Spaces:
Runtime error
Runtime error
Fabrice-TIERCELIN
commited on
Commit
•
f1dff10
1
Parent(s):
d2f25e6
KO -> more comment
Browse files
app.py
CHANGED
@@ -51,21 +51,21 @@ class Tango:
|
|
51 |
self.stft = TacotronSTFT(**stft_config).to(device)
|
52 |
self.model = AudioDiffusion(**main_config).to(device)
|
53 |
|
54 |
-
vae_weights = torch.load("{}/pytorch_model_vae.bin".format(path), map_location = device)
|
55 |
-
stft_weights = torch.load("{}/pytorch_model_stft.bin".format(path), map_location = device)
|
56 |
-
main_weights = torch.load("{}/pytorch_model_main.bin".format(path), map_location = device)
|
57 |
-
|
58 |
-
self.vae.load_state_dict(vae_weights)
|
59 |
-
self.stft.load_state_dict(stft_weights)
|
60 |
-
self.model.load_state_dict(main_weights)
|
61 |
-
|
62 |
-
print ("Successfully loaded checkpoint from:", name)
|
63 |
-
|
64 |
-
self.vae.eval()
|
65 |
-
self.stft.eval()
|
66 |
-
self.model.eval()
|
67 |
-
|
68 |
-
self.scheduler = DDPMScheduler.from_pretrained(main_config["scheduler_name"], subfolder = "scheduler")
|
69 |
|
70 |
def chunks(self, lst, n):
|
71 |
# Yield successive n-sized chunks from a list
|
|
|
51 |
self.stft = TacotronSTFT(**stft_config).to(device)
|
52 |
self.model = AudioDiffusion(**main_config).to(device)
|
53 |
|
54 |
+
# vae_weights = torch.load("{}/pytorch_model_vae.bin".format(path), map_location = device)
|
55 |
+
# stft_weights = torch.load("{}/pytorch_model_stft.bin".format(path), map_location = device)
|
56 |
+
# main_weights = torch.load("{}/pytorch_model_main.bin".format(path), map_location = device)
|
57 |
+
#
|
58 |
+
# self.vae.load_state_dict(vae_weights)
|
59 |
+
# self.stft.load_state_dict(stft_weights)
|
60 |
+
# self.model.load_state_dict(main_weights)
|
61 |
+
#
|
62 |
+
# print ("Successfully loaded checkpoint from:", name)
|
63 |
+
#
|
64 |
+
# self.vae.eval()
|
65 |
+
# self.stft.eval()
|
66 |
+
# self.model.eval()
|
67 |
+
#
|
68 |
+
# self.scheduler = DDPMScheduler.from_pretrained(main_config["scheduler_name"], subfolder = "scheduler")
|
69 |
|
70 |
def chunks(self, lst, n):
|
71 |
# Yield successive n-sized chunks from a list
|