Update app.py

app.py

@@ -1,47 +1,44 @@
 import streamlit as st
 from transformers import AutoTokenizer, AutoModelForCausalLM
 
-tokenizer = AutoTokenizer.from_pretrained("mistralai/Mixtral-8x7B-Instruct-v0.1")
+tokenizer = AutoTokenizer.from_pretrained("gpt2-large")
 @st.cache_data
 def load_model(model_name):
-    model = AutoModelForCausalLM.from_pretrained("mistralai/Mixtral-8x7B-Instruct-v0.1")
-    return model
-
-model = load_model("mistralai/Mixtral-8x7B-Instruct-v0.1")
-
-def infer(input_ids, max_length, temperature, top_k, top_p):
-    input_ids = torch.tensor(input_ids).unsqueeze(0)
-    output_sequences = model.generate(
-        input_ids=input_ids,
-        max_length=max_length,
-        temperature=temperature,
-        top_k=top_k,
-        top_p=top_p,
-        do_sample=True,
-        num_return_sequences=1
-    )
-    return output_sequences
-
-default_value = "Ask me anything!"
+model = AutoModelForCausalLM.from_pretrained(model_name)
+return model
+
+model = load_model("gpt2-large")
+
+def infer(sent, max_length, temperature, top_k, top_p):
+input_ids = tokenizer.encode(sent, return_tensors="pt")
+output_sequences = model.generate(
+input_ids=input_ids,
+max_length=max_length,
+temperature=temperature,
+top_k=top_k,
+top_p=top_p,
+do_sample=True,
+num_return_sequences=1
+)
+return output_sequences
+
+default_value = "You: Ask me anything!"
 
 #prompts
-st.title("Chatbot")
-
-sent = st.text_area("Message", default_value, height = 100)
-max_length = st.sidebar.slider("Max Length", min_value = 10, max_value=30)
-temperature = st.sidebar.slider("Temperature", value = 1.0, min_value = 0.0, max_value=1.0, step=0.05)
-top_k = st.sidebar.slider("Top-k", min_value = 0, max_value=5, value = 0)
-top_p = st.sidebar.slider("Top-p", min_value = 0.0, max_value=1.0, step = 0.05, value = 0.9)
-
-encoded_prompt = tokenizer.encode(sent, add_special_tokens=False, return_tensors="pt")
-if encoded_prompt.size()[-1] == 0:
-    input_ids = None
-else:
-    input_ids = encoded_prompt
-
-output_sequences = infer(input_ids, max_length, temperature, top_k, top_p)
-
-for generated_sequence_idx, generated_sequence in enumerate(output_sequences):
-    generated_sequences = generated_sequence.tolist()
-    text = tokenizer.decode(generated_sequence, clean_up_tokenization_spaces=True)
-    st.write(text)
+st.title("Chat with GPT-2 💬")
+st.write("GPT-2 is a large transformer-based language model with 1.5 billion parameters. It is trained to predict the next word in a sentence, given all of the previous words. This makes it great for text generation and for answering questions about the text it's given.")
+
+messages = [{"role": "system", "content": "You are a helpful assistant."}]
+
+user_input = st.text_input("You:", default_value)
+if user_input:
+messages.append({"role": "user", "content": user_input})
+
+output_sequences = infer(user_input, max_length=100, temperature=0.7, top_k=40, top_p=0.9)
+generated_sequence = output_sequences[0].tolist()
+generated_text = tokenizer.decode(generated_sequence, clean_up_tokenization_spaces=True)
+
+messages.append({"role": "assistant", "content": generated_text})
+
+for message in messages:
+st.write(f"{message['role']}: {message['content']}")