app.py

import os
import torch
import gradio as gr
from transformers import AutoTokenizer, AutoModelForCausalLM, TextGenerationPipeline
from safetensors.torch import load_file  # Import safetensors for loading .safetensors models
import datetime

# Model Constants
MODEL_ID = "FlameF0X/Snowflake-G0-Release"  # HF repo when published
MAX_LENGTH = 384
TEMPERATURE_MIN = 0.1
TEMPERATURE_MAX = 2.0
TEMPERATURE_DEFAULT = 0.7
TOP_P_MIN = 0.1
TOP_P_MAX = 1.0
TOP_P_DEFAULT = 0.9
TOP_K_MIN = 1
TOP_K_MAX = 100
TOP_K_DEFAULT = 40
MAX_NEW_TOKENS_MIN = 16
MAX_NEW_TOKENS_MAX = 1024
MAX_NEW_TOKENS_DEFAULT = 256

# CSS for the app
css = """
.gradio-container {
    background-color: #1e1e2f !important;
    color: #e0e0e0 !important;
}
.header {
    background-color: #2b2b3c;
    padding: 20px;
    margin-bottom: 20px;
    border-radius: 10px;
    text-align: center;
}
.header h1 {
    color: #66ccff;
    margin-bottom: 10px;
}
.snowflake-icon {
    font-size: 24px;
    margin-right: 10px;
}
.footer {
    text-align: center;
    margin-top: 20px;
    font-size: 0.9em;
    color: #999;
}
.parameter-section {
    background-color: #2a2a3a;
    padding: 15px;
    border-radius: 8px;
    margin-bottom: 15px;
}
.parameter-section h3 {
    margin-top: 0;
    color: #66ccff;
}
.example-section {
    background-color: #223344;
    padding: 15px;
    border-radius: 8px;
    margin-bottom: 15px;
}
.example-section h3 {
    margin-top: 0;
    color: #66ffaa;
}
"""

# Helper functions to load model
def load_model_and_tokenizer():
    global model, tokenizer, pipeline  # Add this line

    # Load the tokenizer
    tokenizer = AutoTokenizer.from_pretrained(MODEL_ID)
    
    # Check if the pad_token is None, set it to eos_token if needed
    if tokenizer.pad_token is None:
        tokenizer.pad_token = tokenizer.eos_token

    # Check if the model uses safetensors or pytorch .bin model file
    model_file_path = os.path.join(MODEL_ID, "model.safetensors")  # or model.bin if that's the case

    if os.path.exists(model_file_path):
        # Check if safetensors file exists
        print("Loading model from safetensors file...")
        model = load_file(model_file_path)  # Safetensors loading
    else:
        # Load from standard .bin file
        print("Loading model from .bin file...")
        model = AutoModelForCausalLM.from_pretrained(MODEL_ID, 
                                                     torch_dtype=torch.float16 if torch.cuda.is_available() else torch.float32,
                                                     device_map="auto")
    
    # Initialize the generation pipeline
    pipeline = TextGenerationPipeline(
        model=model,
        tokenizer=tokenizer,
        return_full_text=False,
        max_length=MAX_LENGTH
    )
    
    return model, tokenizer, pipeline

# Helper functions for generation
def generate_text(
    prompt, 
    temperature=TEMPERATURE_DEFAULT, 
    top_p=TOP_P_DEFAULT, 
    top_k=TOP_K_DEFAULT, 
    max_new_tokens=MAX_NEW_TOKENS_DEFAULT,
    history=None
):
    if history is None:
        history = []
    
    # Add current prompt to history
    history.append({"role": "user", "content": prompt})
    
    try:
        # Generate response
        outputs = pipeline(
            prompt,
            do_sample=temperature > 0,
            temperature=temperature,
            top_p=top_p,
            top_k=top_k,
            max_new_tokens=max_new_tokens,
            pad_token_id=tokenizer.pad_token_id,
            num_return_sequences=1
        )
        
        response = outputs[0]["generated_text"]
        
        # Add model response to history
        history.append({"role": "assistant", "content": response})
        
        # Format chat history for display
        formatted_history = []
        for entry in history:
            role_prefix = "👤 User: " if entry["role"] == "user" else "❄️ Snowflake: "
            formatted_history.append(f"{role_prefix}{entry['content']}")
        
        return response, history, "\n\n".join(formatted_history)
    
    except Exception as e:
        error_msg = f"Error generating response: {str(e)}"
        history.append({"role": "assistant", "content": f"[ERROR] {error_msg}"})
        return error_msg, history, str(history)

def clear_conversation():
    return "", [], ""

def apply_preset_example(example, history):
    return example, history

# Example prompts
examples = [
    "Write a short story about a snowflake that comes to life.",
    "Explain the concept of artificial neural networks to a 10-year-old.",
    "What are some interesting applications of natural language processing?",
    "Write a haiku about programming.",
    "Create a dialogue between two AI researchers discussing the future of language models."
]

# Main function
def create_demo():
    with gr.Blocks(css=css) as demo:
        # Header
        gr.HTML("""
        <div class="header">
            <h1><span class="snowflake-icon">❄️</span> Snowflake-G0-Release Demo</h1>
            <p>Experience the capabilities of the Snowflake-G0-Release language model</p>
        </div>
        """)
        
        # Model info
        with gr.Accordion("About Snowflake-G0-Release", open=False):
            gr.Markdown("""
            ## Snowflake-G0-Release
            
            This is the initial release of the Snowflake series language models, trained on the DialogMLM-50K dataset with optimized memory usage.
            
            ### Model details
            - Architecture: SnowflakeCore
            - Hidden size: 384
            - Number of attention heads: 6
            - Number of layers: 4
            - Feed-forward dimension: 768
            - Maximum sequence length: 384
            - Vocabulary size: 30522 (BERT tokenizer)
            
            ### Key Features
            - Efficient memory usage
            - Fused QKV projection for faster inference
            - Pre-norm architecture for stable training
            - Compatible with HuggingFace Transformers
            """)
        
        # Chat interface
        with gr.Column():
            chat_history_display = gr.Textbox(
                value="", 
                label="Conversation History", 
                lines=10, 
                max_lines=30,
                interactive=False
            )
            
            # Invisible state variables
            history_state = gr.State([])
            
            # Input and output
            with gr.Row():
                with gr.Column(scale=4):
                    prompt = gr.Textbox(
                        placeholder="Type your message here...", 
                        label="Your Input",
                        lines=2
                    )
                with gr.Column(scale=1):
                    submit_btn = gr.Button("Send", variant="primary")
                    clear_btn = gr.Button("Clear Conversation")
            
            response_output = gr.Textbox(
                value="", 
                label="Model Response", 
                lines=5,
                max_lines=10,
                interactive=False
            )
        
        # Advanced parameters
        with gr.Accordion("Generation Parameters", open=False):
            with gr.Column(elem_classes="parameter-section"):
                with gr.Row():
                    with gr.Column():
                        temperature = gr.Slider(
                            minimum=TEMPERATURE_MIN, 
                            maximum=TEMPERATURE_MAX,
                            value=TEMPERATURE_DEFAULT,
                            step=0.05,
                            label="Temperature",
                            info="Higher = more creative, Lower = more deterministic"
                        )
                        
                        top_p = gr.Slider(
                            minimum=TOP_P_MIN,
                            maximum=TOP_P_MAX,
                            value=TOP_P_DEFAULT,
                            step=0.05,
                            label="Top-p (nucleus sampling)",
                            info="Controls diversity via cumulative probability"
                        )
                    
                    with gr.Column():
                        top_k = gr.Slider(
                            minimum=TOP_K_MIN,
                            maximum=TOP_K_MAX,
                            value=TOP_K_DEFAULT,
                            step=1,
                            label="Top-k",
                            info="Limits word choice to top k options"
                        )
                        
                        max_new_tokens = gr.Slider(
                            minimum=MAX_NEW_TOKENS_MIN,
                            maximum=MAX_NEW_TOKENS_MAX,
                            value=MAX_NEW_TOKENS_DEFAULT,
                            step=8,
                            label="Maximum New Tokens",
                            info="Controls the length of generated response"
                        )
        
        # Examples
        with gr.Accordion("Example Prompts", open=True):
            with gr.Column(elem_classes="example-section"):
                example_btn = gr.Examples(
                    examples=examples,
                    inputs=prompt,
                    label="Click on an example to try it",
                    examples_per_page=5
                )
        
        # Footer
        gr.HTML(f"""
        <div class="footer">
            <p>Snowflake-G0-Release Demo • Created with Gradio • {datetime.datetime.now().year}</p>
        </div>
        """)
        
        # Set up interactions
        submit_btn.click(
            fn=generate_text,
            inputs=[prompt, temperature, top_p, top_k, max_new_tokens, history_state],
            outputs=[response_output, history_state, chat_history_display]
        )
        
        prompt.submit(
            fn=generate_text,
            inputs=[prompt, temperature, top_p, top_k, max_new_tokens, history_state],
            outputs=[response_output, history_state, chat_history_display]
        )
        
        clear_btn.click(
            fn=clear_conversation,
            inputs=[],
            outputs=[prompt, history_state, chat_history_display]
        )
        
    return demo

# Load model and tokenizer
print("Loading Snowflake-G0-Release model and tokenizer...")
try:
    model, tokenizer, pipeline = load_model_and_tokenizer()
    print("Model loaded successfully!")
except Exception as e:
    print(f"Error loading model: {str(e)}")
    # Create a simple error demo if model fails to load
    with gr.Blocks(css=css) as error_demo:
        gr.HTML(f"""
        <div class="header" style="background-color: #ffebee;">
            <h1><span class="snowflake-icon">⚠️</span> Error Loading Model</h1>
            <p>There was a problem loading the Snowflake-G0-Release model: {str(e)}</p>
        </div>
        """)
    demo = error_demo

# Create and launch the demo
demo = create_demo()

# Launch the app
if __name__ == "__main__":
    demo.launch()