minimind-135m-cpt-instruct

minimind-135m-cpt-instruct

Instruction-tuned MiniMind (135M), fine-tuned from minimind-135m-cpt.

Model details

• Architecture: decoder-only Transformer (GPT-style), 12 layers / 768 dim / 12 heads, block size 512
• Parameters: ~135M
• Tokenizer: 32k byte-level BPE (tokenizer.json)
• Instruction-tuned (SFT) on Alpaca (52k) + Dolly (15k), starting from the minimind-135m-cpt base.
• Uses the Alpaca prompt template (see the training notebook).

Files

• minimind_model.pt — native PyTorch checkpoint (dict with model_state_dict + config)
• tokenizer.json / tokenizer_config.json / special_tokens_map.json

Load (native)

import torch
from <your_module> import MiniMind   # class definition lives in the training notebook
ck = torch.load('minimind_model.pt', map_location='cpu')
model = MiniMind(**ck['config'])
model.load_state_dict(ck['model_state_dict'])