Spaces:
Runtime error
Runtime error
File size: 2,372 Bytes
23161e3 8c8d51f 23161e3 8c8d51f |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 |
import gradio as gr
from transformers import pipeline
import sounddevice as sd # For microphone input
from diffusers import DiffusionPipeline
# Load the diffuser pipeline with LORA weights
pipeline = DiffusionPipeline.from_pretrained("CompVis/stable-diffusion-v1-4")
pipeline.load_lora_weights("MdEndan/tinysketch-fine-tuned")
def generate_image(text):
"""Converts speech to text, generates an image using diffuser pipeline,
and displays the result."""
# Speech-to-text using a pre-trained pipeline (replace with your choice)
speech_pipe = pipeline("automatic-speech-recognition")
try:
# Record audio from microphone (adjust duration and sample rate if needed)
duration = 5 # Record for 5 seconds
fs = 16000 # Sample rate
print("Speak now...")
myrecording = sd.rec(duration * fs, samplerate=fs, channels=1)
sd.wait()
print("Recording stopped")
# Convert audio to WAV for compatibility with some pipelines
sd.write("recording.wav", myrecording, fs)
# Transcribe speech
with open("recording.wav", "rb") as f:
audio_bytes = f.read()
speech_output = speech_pipe(audio_bytes, return_tensors="pt")["sequences"]
text = speech_output[0].tolist() # Extract the transcribed text
except Exception as e:
print(f"Error during speech recognition: {e}")
text = "Error: Speech recognition failed."
# Ensure text input is a string
if not isinstance(text, str):
text = str(text)
# Generate image using diffuser pipeline
try:
image = pipeline(text).images[0]
return image
except Exception as e:
print(f"Error during image generation: {e}")
return None
# Gradio interface with microphone and image display
interface = gr.Interface(
fn=generate_image,
inputs=gr.Audio(sources=["microphone"]),
outputs=gr.Image(thumbnail=True),
title="Speak & Create: Text-to-Image with Microphone Input (LORA)",
description="Speak your description and see an image generated using a fine-tuned model!",
)
# Handle potential errors during Gradio launch
try:
# Request access to the microphone (might require user permission)
interface.launch(share=True, capture_audio=True)
except Exception as e:
print(f"Error launching Gradio interface: {e}")
|