Update app.py

app.py

@@ -5,93 +5,151 @@ import torch
 import wavio
 from tqdm import tqdm
 from huggingface_hub import snapshot_download
-from models import AudioDiffusion, DDPMScheduler
-from audioldm.audio.stft import TacotronSTFT
-from audioldm.variational_autoencoder import AutoencoderKL
 from pydub import AudioSegment
 from gradio import Markdown
-
+import uuid
 import torch
-#from diffusers.models.autoencoder_kl import AutoencoderKL
 from diffusers import DiffusionPipeline,AudioPipelineOutput
 from transformers import CLIPTextModel, T5EncoderModel, AutoModel, T5Tokenizer, T5TokenizerFast
 from typing import Union
 from diffusers.utils.torch_utils import randn_tensor
 from tqdm import tqdm
 from TangoFlux import TangoFluxInference
+import torchaudio
 
 
 
-tangoflux = TangoFluxInference(path="declare-lab/TangoFlux")
+tangoflux = TangoFluxInference(name="declare-lab/TangoFlux")
 
 
 
 @spaces.GPU(duration=15)
-def gradio_generate(prompt, output_format, steps, guidance,duration=10):
+def gradio_generate(prompt, steps, guidance,duration=10):
+
+    output = tangoflux.generate(prompt,steps=steps,guidance_scale=guidance,duration=duration)
+    #output = output.to(torch.float32).div(torch.max(torch.abs(output))).clamp(-1, 1).mul(32767).to(torch.int16).cpu()
+    
+    
+    #wavio.write(output_filename, output_wave, rate=44100, sampwidth=2)
+    filename = 'temp.wav'
+    #print(f"Saving audio to file: {unique_filename}")
+
+    # Save to file
+    torchaudio.save(filename, output, 44100)
+   
 
-    output_wave = tangoflux.generate(prompt,steps=steps,guidance=guidance,duration=duration)
-    output_wave = pipe(prompt,steps,guidance) ## Using pipeliine automatically uses flash attention for torch2.0 above
-    #output_wave = tango.generate(prompt, steps, guidance)
-    # output_filename = f"{prompt.replace(' ', '_')}_{steps}_{guidance}"[:250] + ".wav"
-    output_wave = output_wave.audios[0]
-    output_filename = "temp.wav"
-    wavio.write(output_filename, output_wave, rate=16000, sampwidth=2)
+    # Return the path to the generated audio file
+    return filename
 
-    if (output_format == "mp3"):
-        AudioSegment.from_wav("temp.wav").export("temp.mp3", format = "mp3")
-        output_filename = "temp.mp3"
+    #if (output_format == "mp3"):
+     #   AudioSegment.from_wav("temp.wav").export("temp.mp3", format = "mp3")
+      #  output_filename = "temp.mp3"
 
-    return output_filename
+    #return output_filename
 
 description_text = """
-<p><a href="https://huggingface.co/spaces/declare-lab/tango2/blob/main/app.py?duplicate=true"> <img style="margin-top: 0em; margin-bottom: 0em" src="https://bit.ly/3gLdBN6" alt="Duplicate Space"></a> For faster inference without waiting in queue, you may duplicate the space and upgrade to a GPU in the settings. <br/><br/>
-Generate audio using Tango2 by providing a text prompt. Tango2 was built from Tango and was trained on <a href="https://huggingface.co/datasets/declare-lab/audio-alpaca">Audio-alpaca</a>
-<br/><br/> This is the demo for Tango2 for text to audio generation: <a href="https://arxiv.org/abs/2404.09956">Read our paper.</a>
-<p/>
+Generate high quality and faithful audio in just a few seconds using <b>TangoFlux</b> by providing a text prompt. <b>TangoFlux</b> was trained from scratch and underwent alignment to follow human instructions using a new method called <b>Claped-Ranked Preference Optimization (CRPO)</b>.
+<div style="display: flex; gap: 10px; align-items: center;">
+  <a href="https://arxiv.org/abs/2412.21037">
+    <img src="https://img.shields.io/badge/Read_the_Paper-blue?link=https%3A%2F%2Fopenreview.net%2Fattachment%3Fid%3DtpJPlFTyxd%26name%3Dpdf" alt="arXiv">
+  </a>
+  <a href="https://huggingface.co/declare-lab/TangoFlux">
+    <img src="https://img.shields.io/badge/TangoFlux-Huggingface-violet?logo=huggingface&link=https%3A%2F%2Fhuggingface.co%2Fdeclare-lab%2FTangoFlux" alt="Static Badge">
+  </a>
+  <a href="https://tangoflux.github.io/">
+    <img src="https://img.shields.io/badge/Demos-declare--lab-brightred?style=flat" alt="Static Badge">
+  </a>
+  <a href="https://huggingface.co/spaces/declare-lab/TangoFlux">
+    <img src="https://img.shields.io/badge/TangoFlux-Huggingface_Space-8A2BE2?logo=huggingface&link=https%3A%2F%2Fhuggingface.co%2Fspaces%2Fdeclare-lab%2FTangoFlux" alt="Static Badge">
+  </a>
+  <a href="https://huggingface.co/datasets/declare-lab/CRPO">
+    <img src="https://img.shields.io/badge/TangoFlux_Dataset-Huggingface-red?logo=huggingface&link=https%3A%2F%2Fhuggingface.co%2Fdatasets%2Fdeclare-lab%2FTangoFlux" alt="Static Badge">
+  </a>
+  <a href="https://github.com/declare-lab/TangoFlux">
+    <img src="https://img.shields.io/badge/Github-brown?logo=github&link=https%3A%2F%2Fgithub.com%2Fdeclare-lab%2FTangoFlux" alt="Static Badge">
+  </a>
+</div>
 """
 # Gradio input and output components
 input_text = gr.Textbox(lines=2, label="Prompt")
-output_format = gr.Radio(label = "Output format", info = "The file you can dowload", choices = ["mp3", "wav"], value = "wav")
+#output_format = gr.Radio(label = "Output format", info = "The file you can dowload", choices =  "wav"], value = "wav")
 output_audio = gr.Audio(label="Generated Audio", type="filepath")
-denoising_steps = gr.Slider(minimum=10, maximum=100, value=25, step=1, label="Steps", interactive=True)
-guidance_scale = gr.Slider(minimum=1, maximum=10, value=3, step=0.1, label="Guidance Scale", interactive=True)
+denoising_steps = gr.Slider(minimum=10, maximum=100, value=25, step=5, label="Steps", interactive=True)
+guidance_scale = gr.Slider(minimum=1, maximum=10, value=4.5, step=0.5, label="Guidance Scale", interactive=True)
 duration_scale = gr.Slider(minimum=1, maximum=30, value=10, step=1, label="Duration", interactive=True)
 
+
 # Gradio interface
 gr_interface = gr.Interface(
     fn=gradio_generate,
-    inputs=[input_text, output_format, denoising_steps, guidance_scale,duration_scale],
-    outputs=[output_audio],
-    title="TangoFlux: Aligning Diffusion-based Text-to-Audio Generations through Direct Preference Optimization",
+    inputs=[input_text, denoising_steps, guidance_scale,duration_scale],
+    outputs=output_audio,
+    title="TangoFlux: Super Fast and Faithful Text to Audio Generation with Flow Matching and Clap-Ranked Preference Optimization",
     description=description_text,
     allow_flagging=False,
     examples=[
-        ["Quiet speech and then and airplane flying away"],
-        ["A bicycle peddling on dirt and gravel followed by a man speaking then laughing"],
-        ["Ducks quack and water splashes with some animal screeching in the background"],
-        ["Describe the sound of the ocean"],
-        ["A woman and a baby are having a conversation"],
-        ["A man speaks followed by a popping noise and laughter"],
-        ["A cup is filled from a faucet"],
-        ["An audience cheering and clapping"],
-        ["Rolling thunder with lightning strikes"],
-        ["A dog barking and a cat mewing and a racing car passes by"],
-        ["Gentle water stream, birds chirping and sudden gun shot"],
-        ["A man talking followed by a goat baaing then a metal gate sliding shut as ducks quack and wind blows into a microphone."],
-        ["A dog barking"],
-        ["A cat meowing"],
-        ["Wooden table tapping sound while water pouring"],
-        ["Applause from a crowd with distant clicking and a man speaking over a loudspeaker"],
-        ["two gunshots followed by birds flying away while chirping"],
-        ["Whistling with birds chirping"],
-        ["A person snoring"],
-        ["Motor vehicles are driving with loud engines and a person whistles"],
-        ["People cheering in a stadium while thunder and lightning strikes"],
-        ["A helicopter is in flight"],
-        ["A dog barking and a man talking and a racing car passes by"],
+["A parade marches through a town square, with drumbeats pounding, children clapping, and a horse neighing amidst the commotion"],
+["A soccer ball hits a goalpost with a metallic clang, followed by cheers, clapping, and the distant hum of a commentator’s voice"],
+["The deep growl of an alligator ripples through the swamp as reeds sway with a soft rustle and a turtle splashes into the murky water"],        
+["A basketball bounces rhythmically on a court, shoes squeak against the floor, and a referee’s whistle cuts through the air"],
+["A train conductor blows a sharp whistle, metal wheels screech on the rails, and passengers murmur while settling into their seats"],
+["A fork scrapes a plate, water drips slowly into a sink, and the faint hum of a refrigerator lingers in the background"],
+["Alarms blare with rising urgency as fragments clatter against a metallic hull, interrupted by a faint hiss of escaping air"],
+["Tiny pops and hisses of chemical reactions intermingle with the rhythmic pumping of a centrifuge and the soft whirr of air filtration"],
+["A train conductor blows a sharp whistle, metal wheels screech on the rails, and passengers murmur while settling into their seats"],
+["Simulate a forest ambiance with birds chirping and wind rustling through the leaves"],
+["Quiet whispered conversation gradually fading into distant jet engine roar diminishing into silence"],
+["Clear sound of bicycle tires crunching on loose gravel and dirt, followed by deep male laughter echoing"],
+["Multiple ducks quacking loudly with splashing water and piercing wild animal shriek in background"],
+["Create a serene soundscape of a quiet beach at sunset"],
+["A pile of coins spills onto a wooden table with a metallic clatter, followed by the hushed murmur of a tavern crowd and the creak of a swinging door"],
+["Generate an energetic and bustling city street scene with distant traffic and close conversations"],
+["Powerful ocean waves crashing and receding on sandy beach with distant seagulls"],
+["Gentle female voice cooing and baby responding with happy gurgles and giggles"],
+["Clear male voice speaking, sharp popping sound, followed by genuine group laughter"],
+["Stream of water hitting empty ceramic cup, pitch rising as cup fills up"],
+["Massive crowd erupting in thunderous applause and excited cheering"],
+["Deep rolling thunder with bright lightning strikes crackling through sky"],
+["Aggressive dog barking and distressed cat meowing as racing car roars past at high speed"],
+["Peaceful stream bubbling and birds singing, interrupted by sudden explosive gunshot"],
+["Man speaking outdoors, goat bleating loudly, metal gate scraping closed, ducks quacking frantically, wind howling into microphone"],
+["Series of loud aggressive dog barks echoing"],
+["Multiple distinct cat meows at different pitches"],
+["Rhythmic wooden table tapping overlaid with steady water pouring sound"],
+["Sustained crowd applause with camera clicks and amplified male announcer voice"],
+["Two sharp gunshots followed by panicked birds taking flight with rapid wing flaps"],
+["Melodic human whistling harmonizing with natural birdsong"],
+["Deep rhythmic snoring with clear breathing patterns"],
+["Multiple racing engines revving and accelerating with sharp whistle piercing through"],
+["Massive stadium crowd cheering as thunder crashes and lightning strikes"],
+["Heavy helicopter blades chopping through air with engine and wind noise"],
+["Dog barking excitedly and man shouting as race car engine roars past"],
+["Quiet speech and then and airplane flying away"],
+["A bicycle peddling on dirt and gravel followed by a man speaking then laughing"],
+["Ducks quack and water splashes with some animal screeching in the background"],
+["Describe the sound of the ocean"],
+["A woman and a baby are having a conversation"],
+["A man speaks followed by a popping noise and laughter"],
+["A cup is filled from a faucet"],
+["An audience cheering and clapping"],
+["Rolling thunder with lightning strikes"],
+["A dog barking and a cat mewing and a racing car passes by"],
+["Gentle water stream, birds chirping and sudden gun shot"],
+["A dog barking"],
+["A cat meowing"],
+["Wooden table tapping sound while water pouring"],
+["Applause from a crowd with distant clicking and a man speaking over a loudspeaker"],
+["two gunshots followed by birds flying away while chirping"],
+["Whistling with birds chirping"],
+["A person snoring"],
+["Motor vehicles are driving with loud engines and a person whistles"],
+["People cheering in a stadium while thunder and lightning strikes"],
+["A helicopter is in flight"],
+["A dog barking and a man talking and a racing car passes by"],
     ],
     cache_examples="lazy", # Turn on to cache.
 )
 
-# Launch Gradio app
-gr_interface.queue(10).launch()
+
+
+gr_interface.queue(15).launch()