Spaces:
Running
Running
Started saving latents during voice creation
Browse files
app.py
CHANGED
@@ -1,5 +1,3 @@
|
|
1 |
-
# AGPL: a notification must be added stating that changes have been made to that file.
|
2 |
-
|
3 |
import os
|
4 |
import shutil
|
5 |
from pathlib import Path
|
@@ -24,7 +22,6 @@ from app_utils.funcs import (
|
|
24 |
load_voice_conditionings,
|
25 |
)
|
26 |
|
27 |
-
|
28 |
LATENT_MODES = [
|
29 |
"Tortoise original (bad)",
|
30 |
"average per 4.27s (broken on small files)",
|
@@ -71,11 +68,13 @@ def main():
|
|
71 |
bytes_data = uploaded_file.read()
|
72 |
with open(f"{voices_dir}voice_sample{index}.wav", "wb") as wav_file:
|
73 |
wav_file.write(bytes_data)
|
74 |
-
|
75 |
-
#
|
76 |
-
voice_samples, conditioning_latents =
|
77 |
-
|
78 |
-
|
|
|
|
|
79 |
st.session_state["text_input_key"] = str(randint(1000, 100000000))
|
80 |
st.session_state["file_uploader_key"] = str(randint(1000, 100000000))
|
81 |
st.experimental_rerun()
|
@@ -109,6 +108,7 @@ def main():
|
|
109 |
help="Which voice preset to use.",
|
110 |
index=1,
|
111 |
)
|
|
|
112 |
with st.expander("Advanced"):
|
113 |
col1, col2 = st.columns(2)
|
114 |
with col1:
|
@@ -126,7 +126,6 @@ def main():
|
|
126 |
)
|
127 |
sampler = st.radio(
|
128 |
"Sampler",
|
129 |
-
#SAMPLERS,
|
130 |
["dpm++2m", "p", "ddim"],
|
131 |
help="Diffusion sampler. Note that dpm++2m is experimental and typically requires more steps.",
|
132 |
index=1,
|
|
|
|
|
|
|
1 |
import os
|
2 |
import shutil
|
3 |
from pathlib import Path
|
|
|
22 |
load_voice_conditionings,
|
23 |
)
|
24 |
|
|
|
25 |
LATENT_MODES = [
|
26 |
"Tortoise original (bad)",
|
27 |
"average per 4.27s (broken on small files)",
|
|
|
68 |
bytes_data = uploaded_file.read()
|
69 |
with open(f"{voices_dir}voice_sample{index}.wav", "wb") as wav_file:
|
70 |
wav_file.write(bytes_data)
|
71 |
+
|
72 |
+
# Generate conditioning latents and samples here
|
73 |
+
voice_samples, conditioning_latents = generate_conditioning(voices_dir)
|
74 |
+
|
75 |
+
# Save the conditioning latents and samples
|
76 |
+
save_conditioning(voices_dir, voice_samples, conditioning_latents)
|
77 |
+
|
78 |
st.session_state["text_input_key"] = str(randint(1000, 100000000))
|
79 |
st.session_state["file_uploader_key"] = str(randint(1000, 100000000))
|
80 |
st.experimental_rerun()
|
|
|
108 |
help="Which voice preset to use.",
|
109 |
index=1,
|
110 |
)
|
111 |
+
|
112 |
with st.expander("Advanced"):
|
113 |
col1, col2 = st.columns(2)
|
114 |
with col1:
|
|
|
126 |
)
|
127 |
sampler = st.radio(
|
128 |
"Sampler",
|
|
|
129 |
["dpm++2m", "p", "ddim"],
|
130 |
help="Diffusion sampler. Note that dpm++2m is experimental and typically requires more steps.",
|
131 |
index=1,
|