Update model card: Zero9 Tech copyright, Apache-2.0, dataset links and key test highlights

README.md

@@ -1,21 +1,38 @@
 ---
-base_model: unsloth/Qwen3-4B-unsloth-bnb-4bit
-tags:
-- text-generation-inference
-- transformers
-- unsloth
-- qwen3
-license: apache-2.0
 language:
 - en
+license: apache-2.0
+library_name: transformers
+tags:
+- qwen3
+- english
+- data-mining
+- data-science
+- instruction-tuning
+- sft
+datasets:
+- murataksit34/data-scientist-dialog-8k-en
 ---
 
-# Uploaded finetuned  model
+# Qwen3-4B-DataScience
+
+This model is tuned for decision-oriented data mining and applied data science assistance.
+
+## Training Setup
+1. Domain SFT: `murataksit34/data-scientist-dialog-8k-en`.
+
+## Dataset Test Highlights (murataksit34/data-scientist-dialog-8k-en)
+- Total records: `2,611`
+- Split: `train: 2,098 · test: 255`
+- Purity violations: `train=0, test=0`
+- assistant_first_unique_ratio: `0.8430`
+- assistant_final_unique_ratio: `1.0000`
 
-- **Developed by:** zero9tech
-- **License:** apache-2.0
-- **Finetuned from model :** unsloth/Qwen3-4B-unsloth-bnb-4bit
+## Usage Note
+Model behavior is optimized for decision-focused responses (method choice, alternatives, risk signals, validation planning).
 
-This qwen3 model was trained 2x faster with [Unsloth](https://github.com/unslothai/unsloth) and Huggingface's TRL library.
+## Copyright
+Copyright (c) Zero9 Tech
 
-[<img src="https://raw.githubusercontent.com/unslothai/unsloth/main/images/unsloth%20made%20with%20love.png" width="200"/>](https://github.com/unslothai/unsloth)
+## License
+Apache-2.0