How to use finetune wavlm large model?

[haha010508]

if i want to use finetune wavlm large model in speaker verification job to get best performence, so i need give config_path in this code:

elif model_name == 'wavlm_large':
    config_path = None
    model = ECAPA_TDNN_SMALL(feat_dim=1024, feat_type='wavlm_large', config_path=config_path)

right? but what need i give to config_path? anything is ok?

[haha010508]

i found this method will load two model, one is s3prl upstream model, one is wavlm_large finetune model, total size is 2.4G, how to reduce the size?

[sjoon2455]

have you found out the way out? and, where is the code snippet from?