

what's more, what is the input_size in training stage of pcn datasets?