| schema_version: 1.0.0 |
| report_type: vision-classification |
| task_type: image-classification |
| title: DeiT-Tiny INT8 |
| description: >- |
| INT8-quantized compact vision transformer for 1000-class image classification, designed for |
| efficient inference on ARM edge devices with ExecuTorch and XNNPACK. It uses static symmetric |
| per-channel quantization and ImageNet-style preprocessing, producing raw logits for top-k |
| classification while reducing model size, latency, and memory use with minimal accuracy impact |
| compared with the FP32 baseline. |
| id: Arm/deit-tiny-int8-xnnpack-executorch |
| filename: deit_raspberry_executorch_optimized.pte |
| base_model_id: facebook/deit-tiny-patch16-224 |
| vendor: Meta |
| base_model_url: https://huggingface.co/facebook/deit-tiny-patch16-224 |
| weight_dtype: int8 |
| profile: Arm-Optimized |
| optimization: |
| optimized_by: Arm |
| quantization: |
| method: PTQ-static |
| weight_bits: 8 |
| activation_bits: 8 |
| symmetric: true |
| mode: static |
| weight_granularity: per-channel |
| calibration: |
| dataset_name: ImageNet-1k |
| sample_count: 1000 |
| selection: random |
| model_size_mb: 6.62 |
| parameter_count: 5717416 |
| format: pte |
|
|
