Llama-3.2-3B-Instruct-ET

ExecuTorch .pte model converted from meta-llama/Llama-3.2-3B-Instruct for on-device inference with ToMogo.

Model Details

Property	Value
Source	meta-llama/Llama-3.2-3B-Instruct
Format	executorch
Quantization	int4 weight-only
Backend	xnnpack
Fusion	disabled
Total Size	4588.1 MB

Files

File	Size
README.md	0.0 MB
chat_template.jinja	0.0 MB
export_config.json	0.0 MB
model.pte	4571.6 MB
tokenizer.json	16.4 MB
tokenizer_config.json	0.0 MB

Usage

Download all files into a single directory and load with ExecuTorch on Android:

val engine = ExecuTorchEngine(modelDir = "/path/to/Llama-3.2-3B-Instruct-ET")
engine.prefill(prompt) { token -> print(token) }

---

Auto-generated by ToMogo upload pipeline.