Update README.md

README.md

@@ -7,6 +7,8 @@ base_model:
 pipeline_tag: question-answering
 ---
 
+# LLaMA 3.1-8B Fine-Tuned on ChatDoctor Dataset
+
 ## Model Overview
 This model is a fine-tuned version of the LLaMA 3.1-8B model, trained on a curated selection of 1,122 samples from the **ChatDoctor (HealthCareMagic-100k)** dataset. It has been optimized for tasks related to medical consultations.
 
@@ -21,6 +23,12 @@ This model is designed to assist in:
 - Providing health-related advice
 - Assisting in basic diagnostic reasoning (non-clinical use)
 
+## Datasets
+- **Training Data**: ChatDoctor-HealthCareMagic-100k
+  - **Training Set**: 900 samples
+  - **Validation Set**: 100 samples
+  - **Test Set**: 122 samples
+
 ## Model Details
 | **Feature**                  | **Details**                |
 |------------------------------|----------------------------|
@@ -49,7 +57,7 @@ The model was fine-tuned with the following hyperparameters:
 Validation was performed using a separate subset of the dataset. The final training and validation loss are as follows:
 
 <p align="center">
-  <img src="train-val-curve.png" alt="Training and Validation Loss" width="50%"/>
+  <img src="train-val-curve.png" alt="Training and Validation Loss" width="35%"/>
 </p>
 
 ## Evaluation Results
@@ -64,46 +72,28 @@ Validation was performed using a separate subset of the dataset. The final train
 - **ROUGE-L**: 0.1249
 
 ## Usage
-### Loading the Model
-This model is hosted in **GGUF format** for optimal deployment. You can load and run the model using **LLaMA.cpp**.
-
-#### Steps to Use
-1. Clone the LLaMA.cpp repository:
-    ```bash
-    git clone https://github.com/ggerganov/llama.cpp
-    cd llama.cpp
-    make
-    ```
-
-2. Download the model from Hugging Face:
-    ```bash
-    huggingface-cli login
-    wget https://huggingface.co/your-username/llama-3.1-8B-gguf/resolve/main/output_model.gguf
-    ```
-
-3. Run inference:
-    ```bash
-    ./main -m output_model.gguf -p "What are the symptoms of a common cold?" -t 4 -n 100
-    ```
-
-### Quantization Details
-The model is quantized to **Q4_0** for faster inference while maintaining reasonable accuracy. You can run it efficiently on CPUs with low memory requirements.
-
-## Limitations and Intended Use
-- **Not for Clinical Use**: This model is intended for educational purposes and general health advice. It should not replace professional medical consultation.
-- **Bias and Errors**: The model might exhibit biases present in the training data. Outputs should be interpreted with caution.
-
-## Acknowledgments
-- **Dataset**: ChatDoctor (HealthCareMagic-100k)
-- **Base Model**: LLaMA 3.1-8B
-- **Quantization Tools**: LLaMA.cpp
-
-## Citation
-If you use this model, please cite:
-```
-@article{yourcitation,
-  title={Fine-tuned LLaMA 3.1-8B on ChatDoctor Dataset},
-  author={Your Name},
-  year={2025},
-  publisher={Hugging Face}
-}
+```python
+from transformers import AutoTokenizer, AutoModelForCausalLM
+from bitsandbytes import BitsAndBytesConfig
+
+model_id = "your-model-id"
+
+# Load tokenizer
+tokenizer = AutoTokenizer.from_pretrained(model_id)
+tokenizer.pad_token = tokenizer.eos_token
+
+# Configure quantization
+bnb_config = BitsAndBytesConfig(
+    load_in_4bit=True, 
+    bnb_4bit_quant_type="nf4", 
+    bnb_4bit_compute_dtype="float16", 
+    bnb_4bit_use_double_quant=True
+)
+
+# Load model with quantization
+model = AutoModelForCausalLM.from_pretrained(
+    model_id, 
+    quantization_config=bnb_config, 
+    device_map="auto"
+)
+```