Create app.py

app.py

@@ -0,0 +1,55 @@
+import gradio as gr
+from transformers import AutoModelForCausalLM, AutoTokenizer
+import torch
+
+# Define a dictionary of model names and their corresponding Hugging Face model IDs
+models = {
+    "GPT-Neo-125M": "EleutherAI/gpt-neo-125M",
+    "GPT-J-6B": "EleutherAI/gpt-j-6B",
+    "GPT-NeoX-20B": "EleutherAI/gpt-neox-20b",
+    "GPT-3.5-Turbo": "gpt2",  # Placeholder for illustrative purposes
+}
+
+# Initialize tokenizers and models
+tokenizers = {}
+models_loaded = {}
+
+for model_name, model_id in models.items():
+    tokenizers[model_name] = AutoTokenizer.from_pretrained(model_id)
+    models_loaded[model_name] = AutoModelForCausalLM.from_pretrained(model_id)
+
+def chat(model_name, user_input, history=[]):
+    tokenizer = tokenizers[model_name]
+    model = models_loaded[model_name]
+    
+    # Encode the input
+    input_ids = tokenizer.encode(user_input + tokenizer.eos_token, return_tensors="pt")
+    
+    # Generate a response
+    with torch.no_grad():
+        output = model.generate(input_ids, max_length=150, pad_token_id=tokenizer.eos_token_id)
+    
+    response = tokenizer.decode(output[0], skip_special_tokens=True)
+    
+    # Clean up the response to remove the user input part
+    response = response[len(user_input):].strip()
+    
+    # Append to chat history
+    history.append((user_input, response))
+    
+    return history, history
+
+# Define the Gradio interface
+with gr.Blocks() as demo:
+    gr.Markdown("## Chat with Different Models")
+    
+    model_choice = gr.Dropdown(list(models.keys()), label="Choose a Model")
+    chatbot = gr.Chatbot(label="Chat")
+    message = gr.Textbox(label="Message")
+    submit = gr.Button("Submit")
+    
+    submit.click(chat, inputs=[model_choice, message, chatbot], outputs=[chatbot, chatbot])
+    message.submit(chat, inputs=[model_choice, message, chatbot], outputs=[chatbot, chatbot])
+
+# Launch the demo
+demo.launch()