config.json

{
  "architecture": "nanoGPT (custom, trained from scratch)",
  "model_type": "instruction-tuned (SFT)",
  "model_config": {
    "block_size": 256,
    "vocab_size": 50257,
    "n_layer": 12,
    "n_head": 12,
    "n_embd": 768,
    "dropout": 0.0,
    "bias": true
  },
  "total_parameters": 123.849984,
  "tokenizer": "tiktoken gpt2 (50,257 BPE tokens)",
  "framework": "PyTorch",
  "prompt_format": "Alpaca (### Instruction / ### Input / ### Response)"
}