Update README.md

README.md

@@ -1,22 +1,77 @@
----
-base_model: unsloth/qwen2.5-coder-14b-instruct-bnb-4bit
-tags:
-- text-generation-inference
-- transformers
-- unsloth
-- qwen2
-- trl
-license: apache-2.0
-language:
-- en
----
+[🤗 HF Repo](https://huggingface.co/imsanjoykb/sqlCoder-Qwen2.5-8bit) | [♾️ Colab](https://colab.research.google.com/drive/19e-u32GY2y5lsckNuWhBQExvXgVn8ZjG?usp=sharing)
 
-# Uploaded  model
+Introducing the latest fine-tuned version of Qwen2.5-Coder-14B-Instruct, specifically tailored for SQL code generation. Built on the robust 14-billion parameter Qwen2.5-Coder architecture, this model leverages advanced configurations like bfloat16 precision and a custom quantization setup, optimized for efficient 4-bit computation. With a maximum context window of 32K tokens, this model supports extensive SQL sequences and complex query generation without compromising accuracy or performance.
 
-- **Developed by:** imsanjoykb
-- **License:** apache-2.0
-- **Finetuned from model :** unsloth/qwen2.5-coder-14b-instruct-bnb-4bit
+Our fine-tuning process has enriched this model with domain-specific SQL patterns and nuanced query constructions, making it exceptionally adept at handling real-world SQL requirements, from query creation to debugging and optimization. By combining Qwen2.5's foundational strengths with targeted training on custom SQL data, this model achieves a powerful balance of general-purpose code understanding and SQL-specific precision, making it an ideal tool for developers and data engineers seeking top-tier SQL generation capabilities.
+
+
+## Inference
+
+Here provides a code snippet with `apply_chat_template` to show you how to load the tokenizer and model and how to generate contents.
+
+```python
+# Import necessary libraries
+from unsloth import FastLanguageModel
+import torch
+
+# Define the model name and other parameters
+model_name = "imsanjoykb/sqlCoder-Qwen2.5-8bit"
+max_seq_length = 2048
+dtype = None
+load_in_4bit = True
+
+# Load the model and tokenizer from Hugging Face
+model, tokenizer = FastLanguageModel.from_pretrained(
+    model_name=model_name,
+    max_seq_length=max_seq_length,
+    dtype=dtype,
+    load_in_4bit=load_in_4bit,
+)
+
+# Enable faster inference
+FastLanguageModel.for_inference(model)
+
+# Define the prompt template
+odoo_text2sql_prompt = """Below is an instruction describing a task related to generating a SQL query specifically for Odoo's database structure. The input provides relevant context about Odoo models or data fields from {db_schema}. Write a SQL query that fulfills the given task using Odoo's database schema.
 
-This qwen2 model was trained 2x faster with [Unsloth](https://github.com/unslothai/unsloth) and Huggingface's TRL library.
+### Instruction:
+Generate a SQL query in the context of Odoo to {}
 
-[<img src="https://raw.githubusercontent.com/unslothai/unsloth/main/images/unsloth%20made%20with%20love.png" width="200"/>](https://github.com/unslothai/unsloth)
+### Input:
+{}
+
+### Response:
+{}
+"""
+
+# Optionally, use a TextStreamer for continuous inference
+from transformers import TextStreamer
+
+# Prepare the input text for continuous inference
+instruction = ""
+input_text = "What is the top profitable product?"
+output_text = ""
+
+# Tokenize the input text
+inputs = tokenizer(
+    [
+        odoo_text2sql_prompt.format(instruction, input_text, output_text)
+    ],
+    return_tensors="pt"
+).to("cuda")
+
+# Initialize the TextStreamer
+text_streamer = TextStreamer(tokenizer)
+
+# Generate the output using the model with TextStreamer
+_ = model.generate(**inputs, streamer=text_streamer, max_new_tokens=350)
+```
+## Model Download
+|            **Model**            | **#Total Params** | **#Active Params** | **Context Length** |                         **Download**                         |
+| :-----------------------------: | :---------------: | :----------------: | :----------------: | :----------------------------------------------------------: |
+|   sqlCoder-Qwen2.5-8bit         |        14B        |        2.4B        |        128k        | [🤗 HuggingFace](https://huggingface.co/imsanjoykb/sqlCoder-Qwen2.5-8bit) |
+
+# Uploaded  model
+
+- **Developed by:** [Sanjoy Biswas](https://www.linkedin.com/in/imsanjoykb/) 
+- **License:** apache-2.0