---
tags:
  - coreml
  - apple
  - watchos
  - tinybrain
library_name: coreml
---

# TinyGPT2 for TinyBrain

This repository packages `sshleifer/tiny-gpt2` for the TinyBrain runtime on Apple Watch.

## TinyBrain bundle contract

The conversion script produces a single bundle archive:

- `tinygpt2_only_logits.mlmodelc.zip`

The zip layout is:

- `tinygpt2_only_logits.mlmodelc/`
- `tokenizer/`
- `metadata.json`

## Runtime interface

- Task: causal language modeling
- Inputs:
  - `input_ids` (`Int32`, shape `1 x 16`)
  - `attention_mask` (`Int32`, shape `1 x 16`)
- Output:
  - `logits`

## Files included in `tokenizer/`

- `tokenizer.json` when provided by Transformers
- `vocab.json`
- `merges.txt`
- `special_tokens_map.json`
- `tokenizer_config.json`
- `token_decoder.json`

## Build

Run:

```bash
python3 convert_tinygpt2_to_coreml.py
```

The script:

1. Downloads the model and tokenizer from Hugging Face.
2. Converts the model to Core ML.
3. Compiles the package to `.mlmodelc`.
4. Writes TinyBrain runtime metadata.
5. Packages the compiled model, tokenizer, and metadata into one zip.