commit | 3e9cd02d744832be48f8ccdc9aee6c5b442ef9f4 | [log] [tgz] |
---|---|---|
author | Lun Dong <lundong@google.com> | Thu Mar 27 12:35:49 2025 -0700 |
committer | Lun Dong <lundong@google.com> | Thu Mar 27 12:35:49 2025 -0700 |
tree | 89023f80ce3fae437d85e7c47066891bbb2b4b63 | |
parent | a3240e0f75385a730484778456db8f00132ea5a1 [diff] |
ml/ml-models-public: changes to accommodate IREE API for the 12052024 revision Context: due to a long-standing bug on IREE importer/compiler (https://github.com/iree-org/iree/issues/19402) and a latest change on IREE compiler/runtime (https://github.com/iree-org/iree/commit/487c5217e4b7a931eb288b35b11c043c7567a7e3), we have to pin IREE to a 12/05/2024 revision to avoid breakage. This change makes changes to accommodate IREE API for this specific revision Change-Id: I5d916ec31c8add20db82c8f92ef71f8dc32ab069
This is the model zoo for public models used in Shodan vector core examples.
The person presence detection quantized model from https://github.com/tensorflow/tflite-micro/blob/main/tensorflow/lite/micro/models/person_detect.tflite
Mobilenet V1 float-point model from https://tfhub.dev/tensorflow/lite-model/mobilenet_v1_0.25_224/1/default/1?lite-format=tflite
Mobilenet V1 quantized model from https://tfhub.dev/tensorflow/lite-model/mobilenet_v1_0.25_224_quantized/1/default/1?lite-format=tflite
Mobilenet V2 quantized model from https://tfhub.dev/tensorflow/lite-model/mobilenet_v2_1.0_224_quantized/1/default/1?lite-format=tflite
HPS (Human Presence Sensor) non-tiled quantized model from https://chromium.googlesource.com/chromiumos/platform/hps-firmware/+/6cdea6d1158a8cd3238b8ae4f744fdb494779c80/models/shared.tflite
For tflite models, use the web-based visualization tool to inspect the file. The tool supports drag and drop or file GUI.
Test data for sample model inputs are in the “test_data” directory. Currently, only test data for the Human Presence Sensor model are included.
Test image files for Human Presence Sensor model, from ChromeOS codebase.
For each model, the corresponding model executable under iree_exec
can be built with the .c/.h files using sw/vec_iree
project as library. They can be run as unit tests via lit framework using Renode or QEMU.