Update README.md

README.md

@@ -14,93 +14,3 @@ datasets:
   - custom-legal-dataset-pakistan
 inference: true
 ---
-
-# 🧠 LLaMA 3.2 3B Instruct – Legal QA (Pakistan Law) – Fine-Tuned LoRA
-
-This model is a fine-tuned version of **Unsloth's LLaMA 3.2 3B Instruct** on a custom dataset of Pakistani laws, including:
-
-- 📜 Family Law
-- 🏠 Property Law
-- ⚖️ Criminal Law
-
-It is optimized using **LoRA** (Low-Rank Adaptation) with `unsloth`, making it highly efficient for legal question answering and chatbot use cases within the context of Pakistani law.
-
----
-
-## 📊 Dataset
-
-The model was trained on a custom dataset created by parsing official PDFs of Pakistani legal acts. The structure followed a conversational format like:
-
-\`\`\`text
-### Question:
-What is the procedure of talaq under Pakistani law?
-
-### Answer:
-Under Pakistani law, talaq (divorce) must be initiated by the husband in writing and sent to the relevant Union Council...
-\`\`\`
-
----
-
-## 🛠️ How to Use
-
-> Requires `unsloth` + `transformers`.
-
-\`\`\`python
-from unsloth import FastLanguageModel
-
-model, tokenizer = FastLanguageModel.from_pretrained(
-    model_name = "heyIamUmair/llama3-3b-instruct-legal-pakistan",
-    max_seq_length = 2048,
-    dtype = None,
-    load_in_4bit = True,
-)
-
-FastLanguageModel.for_inference(model)
-
-messages = [{"role": "user", "content": "What are the rights of women after divorce in Pakistani law?"}]
-inputs = tokenizer.apply_chat_template(messages, tokenize=True, add_generation_prompt=True, return_tensors="pt").to("cuda")
-
-outputs = model.generate(**inputs, max_new_tokens=200)
-response = tokenizer.decode(outputs[0], skip_special_tokens=True)
-print(response)
-\`\`\`
-
----
-
-## 🔧 Technical Details
-
-| Setting | Value |
-|--------|-------|
-| Base Model | `unsloth/Llama-3.2-3B-Instruct` |
-| LoRA Rank | 16 |
-| Optimizer | `adamw_8bit` |
-| Quantization | 4-bit |
-| Sequence Length | 2048 |
-| Batch Size | 7 |
-| Epochs | 3 |
-
----
-
-## 📈 Performance Metrics (On Test Set)
-
-| Metric     | Score   |
-|------------|---------|
-| ROUGE-1    | `0.XX`  |
-| BLEU       | `0.XX`  |
-| METEOR     | `0.XX`  |
-
-_(Replace scores with your actual numbers)_
-
----
-
-## 🙋 Author
-
-This model was fine-tuned by **Umair** as part of a legal AI project focused on improving access to Pakistani legal information via LLMs.
-
----
-
-## 📌 License
-
-Apache 2.0 — free to use, modify, and distribute.
-
----