| ---
|
| license: apache-2.0
|
| library_name: pytorch
|
| pipeline_tag: image-classification
|
| tags:
|
| - executorch
|
| - zephyr
|
| - tinyml
|
| - mnist
|
| - cortex-m
|
| - esp32
|
| - int8
|
| datasets:
|
| - ylecun/mnist
|
| metrics:
|
| - accuracy
|
| model-index:
|
| - name: TinyDigitNet
|
| results:
|
| - task:
|
| type: image-classification
|
| name: Handwritten digit recognition
|
| dataset:
|
| type: ylecun/mnist
|
| name: MNIST test set
|
| metrics:
|
| - type: accuracy
|
| value: 0.9746
|
| name: Top-1 accuracy (fp32, 10,000 test images)
|
| ---
|
|
|
| # TinyDigitNet
|
|
|
| A 14,410-parameter CNN for 28x28 grayscale digits, sized to run under
|
| [ExecuTorch](https://github.com/pytorch/executorch) on Zephyr RTOS boards with no
|
| delegate: it has been run on an RP2040 (Cortex-M0+, Arducam Pico4ML-BLE, reading digits
|
| from the on-board camera), an ESP32-S3 (Seeed XIAO ESP32S3 Sense), an ESP32 (M5Stack ATOM)
|
| and Cortex-M33 parts (Arduino UNO Q, nRF5340, MCX N947).
|
|
|
| ```
|
| conv1 1->8 3x3, ReLU, maxpool 2 28x28 -> 13x13
|
| conv2 8->16 3x3, ReLU, maxpool 2 13x13 -> 5x5
|
| fc1 400->32, ReLU
|
| fc2 32->10
|
| ```
|
|
|
| Input: `(1, 1, 28, 28)` float, pixel/255 normalised with mean 0.1307 and std 0.3081
|
| (the usual MNIST constants; the firmware applies the same transform to camera crops).
|
| Output: 10 logits.
|
|
|
| ## Files
|
|
|
| | File | What |
|
| |---|---|
|
| | `tiny_digit_net.py` | model definition (PyTorch) |
|
| | `tiny_digit_net.pt` | trained state dict, fp32 |
|
| | `executorch/fp32.pte` | ExecuTorch program, portable fp32 kernels (62.8 KB) |
|
| | `executorch/int8-cmsis-m0plus.pte`, `-m4.pte`, `-m33.pte`, `-m7.pte` | int8 via the ExecuTorch Cortex-M quantizer and `CortexMPassManager`, one per core (25.4 KB) |
|
| | `tflite/fp32.tflite`, `tflite/int8.tflite` | the same weights through a Keras twin, for TensorFlow Lite Micro comparisons (post-training int8 quantization) |
|
| | `test_vectors.h` | one test image plus reference logits for an on-device self-test |
|
|
|
| ## Results
|
|
|
| | Metric | Value |
|
| |---|---|
|
| | MNIST test accuracy (fp32) | 97.46 % |
|
| | Parameters | 14,410 |
|
| | Inference, RP2040 @ 133 MHz, fp32 portable kernels | 766 ms |
|
|
|
| ## Training
|
|
|
| Standard MNIST, Adam, a few epochs on CPU; the script is `train/train_mnist.py` in the
|
| [Pico4ML-BLE repo](https://github.com/eoinjordan/executorch-zephyr-arducam-pico4ML-BLE).
|
| Export with `torch.export` then `to_edge_transform_and_lower(...).to_executorch()`; the
|
| int8 variants use `CortexMQuantizer` + `prepare_pt2e`/`convert_pt2e` and the Cortex-M pass
|
| manager from ExecuTorch 1.5.
|
|
|
| ## Where it runs
|
|
|
| - [executorch-zephyr-arducam-pico4ML-BLE](https://github.com/eoinjordan/executorch-zephyr-arducam-pico4ML-BLE): camera to screen digit reader
|
| - [executorch-zephyr-xiao-esp32s3](https://github.com/eoinjordan/executorch-zephyr-xiao-esp32s3): same on the XIAO ESP32-S3 Sense
|
| - [executorch-vs-tflm-benchmarks](https://github.com/eoinjordan/executorch-vs-tflm-benchmarks): ExecuTorch vs TFLM, fp32 and int8, across boards
|
|
|
| ## License
|
|
|
| Apache-2.0. MNIST is the public dataset by LeCun, Cortes and Burges.
|
|
|
