Update README.md

README.md

@@ -3,7 +3,6 @@ language:
   - en
 license: apache-2.0
 library_name: peft
-pipeline_tag: text-generation
 tags:
   - text-generation
   - sql
@@ -11,61 +10,80 @@ tags:
   - peft
   - lora
   - transformers
-base_model: "unsloth/Phi-3-mini-4k-instruct-bnb-4bit"
-datasets:
-  - your-dataset-name
+  - phi-3
+  - instruction-tuning
+base_model: unsloth/Phi-3-mini-4k-instruct-bnb-4bit
+pipeline_tag: text-generation
+inference: false
 ---
 
-# 🧠 SQL Chat Finetuned Model
+# 🧠 SQL Chat – Phi-3-mini SQL Assistant
 
 **Model ID:** `saadkhi/SQL_Chat_finetuned_model`  
-**Task:** Conversational SQL query generation and response  
-**Base Model:** `unsloth/Phi-3-mini-4k-instruct-bnb-4bit`
+**Base model:** `unsloth/Phi-3-mini-4k-instruct-bnb-4bit`  
+**Model type:** LoRA (merged)  
+**Task:** Natural Language → SQL query generation + conversational SQL assistance  
+**Language:** English  
+**License:** Apache 2.0
 
-This model is fine-tuned to assist with generating SQL queries and provide conversational responses related to SQL tasks.
+This model is a fine-tuned version of **Phi-3-mini-4k-instruct** (4-bit quantized) specialized in understanding natural language questions about databases and generating correct, clean SQL queries.
 
----
+## ✨ Key Features
 
-## 🔍 Model Description
+- Very good balance between size, speed and SQL generation quality  
+- Works well with common database dialects (PostgreSQL, MySQL, SQLite, SQL Server, etc.)  
+- Can explain queries, suggest improvements and handle follow-up questions  
+- Fast inference even on consumer hardware (especially with 4-bit quantization)
 
-This model was fine-tuned using PEFT/LoRA techniques on SQL question–answering and instruction data. It’s designed to understand natural language questions about databases and generate corresponding SQL queries or explanations.
+## 🎯 Intended Use & Capabilities
 
-**Intended uses:**
-- Assist developers in writing SQL queries from natural language descriptions.
-- Educational SQL support and explanations.
+**Best for:**
+- Converting natural language questions → SQL queries  
+- Helping beginners learn SQL through explanations  
+- Quick prototyping of SQL queries in development  
+- Building SQL chat interfaces / tools / assistants  
+- Educational purposes
 
-**Out-of-scope uses:**
-- Complex query optimization advice
-- Handling proprietary or highly specialized SQL dialects
+**Limitations / Not recommended for:**
+- Extremely complex analytical/business intelligence queries  
+- Real-time query optimization advice  
+- Very database-specific or proprietary SQL extensions  
+- Production systems without human review (always validate generated SQL!)
 
----
-
-## 🚀 How to Use
-
-### Python (Transformers + PEFT)
+## 🛠️ Quick Start (merged LoRA version)
 
 ```python
-from transformers import AutoModelForCausalLM, AutoTokenizer
+from transformers import AutoTokenizer, AutoModelForCausalLM
 import torch
 
 model_id = "saadkhi/SQL_Chat_finetuned_model"
 
 tokenizer = AutoTokenizer.from_pretrained(model_id)
-model = AutoModelForCausalLM.from_pretrained(model_id, torch_dtype=torch.float16)
-
-prompt = "Show me the SQL to select all users older than 30 from the users table."
-inputs = tokenizer(prompt, return_tensors="pt")
-
-outputs = model.generate(**inputs, max_new_tokens=150)
-print(tokenizer.decode(outputs[0], skip_special_tokens=True))
-
----
-
-### 🧾 Notes
-
-✔️ The YAML metadata at the top (between `---`) helps the Hub display model tags, license, task type, languages, and base model. :contentReference[oaicite:1]{index=1}  
-✔️ You can fill in any missing **More Information Needed** parts from your own fine-tuning logs, dataset info, evaluation results, etc.  
-✔️ Once updated, save this as `README.md` in your repo root. The Hugging Face Hub will render it as your **model card**.
-
-Would you like me to **fill this with real data** from your training logs/parameters if you share them? (e.g., datasets used, evaluation results, hyperparameters)
-::contentReference[oaicite:2]{index=2}
+model = AutoModelForCausalLM.from_pretrained(
+    model_id,
+    torch_dtype=torch.bfloat16,
+    device_map="auto",
+    trust_remote_code=True
+)
+
+# Simple prompt style (chat template is recommended)
+prompt = """Show all customers who placed more than 5 orders in 2024"""
+
+messages = [{"role": "user", "content": prompt}]
+
+inputs = tokenizer.apply_chat_template(
+    messages,
+    tokenize=True,
+    add_generation_prompt=True,
+    return_tensors="pt"
+).to(model.device)
+
+outputs = model.generate(
+    inputs,
+    max_new_tokens=180,
+    do_sample=False,
+    temperature=0.0,
+    pad_token_id=tokenizer.eos_token_id
+)
+
+print(tokenizer.decode(outputs[0], skip_special_tokens=True))