chat.py

import torch
from transformers import AutoTokenizer, AutoModelForCausalLM

# Load the tokenizer and model
model_name = "synCAI-144k-gpt2.5"
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForCausalLM.from_pretrained(model_name)

# Check if GPU is available and move model to GPU
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
model.to(device)

def generate_text(prompt, model, tokenizer, device, max_length=100, temperature=0.7, top_p=0.9, top_k=50):
    try:
        # Tokenize the input prompt
        inputs = tokenizer(prompt, return_tensors="pt")
        inputs = {key: value.to(device) for key, value in inputs.items()}

        # Generate text
        outputs = model.generate(
            inputs['input_ids'], 
            max_length=max_length, 
            temperature=temperature, 
            top_p=top_p, 
            top_k=top_k
        )

        # Decode and return the generated text
        generated_text = tokenizer.decode(outputs[0], skip_special_tokens=True)
        return generated_text
    except Exception as e:
        print(f"Error generating text for prompt '{prompt}': {e}")
        return None

# Example input prompts
input_prompts = [
    "Explain the significance of the project:",
    "What methodologies were used in the research?",
    "What are the future implications of the findings?"
]

# Generate and print texts for each prompt
for prompt in input_prompts:
    generated_text = generate_text(prompt, model, tokenizer, device)
    if generated_text:
        print(f"Prompt: {prompt}")
        print(f"Generated Text: {generated_text}\n")