ClearMind-Small / README.md
Perlous's picture
smoke test: ClearMind-Small (13.23M dense, pretrain+sft+dpo, minimind tokenizer/data)
8ebc625 verified
|
Raw History Blame Contribute Delete
183 Bytes

ClearMind-Small

Trained from scratch with ClearMind. Architecture: GPT (RoPE + RMSNorm + SwiGLU + GQA + QK-Norm).

  • Parameters: 15.7M
  • Hidden size: 384
  • Layers: 6
  • Vocab: 6400