First push

.gitattributes

@@ -0,0 +1,35 @@
+*.7z filter=lfs diff=lfs merge=lfs -text
+*.arrow filter=lfs diff=lfs merge=lfs -text
+*.bin filter=lfs diff=lfs merge=lfs -text
+*.bz2 filter=lfs diff=lfs merge=lfs -text
+*.ckpt filter=lfs diff=lfs merge=lfs -text
+*.ftz filter=lfs diff=lfs merge=lfs -text
+*.gz filter=lfs diff=lfs merge=lfs -text
+*.h5 filter=lfs diff=lfs merge=lfs -text
+*.joblib filter=lfs diff=lfs merge=lfs -text
+*.lfs.* filter=lfs diff=lfs merge=lfs -text
+*.mlmodel filter=lfs diff=lfs merge=lfs -text
+*.model filter=lfs diff=lfs merge=lfs -text
+*.msgpack filter=lfs diff=lfs merge=lfs -text
+*.npy filter=lfs diff=lfs merge=lfs -text
+*.npz filter=lfs diff=lfs merge=lfs -text
+*.onnx filter=lfs diff=lfs merge=lfs -text
+*.ot filter=lfs diff=lfs merge=lfs -text
+*.parquet filter=lfs diff=lfs merge=lfs -text
+*.pb filter=lfs diff=lfs merge=lfs -text
+*.pickle filter=lfs diff=lfs merge=lfs -text
+*.pkl filter=lfs diff=lfs merge=lfs -text
+*.pt filter=lfs diff=lfs merge=lfs -text
+*.pth filter=lfs diff=lfs merge=lfs -text
+*.rar filter=lfs diff=lfs merge=lfs -text
+*.safetensors filter=lfs diff=lfs merge=lfs -text
+saved_model/**/* filter=lfs diff=lfs merge=lfs -text
+*.tar.* filter=lfs diff=lfs merge=lfs -text
+*.tar filter=lfs diff=lfs merge=lfs -text
+*.tflite filter=lfs diff=lfs merge=lfs -text
+*.tgz filter=lfs diff=lfs merge=lfs -text
+*.wasm filter=lfs diff=lfs merge=lfs -text
+*.xz filter=lfs diff=lfs merge=lfs -text
+*.zip filter=lfs diff=lfs merge=lfs -text
+*.zst filter=lfs diff=lfs merge=lfs -text
+*tfevents* filter=lfs diff=lfs merge=lfs -text

README.md

@@ -0,0 +1,13 @@
+---
+title: Deepseek R1 PF
+emoji: 💬
+colorFrom: yellow
+colorTo: purple
+sdk: gradio
+sdk_version: 5.0.1
+app_file: app.py
+pinned: false
+license: apache-2.0
+---
+
+An example chatbot using [Gradio](https://gradio.app), [`huggingface_hub`](https://huggingface.co/docs/huggingface_hub/v0.22.2/en/index), and the [Hugging Face Inference API](https://huggingface.co/docs/api-inference/index).

app.py

@@ -0,0 +1,229 @@
+import os
+import json
+import gradio as gr
+import torch
+from transformers import (
+    TrainingArguments,
+    Trainer,
+    AutoModelForCausalLM,
+    AutoTokenizer,
+    DataCollatorForLanguageModeling
+)
+from datasets import Dataset
+from peft import (
+    prepare_model_for_kbit_training,
+    LoraConfig,
+    get_peft_model
+)
+
+# Constants
+MODEL_NAME = "deepseek-ai/DeepSeek-R1"
+OUTPUT_DIR = "finetuned_models"
+LOGS_DIR = "training_logs"
+
+def save_uploaded_file(file):
+    """Save uploaded file and return its path"""
+    os.makedirs('uploads', exist_ok=True)
+    file_path = os.path.join('uploads', file.name)
+    with open(file_path, 'wb') as f:
+        f.write(file.read())
+    return file_path
+
+def prepare_training_components(
+    data_path,
+    learning_rate,
+    num_epochs,
+    batch_size,
+    model_name=MODEL_NAME
+):
+    """Prepare model, tokenizer, and training arguments"""
+    
+    # Create output directory with timestamp
+    import time
+    timestamp = time.strftime("%Y%m%d_%H%M%S")
+    specific_output_dir = os.path.join(OUTPUT_DIR, f"run_{timestamp}")
+    os.makedirs(specific_output_dir, exist_ok=True)
+    os.makedirs(LOGS_DIR, exist_ok=True)
+
+    # Load tokenizer and model
+    tokenizer = AutoTokenizer.from_pretrained(model_name, trust_remote_code=True)
+    model = AutoModelForCausalLM.from_pretrained(
+        model_name,
+        trust_remote_code=True,
+        torch_dtype=torch.float16,
+        load_in_8bit=True
+    )
+
+    # LoRA Configuration
+    lora_config = LoraConfig(
+        r=16,
+        lora_alpha=32,
+        target_modules=[
+            "q_proj", "k_proj", "v_proj", "o_proj",
+            "gate_proj", "up_proj", "down_proj"
+        ],
+        lora_dropout=0.05,
+        bias="none",
+        task_type="CAUSAL_LM"
+    )
+
+    # Prepare model
+    model = prepare_model_for_kbit_training(model)
+    model = get_peft_model(model, lora_config)
+
+    # Training Arguments
+    training_args = TrainingArguments(
+        output_dir=specific_output_dir,
+        num_train_epochs=num_epochs,
+        per_device_train_batch_size=batch_size,
+        learning_rate=learning_rate,
+        fp16=True,
+        gradient_accumulation_steps=8,
+        gradient_checkpointing=True,
+        logging_dir=os.path.join(LOGS_DIR, f"run_{timestamp}"),
+        logging_steps=10,
+        save_strategy="epoch",
+        evaluation_strategy="epoch",
+        save_total_limit=2,
+    )
+
+    # Load and prepare dataset
+    with open(data_path, 'r') as f:
+        raw_data = json.load(f)
+    
+    # Convert to datasets format
+    dataset = Dataset.from_dict({
+        'text': [item['text'] for item in raw_data]
+    })
+
+    # Create data collator
+    data_collator = DataCollatorForLanguageModeling(
+        tokenizer=tokenizer,
+        mlm=False
+    )
+
+    return {
+        'model': model,
+        'tokenizer': tokenizer,
+        'training_args': training_args,
+        'dataset': dataset,
+        'data_collator': data_collator,
+        'output_dir': specific_output_dir
+    }
+
+def train_model(
+    file,
+    learning_rate=2e-4,
+    num_epochs=3,
+    batch_size=4,
+    progress=gr.Progress()
+):
+    """Training function for Gradio interface"""
+    try:
+        # Save uploaded file
+        file_path = save_uploaded_file(file)
+        
+        # Prepare components
+        progress(0.2, desc="Preparing training components...")
+        components = prepare_training_components(
+            file_path,
+            learning_rate,
+            num_epochs,
+            batch_size
+        )
+        
+        # Initialize trainer
+        progress(0.4, desc="Initializing trainer...")
+        trainer = Trainer(
+            model=components['model'],
+            args=components['training_args'],
+            train_dataset=components['dataset'],
+            data_collator=components['data_collator'],
+        )
+        
+        # Train
+        progress(0.5, desc="Training model...")
+        trainer.train()
+        
+        # Save model and tokenizer
+        progress(0.9, desc="Saving model...")
+        trainer.save_model()
+        components['tokenizer'].save_pretrained(components['output_dir'])
+        
+        progress(1.0, desc="Training complete!")
+        return f"Training completed! Model saved in {components['output_dir']}"
+    
+    except Exception as e:
+        return f"Error during training: {str(e)}"
+
+# Create Gradio interface
+def create_interface():
+    with gr.Blocks() as demo:
+        gr.Markdown("# DeepSeek-R1 Model Finetuning Interface")
+        
+        with gr.Row():
+            with gr.Column():
+                file_input = gr.File(
+                    label="Upload Training Data (JSON)",
+                    type="binary",
+                    file_types=[".json"]
+                )
+                
+                learning_rate = gr.Slider(
+                    minimum=1e-5,
+                    maximum=1e-3,
+                    value=2e-4,
+                    label="Learning Rate"
+                )
+                
+                num_epochs = gr.Slider(
+                    minimum=1,
+                    maximum=10,
+                    value=3,
+                    step=1,
+                    label="Number of Epochs"
+                )
+                
+                batch_size = gr.Slider(
+                    minimum=1,
+                    maximum=8,
+                    value=4,
+                    step=1,
+                    label="Batch Size"
+                )
+                
+                train_button = gr.Button("Start Training")
+                
+            with gr.Column():
+                output = gr.Textbox(label="Training Status")
+                
+        train_button.click(
+            fn=train_model,
+            inputs=[file_input, learning_rate, num_epochs, batch_size],
+            outputs=output
+        )
+        
+        gr.Markdown("""
+        ## Instructions
+        1. Upload your training data in JSON format:
+        ```json
+        [
+            {"text": "User: Question\nAssistant: Answer"},
+            {"text": "User: Another question\nAssistant: Another answer"}
+        ]
+        ```
+        2. Adjust training parameters if needed
+        3. Click 'Start Training'
+        4. Wait for training to complete
+        """)
+    
+    return demo
+
+if __name__ == "__main__":
+    # Create necessary directories
+    os.makedirs(OUTPUT_DIR, exist_ok=True)
+    os.makedirs(LOGS_DIR, exist_ok=True)
+    
+    # Launch Gradio interface
+    demo = create_interface()
+    demo.launch(share=True)

requirements.txt

@@ -0,0 +1 @@
+huggingface_hub==0.25.2