Converting model to DistilGPT2

app.py

@@ -1,28 +1,32 @@
 import gradio as gr
-from transformers import T5Tokenizer, T5ForConditionalGeneration
+from transformers import GPT2Tokenizer, GPT2LMHeadModel
+import torch
 from langchain.memory import ConversationBufferMemory
-from langchain.prompts import PromptTemplate
 
-# Load the tokenizer and model for flan-t5
-tokenizer = T5Tokenizer.from_pretrained("google/flan-t5-base")
-model = T5ForConditionalGeneration.from_pretrained("google/flan-t5-base")
+# Load the tokenizer and model for DistilGPT-2
+tokenizer = GPT2Tokenizer.from_pretrained("distilgpt2")
+model = GPT2LMHeadModel.from_pretrained("distilgpt2")
+
+# Move model to device (GPU if available)
+device = torch.device("cuda") if torch.cuda.is_available() else torch.device("cpu")
+model.to(device)
 
 # Set up conversational memory using LangChain's ConversationBufferMemory
 memory = ConversationBufferMemory()
 
 # Define the chatbot function with memory
-def chat_with_flan(input_text):
+def chat_with_distilgpt2(input_text):
     # Retrieve conversation history and append the current user input
     conversation_history = memory.load_memory_variables({})['history']
     
     # Combine the history with the current user input
     full_input = f"{conversation_history}\nUser: {input_text}\nAssistant:"
     
-    # Tokenize the input for the model
-    input_ids = tokenizer.encode(full_input, return_tensors="pt")
+    # Tokenize the input and convert to tensor
+    input_ids = tokenizer.encode(full_input, return_tensors="pt").to(device)
     
-    # Generate the response from the model
-    outputs = model.generate(input_ids, max_length=200, num_return_sequences=1)
+    # Generate the response using the model
+    outputs = model.generate(input_ids, max_length=200, num_return_sequences=1, pad_token_id=tokenizer.eos_token_id)
     
     # Decode the model output
     response = tokenizer.decode(outputs[0], skip_special_tokens=True)
@@ -34,13 +38,14 @@ def chat_with_flan(input_text):
 
 # Set up the Gradio interface
 interface = gr.Interface(
-    fn=chat_with_flan,
-    inputs=gr.Textbox(label="Chat with FLAN-T5"),
-    outputs=gr.Textbox(label="FLAN-T5's Response"),
-    title="FLAN-T5 Chatbot with Memory",
-    description="This is a simple chatbot powered by the FLAN-T5 model with conversational memory, using LangChain.",
+    fn=chat_with_distilgpt2,
+    inputs=gr.Textbox(label="Chat with DistilGPT-2"),
+    outputs=gr.Textbox(label="DistilGPT-2's Response"),
+    title="DistilGPT-2 Chatbot with Memory",
+    description="This is a simple chatbot powered by the DistilGPT-2 model with conversational memory, using LangChain.",
 )
 
 # Launch the Gradio app
 interface.launch()
 
+