Update app.py

app.py

@@ -10,43 +10,6 @@ from transformers import Qwen2VLForConditionalGeneration, AutoProcessor, TextIte
 from transformers.image_utils import load_image
 import time
 
-# =============================================================================
-# New imports and helper classes for image generation
-# =============================================================================
-try:
-    # We use Hugging Face’s InferenceClient as a generic image-generation API client.
-    from huggingface_hub import InferenceClient as HFInferenceClient
-except ImportError:
-    HFInferenceClient = None
-
-# A simple wrapper client for our primary image-generation space.
-class Client:
-    def __init__(self, repo_id):
-        self.repo_id = repo_id
-        if HFInferenceClient is not None:
-            self.client = HFInferenceClient(repo_id)
-        else:
-            self.client = None
-
-    def predict(self, task, arg2, prompt, api_name):
-        if self.client is not None:
-            # Here we assume that calling the client with the prompt returns an image.
-            # (Depending on your API, you might need to adjust parameters.)
-            return self.client(prompt)
-        else:
-            raise Exception("HFInferenceClient not available")
-
-def image_gen(prompt):
-    """
-    Uses the STABLE-HAMSTER space to generate an image based on the prompt.
-    """
-    client = Client("prithivMLmods/STABLE-HAMSTER")
-    return client.predict("Image Generation", None, prompt, api_name="/stable_hamster")
-
-# =============================================================================
-# Original Code (with modifications below)
-# =============================================================================
-
 DESCRIPTION = """
 # QwQ Edge 💬
 """
@@ -123,46 +86,13 @@ def generate(
     repetition_penalty: float = 1.2,
 ):
     """
-    Generates chatbot responses with support for multimodal input, TTS, and now image generation.
+    Generates chatbot responses with support for multimodal input and TTS.
     If the query starts with an @tts command (e.g. "@tts1"), previous chat history is cleared.
-    If the query starts with an @image command, the image generation branch is used.
     """
     text = input_dict["text"]
     files = input_dict.get("files", [])
 
-    # -------------------------------------------------------------------------
-    # NEW: Check for image generation command (@image)
-    # -------------------------------------------------------------------------
-    image_prefix = "@image"
-    if text.strip().lower().startswith(image_prefix):
-        # Remove the prefix and any extra whitespace
-        query = text[len(image_prefix):].strip()
-        yield "Generating Image, Please wait 10 sec..."
-        try:
-            image = image_gen(query)
-            # If the API returns a tuple (as in the snippet) use the second element;
-            # otherwise assume it returns an image directly.
-            if isinstance(image, (list, tuple)) and len(image) > 1:
-                yield gr.Image(image[1])
-            else:
-                yield gr.Image(image)
-        except Exception as e:
-            yield "Error in primary image generation, trying fallback..."
-            try:
-                # Use the fallback image generation client.
-                if HFInferenceClient is not None:
-                    client_flux = HFInferenceClient("black-forest-labs/FLUX.1-schnell")
-                    image = client_flux.text_to_image(query)
-                    yield gr.Image(image)
-                else:
-                    yield "Fallback client not available."
-            except Exception as fallback_error:
-                yield f"Error in image generation: {str(fallback_error)}"
-        return  # End execution after processing the image-generation request.
-
-    # -------------------------------------------------------------------------
-    # Continue with the original processing (image files, TTS, or text conversation)
-    # -------------------------------------------------------------------------
+    # Process image files if provided
     if len(files) > 1:
         images = [load_image(image) for image in files]
     elif len(files) == 1:
@@ -173,7 +103,7 @@ def generate(
     tts_prefix = "@tts"
     is_tts = any(text.strip().lower().startswith(f"{tts_prefix}{i}") for i in range(1, 3))
     voice_index = next((i for i in range(1, 3) if text.strip().lower().startswith(f"{tts_prefix}{i}")), None)
-
+    
     if is_tts and voice_index:
         voice = TTS_VOICES[voice_index - 1]
         text = text.replace(f"{tts_prefix}{voice_index}", "").strip()
@@ -258,7 +188,6 @@ demo = gr.ChatInterface(
         ["A train travels 60 kilometers per hour. If it travels for 5 hours, how far will it travel in total?"],
         ["Write a Python function to check if a number is prime."],
         ["@tts2 What causes rainbows to form?"],
-        ["@image A beautiful sunset over a mountain range"],
     ],
     cache_examples=False,
     type="messages",