Spaces:
Sleeping
Sleeping
File size: 4,199 Bytes
5060492 fbd0d2a 5060492 fbd0d2a 5060492 fbd0d2a 5060492 b141261 5060492 4261d7f 5060492 |
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 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 |
import whisper
import gradio as gr
from groq import Groq
from deep_translator import GoogleTranslator
from diffusers import StableDiffusionPipeline
import os
import torch
import openai
from huggingface_hub import InferenceApi
from PIL import Image
import requests
import io
import time
# Set up Groq API key
api_key = os.getenv("groq_key")
client = Groq(api_key=api_key)
# Hugging Face API details for image generation
H_key = os.getenv("hf_key")
API_URL = "https://api-inference.huggingface.co/models/Artples/LAI-ImageGeneration-vSDXL-2"
headers = {"Authorization": f"Bearer {H_key}"}
# Function for querying image generation with retries
def query_image_generation(payload, max_retries=5):
for attempt in range(max_retries):
response = requests.post(API_URL, headers=headers, json=payload)
return response.content
return None
# Function for generating an image from text
def generate_image(prompt):
image_bytes = query_image_generation({"inputs": prompt})
image = Image.open(io.BytesIO(image_bytes)) # Opening the image from bytes
return image
# Updated function for text generation using the new API structure
def generate_creative_text(prompt):
chat_completion = client.chat.completions.create(
messages=[
{"role": "user", "content":prompt}
],
model="llama-3.2-90b-text-preview"
)
chatbot_response = chat_completion.choices[0].message.content
return chatbot_response
def process_audio(audio_path, image_option, creative_text_option):
if audio_path is None:
return "Please upload an audio file.", None, None, None
# Step 1: Transcribe audio
try:
with open(audio_path, "rb") as file:
transcription = client.audio.transcriptions.create(
file=(os.path.basename(audio_path), file.read()),
model="whisper-large-v3",
language="ta",
response_format="verbose_json",
)
tamil_text = transcription.text
except Exception as e:
return f"An error occurred during transcription: {str(e)}", None, None, None
# Step 2: Translate Tamil to English
try:
translator = GoogleTranslator(source='ta', target='en')
translation = translator.translate(tamil_text)
except Exception as e:
return tamil_text, f"An error occurred during translation: {str(e)}", None, None
# Step 3: Generate creative text (if selected)
creative_text = None
if creative_text_option == "Generate Creative Text":
creative_text = generate_creative_text(translation)
# Step 4: Generate image (if selected)
image = None
if image_option == "Generate Image":
image = generate_image(translation)
if image is None:
return tamil_text, translation, creative_text, f"An error occurred during image generation"
return tamil_text, translation, creative_text, image
# Create Gradio interface
with gr.Blocks(theme=gr.themes.Base()) as iface:
gr.Markdown("# Multimodal Application for Vernacular Language Translation and Image Synthesis")
with gr.Row():
with gr.Column():
audio_input = gr.Audio(type="filepath", label="Upload Audio File")
image_option = gr.Dropdown(["Generate Image", "Skip Image"], label="Image Generation", value="Generate Image")
creative_text_option = gr.Dropdown(["Generate Creative Text", "Skip Creative Text"], label="Creative Text Generation", value="Generate Creative Text")
submit_button = gr.Button("Click here to generate")
with gr.Column():
tamil_text_output = gr.Textbox(label="Tamil Transcription")
translation_output = gr.Textbox(label="English Translation")
creative_text_output = gr.Textbox(label="Creative Text")
image_output = gr.Image(label="Generated Image")
submit_button.click(
fn=process_audio,
inputs=[audio_input, image_option, creative_text_option],
outputs=[tamil_text_output, translation_output, creative_text_output, image_output]
)
# Launch the interface
iface.launch() |