Update README.md

README.md

@@ -5,35 +5,119 @@ language:
 pipeline_tag: text-generation
 tags:
 - text-generation-inference
+- reasoning
+- grpo
+- qwen3
+- atomight
+- atomightv2-5
 ---
 
-## Atomight V2.5 1.7B
+<div align="center">
 
-<p align="center">
-  <img src="NewAtomight.png" alt="Atomight v2 Logo" width="500" style="max-width: 100%;">
-</p>
+<img src="NewAtomight.png" alt="Atomight v2 Logo" width="500" style="max-width: 100%;">
 
-Most **powerful and capable model** in the current and *newest* Atomight family variant. 
+# Atomight V2.5 · 1.7B
+
+**Reasoning-first · Zero benchmark contamination · Trained on a free Colab T4**
+
+
+
+![License](https://img.shields.io/badge/license-CC--BY--4.0-blue)
+
+
+
+
+![Base](https://img.shields.io/badge/base-Qwen3--1.7B-orange)
+
+
+
+
+![Method](https://img.shields.io/badge/method-GRPO-purple)
+
+
+
+
+![Params](https://img.shields.io/badge/params-1.7B-green)
+
+
+
+</div>
+
+---
+
+Most **powerful and capable model** in the current and *newest* Atomight family variant.
 
 > [!Note]
 > Use this model when you want explicit chain-of-thought before the final answer — complex debugging, multi-step planning, agentic workflows, and math- or reasoning-heavy tasks.
 
-Other details for this model soon. Wait for further information and details. 
+Other details for this model soon. Wait for further information and details.
+
+---
+
+## Quick Facts
+
+| | |
+|---|---|
+| **Base model** | [Qwen/Qwen3-1.7B](https://huggingface.co/Qwen/Qwen3-1.7B) (Apache 2.0) |
+| **Training method** | GRPO via LoRA/PEFT, merged to 16-bit |
+| **Trained on** | Free-tier Google Colab T4 — no paid compute |
+| **Training data** | ~2,000 curated samples per domain, 6 premium open datasets |
+| **Contamination** | Zero — no benchmark data used in training |
+| **License** | CC-BY-4.0 |
+
+---
+
+## Training Data
+
+Curated (not scraped) from frontier 2025-era open datasets across STEM, science, math, and code:
+
+| Domain | Dataset | License |
+|---|---|---|
+| STEM | [Logics-STEM-SFT-Dataset-Open-5.3M](https://huggingface.co/datasets/Logics-MLLM/Logics-STEM-SFT-Dataset-Open-5.3M) | Mixed (aggregated open sources) |
+| Science | [MegaScience](https://huggingface.co/datasets/MegaScience/MegaScience) | CC-BY / Academic Use |
+| Math | [OpenMathReasoning](https://huggingface.co/datasets/nvidia/OpenMathReasoning) | CC-BY-4.0 |
+| Math | [DeepMath-103K](https://huggingface.co/datasets/zwhe99/DeepMath-103K) | CC-BY-4.0 |
+| Code | [OpenCodeInstruct](https://huggingface.co/datasets/nvidia/OpenCodeInstruct) | CC-BY-4.0 |
+| Code | [Nemotron-SFT-Competitive-Programming-v2](https://huggingface.co/datasets/nvidia/Nemotron-SFT-Competitive-Programming-v2) | NVIDIA Open Model License |
+
+> [!Important]
+> **No benchmark data was used in training.** None of the sources above overlap with MMLU, GSM8K, HumanEval, MBPP, HellaSwag, WinoGrande, ARC-Challenge, or TruthfulQA. Every score below was earned, not leaked.
+
+---
 
 ## Self-reported benchmark results
 
-These results are self-reported for `NovatasticRoScript/Atomight-V2.5-1.7B`.
+These results are self-reported for `NovatasticRoScript/Atomight-V2.5-1.7B`, evaluated with `lm-evaluation-harness`.
 
 | Benchmark | Metric | Score |
 |---|---:|---:|
+| MMLU | Accuracy | **55.68** |
+| GSM8K | Accuracy | **69.60** |
 | ARC-Challenge | Accuracy (normalized) | 43.00 |
 | HellaSwag | Accuracy (normalized) | 60.43 |
 | WinoGrande | Accuracy | 61.09 |
 | TruthfulQA MC2 | Accuracy | 45.89 |
 | HumanEval | Pass@1 | 40.24 |
 | MBPP | Pass@1 | 42.80 |
-| MMLU | Accuracy | 55.68 |
 
-### Notes
-- GSM8K is tracked separately in Hugging Face evaluation results.
-- The other scores are listed here as self-reported model card results.
+<sub>All scores self-reported; no training data overlaps with these benchmarks — see Training Data section above.</sub>
+
+---
+
+## Quick Start
+
+```python
+from transformers import AutoModelForCausalLM, AutoTokenizer
+
+model_id = "NovatasticRoScript/Atomight-V2.5-1.7B"
+tokenizer = AutoTokenizer.from_pretrained(model_id)
+model = AutoModelForCausalLM.from_pretrained(model_id, device_map="auto")
+
+messages = [
+    {"role": "system", "content": "You are a reasoning model. Think step-by-step inside <thinking> tags, then give your final answer inside <answer> tags."},
+    {"role": "user", "content": "If a train travels 60 miles in 45 minutes, what is its speed in mph?"}
+]
+
+inputs = tokenizer.apply_chat_template(messages, add_generation_prompt=True, return_tensors="pt").to(model.device)
+outputs = model.generate(inputs, max_new_tokens=512)
+print(tokenizer.decode(outputs[0], skip_special_tokens=True))