A 4-bit (MLX) quantized build of meta-llama/Llama-3.2-3B-Instruct, repackaged into fixed-size shards for delivery as iOS On-Demand Resources. This is the model behind the free Carryover Compact on-device intelligence engine in the Carryover app.

What this is

Base model: Llama 3.2 3B Instruct (Meta)
Quantization: 4-bit, via MLX
Packaging: weights split into compact-shard-1.bin … compact-shard-5.bin plus tokenizer and config files; the app reassembles and SHA-256-verifies them on device.
Runtime: MLX on Apple silicon. Everything runs locally — no data leaves the device.

Intended use

Powers the free Carryover Compact engine in the Carryover iOS app, turning meeting and voice notes into structured artifacts (recaps, tickets, decision logs, and more), entirely on-device.

License

This is a derivative of Llama 3.2, distributed under the Llama 3.2 Community License.

Llama 3.2 is licensed under the Llama 3.2 Community License, Copyright © Meta Platforms, Inc. All Rights Reserved.

Use is also subject to Meta's Acceptable Use Policy. "Llama" is a trademark of Meta Platforms, Inc.

madjagstudios/Llama-3.2-3B-Carryover-Compact

Llama-3.2-3B-Carryover-Compact

What this is

Intended use

License

Acknowledgements

madjagstudios/
Llama-3.2-3B-Carryover-Compact