Update app.py

app.py

@@ -22,7 +22,6 @@ from autogen_ext.models.openai import OpenAIChatCompletionClient
 from autogen_ext.models.ollama import OllamaChatCompletionClient
 from autogen_ext.models.azure import AzureAIChatCompletionClient
 from azure.core.credentials import AzureKeyCredential
-from transformers import AutoModelForCausalLM, AutoTokenizer
 import traceback
 import soundfile as sf
 import tempfile
@@ -157,41 +156,6 @@ def html_with_progress(label, progress):
     </div>
     """
 
-# Qwen2.5-3B-Instruct Client
-class QwenChatCompletionClient:
-    def __init__(self, model_name="Qwen/Qwen2.5-3B-Instruct"):
-        self.model = AutoModelForCausalLM.from_pretrained(
-            model_name,
-            torch_dtype="auto",
-            device_map="auto"
-        )
-        self.tokenizer = AutoTokenizer.from_pretrained(model_name)
-        logger.info(f"Initialized Qwen model: {model_name}")
-
-    def create_chat_completion(self, messages, max_tokens=512, temperature=0.7, top_p=0.9):
-        try:
-            text = self.tokenizer.apply_chat_template(
-                messages,
-                tokenize=False,
-                add_generation_prompt=True
-            )
-            model_inputs = self.tokenizer([text], return_tensors="pt").to(self.model.device)
-
-            generated_ids = self.model.generate(
-                **model_inputs,
-                max_new_tokens=max_tokens,
-                temperature=temperature,
-                top_p=top_p
-            )
-            generated_ids = [
-                output_ids[len(input_ids):] for input_ids, output_ids in zip(model_inputs.input_ids, generated_ids)
-            ]
-            response = self.tokenizer.batch_decode(generated_ids, skip_special_tokens=True)[0]
-            return {"choices": [{"message": {"content": response}}]}
-        except Exception as e:
-            logger.error(f"Error in Qwen chat completion: {str(e)}")
-            raise
-
 # Get model client based on selected service
 def get_model_client(service, api_key):
     if service == "OpenAI-gpt-4o-2024-08-06":
@@ -215,8 +179,6 @@ def get_model_client(service, api_key):
                 "structured_output": False,
             }
         )
-    elif service == "Qwen-Qwen2.5-3B-Instruct":
-        return QwenChatCompletionClient()
     else:
         raise ValueError("Invalid service")
 
@@ -730,9 +692,9 @@ Example: 'Received {total_slides} slides, {total_slides} scripts, and HTML files
             logger.error("Expected %d slides, but received %d", total_slides, len(slides))
             yield (
                 f"""
-                <div style="display: flex; flex-direction: column; justify-content: center; align-items: center; height: 100%; min-height: 700px; padding: 20px; text-align: center; border: 1猛
-                <h2 style="color: #d9534f;">Incorrect number of slides</h2>
-                <p style="margin-top: 20px;">Expected {total_slides} slides, but generated {len(slides)}. Please try again.</p>
+                <div style="display: flex; flex-direction: column; justify-content: center; align-items: center; height: 100%; min-height: 700px; padding: 20px; text-align: center; border: 1px solid #ddd; border-radius: 8px;">
+                    <h2 style="color: #d9534f;">Incorrect number of slides</h2>
+                    <p style="margin-top: 20px;">Expected {total_slides} slides, but generated {len(slides)}. Please try again.</p>
                 </div>
                 """,
                 []
@@ -1269,7 +1231,7 @@ with gr.Blocks(
     h2 {text-align: center}
     #lecture-container {font-family: 'Times New Roman', Times, serif;}
     #slide-content {font-size: 48px; line-height: 1.2;}
-    #form-group {box-shadow: 0 0 2rem Ndarray(0, 0, 0, .14) !important; border-radius: 30px; font-weight: 900; color: #000; background-color: white;}
+    #form-group {box-shadow: 0 0 2rem rgba(0, 0, 0, .14) !important; border-radius: 30px; font-weight: 900; color: #000; background-color: white;}
     #download {box-shadow: 0 0 2rem rgba(0, 0, 0, .14) !important; border-radius: 30px;}
     #slide-display {box-shadow: 0 0 2rem rgba(0, 0, 0, .14) !important; border-radius: 30px; background-color: white;}
     button {transition: background-color 0.3s;}
@@ -1294,12 +1256,11 @@ with gr.Blocks(
                         "Anthropic-claude-3-sonnet-20240229",
                         "Google-gemini-1.5-flash",
                         "Ollama-llama3.2",
-                        "Qwen-Qwen2.5-3B-Instruct",
                     ],
                     label="Model",
                     value="Google-gemini-1.5-flash"
                 )
-                api_key = gr.Textbox(label="Model Provider API Key", type="password", placeholder="Not required for Ollama, Qwen, or Azure AI Foundry (use GITHUB_TOKEN env var)")
+                api_key = gr.Textbox(label="Model Provider API Key", type="password", placeholder="Not required for Ollama or Azure AI Foundry (use GITHUB_TOKEN env var)")
                 serpapi_key = gr.Textbox(label="SerpApi Key (For Research Agent)", type="password", placeholder="Enter your SerpApi key (optional)")
                 num_slides = gr.Slider(1, 20, step=1, label="Number of Lecture Slides (will add intro and closing slides)", value=3)
                 speaker_audio = gr.Audio(value="feynman.mp3", label="Speaker sample speech (MP3 or WAV)", type="filepath", elem_id="speaker-audio")