Spaces:
Running
on
L40S
Running
on
L40S
import torch | |
from tqdm import tqdm | |
import torchaudio | |
from third_party.stable_audio_tools.stable_audio_tools.models.autoencoders import create_autoencoder_from_config | |
import numpy as np | |
import os | |
import json | |
def get_model(model_config, path): | |
with open(model_config) as f: | |
model_config = json.load(f) | |
state_dict = torch.load(path) | |
model = create_autoencoder_from_config(model_config) | |
model.load_state_dict(state_dict['state_dict']) | |
return model |