Update README.md

README.md

@@ -22,11 +22,8 @@ Developed by **{ORGANIZATION}**.
 - **Total parameters:** {model.n_params:,}
 
 ## Training data (3-way mix)
-- **60% distilled** (`cofos_distilled.jsonl` — multi-LLM distillation:
-  Anthropic, Google, Z.ai, DeepSeek, Qwen) with FR + EN instructions and
-  optional chain-of-thought
 - **30% real HF Python** (`iamtarun/python_code_instructions_18k_alpaca`)
-- **10% synthetic** (procedural code generators)
+
 
 ## Result
 - **Best `real_syntax_valid`:** {best_syntax:.1f}% on held-out real Python instructions