Ashx098
/

Mini-LLM

@@ -22,9 +22,8 @@ inference:
 # 🧠 Mini-LLM — 80M Parameter Transformer (Pretrained From Scratch)
-<p align="center">
-  <img src="https://huggingface.co/datasets/huggingface/badges/resolve/main/badges-thumbnail.png" width="120"/>
-</p>
 **Mini-LLM** is an 80M parameter decoder-only transformer trained **fully from scratch** using a custom tokenizer, custom architecture, and custom training loop.
 It is designed as an educational + research-friendly minimal LLM that demonstrates how modern LLM components are built end-to-end.

 # 🧠 Mini-LLM — 80M Parameter Transformer (Pretrained From Scratch)
+[![MIT License](https://img.shields.io/badge/license-MIT-green.svg)]()
+[![Model Size](https://img.shields.io/badge/params-80M-blue.svg)]()
 **Mini-LLM** is an 80M parameter decoder-only transformer trained **fully from scratch** using a custom tokenizer, custom architecture, and custom training loop.
 It is designed as an educational + research-friendly minimal LLM that demonstrates how modern LLM components are built end-to-end.