Handle seed (#5)

- Handle seed (133ca117fa060cb441341f74a339b42d01f99a61)


Co-authored-by: Fabrice TIERCELIN <[email protected]>

app.py

@@ -1,5 +1,6 @@
 import gradio as gr
 import json
+import random
 import torch
 import wavio
 from tqdm import tqdm
@@ -11,6 +12,8 @@ from pydub import AudioSegment
 from gradio import Markdown
 import spaces
 
+max_64_bit_int = 2**63 - 1
+
 # Automatic device detection
 if torch.cuda.is_available():
     device_type = "cuda"
@@ -89,8 +92,16 @@ def gradio_generate(
     output_format="wav",
     output_number=3,
     steps=100,
-    guidance=3
+    guidance=3,
+    is_randomize_seed=True,
+    seed=123
 ):
+    if is_randomize_seed:
+        seed = random.randint(0, max_64_bit_int)
+
+    random.seed(seed)
+    torch.manual_seed(seed)
+
     output_wave = tango.generate(prompt, steps, guidance, output_number)
     # output_filename = f"{prompt.replace(' ', '_')}_{steps}_{guidance}"[:250] + ".wav"
 
@@ -161,39 +172,41 @@ output_audio_2 = gr.Audio(label="Generated Audio #2/3", type="filepath")
 output_audio_3 = gr.Audio(label="Generated Audio #3/3", type="filepath")
 denoising_steps = gr.Slider(minimum=10, maximum=200, value=100, step=1, label="Steps", interactive=True)
 guidance_scale = gr.Slider(minimum=1, maximum=10, value=3, step=0.1, label="Guidance Scale", interactive=True)
+randomize_seed = gr.Checkbox(label = "\U0001F3B2 Randomize seed", value = True, info = "If checked, result is always different")
+seed = gr.Slider(minimum = 0, maximum = max_64_bit_int, step = 1, randomize = True, label = "Seed")
 
 # Gradio interface
 gr_interface = gr.Interface(
     fn=gradio_generate,
-    inputs=[input_text, output_format, output_number, denoising_steps, guidance_scale],
+    inputs=[input_text, output_format, output_number, denoising_steps, guidance_scale, randomize_seed, seed],
     outputs=[output_audio_1, output_audio_2, output_audio_3],
     title="Tango 2: Aligning Diffusion-based Text-to-Audio Generations through Direct Preference Optimization",
     description=description_text,
     allow_flagging=False,
     examples=[
-        ["Quiet speech and then and airplane flying away", "wav", 1, 100, 3],
-        ["A bicycle peddling on dirt and gravel followed by a man speaking then laughing", "wav", 1, 100, 3],
-        ["Ducks quack and water splashes with some animal screeching in the background", "wav", 1, 100, 3],
-        ["Describe the sound of the ocean", "wav", 1, 100, 3],
-        ["A woman and a baby are having a conversation", "wav", 1, 100, 3],
-        ["A man speaks followed by a popping noise and laughter", "wav", 1, 100, 3],
-        ["A cup is filled from a faucet", "wav", 1, 100, 3],
-        ["An audience cheering and clapping", "wav", 1, 100, 3],
-        ["Rolling thunder with lightning strikes", "wav", 1, 100, 3],
-        ["A dog barking and a cat mewing and a racing car passes by", "wav", 1, 100, 3],
-        ["Gentle water stream, birds chirping and sudden gun shot", "wav", 1, 100, 3],
-        ["A man talking followed by a goat baaing then a metal gate sliding shut as ducks quack and wind blows into a microphone.", "wav", 1, 100, 3],
-        ["A dog barking", "wav", 1, 100, 3],
-        ["A cat meowing", "wav", 1, 100, 3],
-        ["Wooden table tapping sound while water pouring", "wav", 1, 100, 3],
-        ["Applause from a crowd with distant clicking and a man speaking over a loudspeaker", "wav", 1, 100, 3],
-        ["two gunshots followed by birds flying away while chirping", "wav", 1, 100, 3],
-        ["Whistling with birds chirping", "wav", 1, 100, 3],
-        ["A person snoring", "wav", 1, 100, 3],
-        ["Motor vehicles are driving with loud engines and a person whistles", "wav", 1, 100, 3],
-        ["People cheering in a stadium while thunder and lightning strikes", "wav", 1, 100, 3],
-        ["A helicopter is in flight", "wav", 1, 100, 3],
-        ["A dog barking and a man talking and a racing car passes by", "wav", 1, 100, 3],
+        ["Quiet speech and then and airplane flying away", "wav", 1, 100, 3, False, 123],
+        ["A bicycle peddling on dirt and gravel followed by a man speaking then laughing", "wav", 1, 100, 3, False, 123],
+        ["Ducks quack and water splashes with some animal screeching in the background", "wav", 1, 100, 3, False, 123],
+        ["Describe the sound of the ocean", "wav", 1, 100, 3, False, 123],
+        ["A woman and a baby are having a conversation", "wav", 1, 100, 3, False, 123],
+        ["A man speaks followed by a popping noise and laughter", "wav", 1, 100, 3, False, 123],
+        ["A cup is filled from a faucet", "wav", 1, 100, 3, False, 123],
+        ["An audience cheering and clapping", "wav", 1, 100, 3, False, 123],
+        ["Rolling thunder with lightning strikes", "wav", 1, 100, 3, False, 123],
+        ["A dog barking and a cat mewing and a racing car passes by", "wav", 1, 100, 3, False, 123],
+        ["Gentle water stream, birds chirping and sudden gun shot", "wav", 1, 100, 3, False, 123],
+        ["A man talking followed by a goat baaing then a metal gate sliding shut as ducks quack and wind blows into a microphone.", "wav", 1, 100, 3, False, 123],
+        ["A dog barking", "wav", 1, 100, 3, False, 123],
+        ["A cat meowing", "wav", 1, 100, 3, False, 123],
+        ["Wooden table tapping sound while water pouring", "wav", 1, 100, 3, False, 123],
+        ["Applause from a crowd with distant clicking and a man speaking over a loudspeaker", "wav", 1, 100, 3, False, 123],
+        ["two gunshots followed by birds flying away while chirping", "wav", 1, 100, 3, False, 123],
+        ["Whistling with birds chirping", "wav", 1, 100, 3, False, 123],
+        ["A person snoring", "wav", 1, 100, 3, False, 123],
+        ["Motor vehicles are driving with loud engines and a person whistles", "wav", 1, 100, 3, False, 123],
+        ["People cheering in a stadium while thunder and lightning strikes", "wav", 1, 100, 3, False, 123],
+        ["A helicopter is in flight", "wav", 1, 100, 3, False, 123],
+        ["A dog barking and a man talking and a racing car passes by", "wav", 1, 100, 3, False, 123],
     ],
     cache_examples="lazy", # Turn on to cache.
 )