Update app.py

app.py

@@ -1,34 +1,57 @@
+# app.py
 import gradio as gr
-from transformers import pipeline
-import requests
-from io import BytesIO
+from transformers import BlipProcessor, BlipForConditionalGeneration
+from gtts import gTTS
+import io
 from PIL import Image
-import pyttsx3  # Text-to-speech (optional)
 
-# --- Load hosted model ---
-captioner = pipeline("image-to-text", model="Salesforce/blip-image-captioning-base")
+# -------------------------------
+# Load BLIP-base model (lighter version)
+# -------------------------------
+processor = BlipProcessor.from_pretrained("Salesforce/blip-image-captioning-base")
+model = BlipForConditionalGeneration.from_pretrained("Salesforce/blip-image-captioning-base")
 
-# --- Caption & TTS function ---
-def generate_caption_tts(image):
-    # If user uploads a URL
-    if isinstance(image, str):
-        image = Image.open(BytesIO(requests.get(image).content))
-    caption = captioner(image)[0]['generated_text']
+# -------------------------------
+# Generate caption function
+# -------------------------------
+def generate_caption_fn(image):
+    # Convert uploaded image to PIL
+    if not isinstance(image, Image.Image):
+        image = Image.fromarray(image)
+    
+    # BLIP preprocessing
+    inputs = processor(images=image, return_tensors="pt")
     
-    # TTS (optional)
-    tts = pyttsx3.init()
-    tts.say(caption)
-    tts.runAndWait()
+    # Generate caption
+    out = model.generate(**inputs)
+    caption = processor.decode(out[0], skip_special_tokens=True)
     
     return caption
 
-# --- Gradio interface ---
-iface = gr.Interface(
+# -------------------------------
+# Convert text to speech using gTTS
+# -------------------------------
+def text_to_speech(caption):
+    tts = gTTS(text=caption, lang='en')
+    mp3_fp = io.BytesIO()
+    tts.write_to_fp(mp3_fp)
+    mp3_fp.seek(0)
+    return mp3_fp
+
+# -------------------------------
+# Gradio interface: Caption + Audio
+# -------------------------------
+def generate_caption_tts(image):
+    caption = generate_caption_fn(image)
+    audio = text_to_speech(caption)
+    return caption, audio
+
+interface = gr.Interface(
     fn=generate_caption_tts,
-    inputs=gr.Image(type="pil"),
-    outputs="text",
-    title="Image Captioning for Visually Impaired",
-    description="Upload any image and get a descriptive caption."
+    inputs=gr.Image(type="numpy"),
+    outputs=[gr.Textbox(label="Generated Caption"), gr.Audio(type="file", label="TTS Audio")],
+    title="Blind Assistant: Image Captioning",
+    description="Upload an image and get a descriptive caption + speech."
 )
 
-iface.launch()
+interface.launch()