Spaces:
Running
Running
yuancwang
commited on
Commit
•
f8b1a1a
1
Parent(s):
f3af09b
commit
Browse files
app.py
CHANGED
@@ -19,6 +19,7 @@ from scipy.io.wavfile import write
|
|
19 |
from utils.util import load_config
|
20 |
import gradio as gr
|
21 |
|
|
|
22 |
class AttrDict(dict):
|
23 |
def __init__(self, *args, **kwargs):
|
24 |
super(AttrDict, self).__init__(*args, **kwargs)
|
@@ -35,16 +36,20 @@ def build_autoencoderkl(cfg, device):
|
|
35 |
autoencoderkl.eval()
|
36 |
return autoencoderkl
|
37 |
|
|
|
38 |
def build_textencoder(device):
|
39 |
-
|
40 |
-
|
41 |
-
|
42 |
-
|
|
|
|
|
43 |
text_encoder = text_encoder.to(device=device)
|
44 |
text_encoder.requires_grad_(requires_grad=False)
|
45 |
text_encoder.eval()
|
46 |
return tokenizer, text_encoder
|
47 |
|
|
|
48 |
def build_vocoder(device):
|
49 |
config_file = os.path.join("ckpts/tta/hifigan_checkpoints/config.json")
|
50 |
with open(config_file) as f:
|
@@ -58,12 +63,13 @@ def build_vocoder(device):
|
|
58 |
vocoder.load_state_dict(checkpoint_dict["generator"])
|
59 |
return vocoder
|
60 |
|
|
|
61 |
def build_model(cfg):
|
62 |
model = AudioLDM(cfg.model.audioldm)
|
63 |
return model
|
64 |
|
65 |
-
def get_text_embedding(text, tokenizer, text_encoder, device):
|
66 |
|
|
|
67 |
prompt = [text]
|
68 |
|
69 |
text_input = tokenizer(
|
@@ -73,28 +79,24 @@ def get_text_embedding(text, tokenizer, text_encoder, device):
|
|
73 |
padding="do_not_pad",
|
74 |
return_tensors="pt",
|
75 |
)
|
76 |
-
text_embeddings = text_encoder(
|
77 |
-
text_input.input_ids.to(device)
|
78 |
-
)[0]
|
79 |
|
80 |
max_length = text_input.input_ids.shape[-1]
|
81 |
uncond_input = tokenizer(
|
82 |
[""] * 1, padding="max_length", max_length=max_length, return_tensors="pt"
|
83 |
)
|
84 |
-
uncond_embeddings = text_encoder(
|
85 |
-
uncond_input.input_ids.to(device)
|
86 |
-
)[0]
|
87 |
text_embeddings = torch.cat([uncond_embeddings, text_embeddings])
|
88 |
|
89 |
return text_embeddings
|
90 |
-
|
|
|
91 |
def tta_inference(
|
92 |
-
|
93 |
-
|
94 |
-
|
95 |
):
|
96 |
-
|
97 |
-
device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
|
98 |
|
99 |
os.environ["WORK_DIR"] = "./"
|
100 |
cfg = load_config("egs/tta/audioldm/exp_config.json")
|
@@ -126,7 +128,6 @@ def tta_inference(
|
|
126 |
|
127 |
noise_scheduler.set_timesteps(num_steps)
|
128 |
|
129 |
-
|
130 |
latents = torch.randn(
|
131 |
(
|
132 |
1,
|
@@ -189,6 +190,7 @@ def tta_inference(
|
|
189 |
|
190 |
return os.path.join("result", text + ".wav")
|
191 |
|
|
|
192 |
demo_inputs = [
|
193 |
gr.Textbox(
|
194 |
value="birds singing and a man whistling",
|
@@ -218,15 +220,8 @@ demo = gr.Interface(
|
|
218 |
fn=tta_inference,
|
219 |
inputs=demo_inputs,
|
220 |
outputs=demo_outputs,
|
221 |
-
title="Amphion Text to Audio"
|
222 |
)
|
223 |
|
224 |
if __name__ == "__main__":
|
225 |
demo.launch()
|
226 |
-
|
227 |
-
|
228 |
-
|
229 |
-
|
230 |
-
|
231 |
-
|
232 |
-
|
|
|
19 |
from utils.util import load_config
|
20 |
import gradio as gr
|
21 |
|
22 |
+
|
23 |
class AttrDict(dict):
|
24 |
def __init__(self, *args, **kwargs):
|
25 |
super(AttrDict, self).__init__(*args, **kwargs)
|
|
|
36 |
autoencoderkl.eval()
|
37 |
return autoencoderkl
|
38 |
|
39 |
+
|
40 |
def build_textencoder(device):
|
41 |
+
try:
|
42 |
+
tokenizer = AutoTokenizer.from_pretrained("t5-base", model_max_length=512)
|
43 |
+
text_encoder = T5EncoderModel.from_pretrained("t5-base")
|
44 |
+
except:
|
45 |
+
tokenizer = AutoTokenizer.from_pretrained("ckpts/tta/tokenizer")
|
46 |
+
text_encoder = T5EncoderModel.from_pretrained("ckpts/tta/text_encoder")
|
47 |
text_encoder = text_encoder.to(device=device)
|
48 |
text_encoder.requires_grad_(requires_grad=False)
|
49 |
text_encoder.eval()
|
50 |
return tokenizer, text_encoder
|
51 |
|
52 |
+
|
53 |
def build_vocoder(device):
|
54 |
config_file = os.path.join("ckpts/tta/hifigan_checkpoints/config.json")
|
55 |
with open(config_file) as f:
|
|
|
63 |
vocoder.load_state_dict(checkpoint_dict["generator"])
|
64 |
return vocoder
|
65 |
|
66 |
+
|
67 |
def build_model(cfg):
|
68 |
model = AudioLDM(cfg.model.audioldm)
|
69 |
return model
|
70 |
|
|
|
71 |
|
72 |
+
def get_text_embedding(text, tokenizer, text_encoder, device):
|
73 |
prompt = [text]
|
74 |
|
75 |
text_input = tokenizer(
|
|
|
79 |
padding="do_not_pad",
|
80 |
return_tensors="pt",
|
81 |
)
|
82 |
+
text_embeddings = text_encoder(text_input.input_ids.to(device))[0]
|
|
|
|
|
83 |
|
84 |
max_length = text_input.input_ids.shape[-1]
|
85 |
uncond_input = tokenizer(
|
86 |
[""] * 1, padding="max_length", max_length=max_length, return_tensors="pt"
|
87 |
)
|
88 |
+
uncond_embeddings = text_encoder(uncond_input.input_ids.to(device))[0]
|
|
|
|
|
89 |
text_embeddings = torch.cat([uncond_embeddings, text_embeddings])
|
90 |
|
91 |
return text_embeddings
|
92 |
+
|
93 |
+
|
94 |
def tta_inference(
|
95 |
+
text,
|
96 |
+
guidance_scale=4,
|
97 |
+
diffusion_steps=100,
|
98 |
):
|
99 |
+
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
|
|
|
100 |
|
101 |
os.environ["WORK_DIR"] = "./"
|
102 |
cfg = load_config("egs/tta/audioldm/exp_config.json")
|
|
|
128 |
|
129 |
noise_scheduler.set_timesteps(num_steps)
|
130 |
|
|
|
131 |
latents = torch.randn(
|
132 |
(
|
133 |
1,
|
|
|
190 |
|
191 |
return os.path.join("result", text + ".wav")
|
192 |
|
193 |
+
|
194 |
demo_inputs = [
|
195 |
gr.Textbox(
|
196 |
value="birds singing and a man whistling",
|
|
|
220 |
fn=tta_inference,
|
221 |
inputs=demo_inputs,
|
222 |
outputs=demo_outputs,
|
223 |
+
title="Amphion Text to Audio",
|
224 |
)
|
225 |
|
226 |
if __name__ == "__main__":
|
227 |
demo.launch()
|
|
|
|
|
|
|
|
|
|
|
|
|
|