--- library_name: pytorch license: other tags: - bu_auto - android pipeline_tag: image-classification --- ![](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/nasnet/web-assets/model_demo.png) # NASNet: Optimized for Qualcomm Devices NASNet is a vision transformer model that can classify images from the Imagenet dataset. This is based on the implementation of NASNet found [here](https://github.com/huggingface/pytorch-image-models/tree/main). This repository contains pre-exported model files optimized for Qualcomm® devices. You can use the [Qualcomm® AI Hub Models](https://github.com/quic/ai-hub-models/blob/main/qai_hub_models/models/nasnet) library to export with custom configurations. More details on model performance across various devices, can be found [here](#performance-summary). Qualcomm AI Hub Models uses [Qualcomm AI Hub Workbench](https://workbench.aihub.qualcomm.com) to compile, profile, and evaluate this model. [Sign up](https://myaccount.qualcomm.com/signup) to run these models on a hosted Qualcomm® device. ## Getting Started There are two ways to deploy this model on your device: ### Option 1: Download Pre-Exported Models Below are pre-exported model assets ready for deployment. | Runtime | Precision | Chipset | SDK Versions | Download | |---|---|---|---|---| | ONNX | float | Universal | QAIRT 2.42, ONNX Runtime 1.24.1 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/nasnet/releases/v0.47.0/nasnet-onnx-float.zip) | ONNX | w8a8_mixed_fp16 | Universal | QAIRT 2.42, ONNX Runtime 1.24.1 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/nasnet/releases/v0.47.0/nasnet-onnx-w8a8_mixed_fp16.zip) | QNN_DLC | float | Universal | QAIRT 2.43 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/nasnet/releases/v0.47.0/nasnet-qnn_dlc-float.zip) | QNN_DLC | w8a8_mixed_fp16 | Universal | QAIRT 2.43 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/nasnet/releases/v0.47.0/nasnet-qnn_dlc-w8a8_mixed_fp16.zip) | TFLITE | float | Universal | QAIRT 2.43, TFLite 2.17.0 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/nasnet/releases/v0.47.0/nasnet-tflite-float.zip) For more device-specific assets and performance metrics, visit **[NASNet on Qualcomm® AI Hub](https://aihub.qualcomm.com/models/nasnet)**. ### Option 2: Export with Custom Configurations Use the [Qualcomm® AI Hub Models](https://github.com/quic/ai-hub-models/blob/main/qai_hub_models/models/nasnet) Python library to compile and export the model with your own: - Custom weights (e.g., fine-tuned checkpoints) - Custom input shapes - Target device and runtime configurations This option is ideal if you need to customize the model beyond the default configuration provided here. See our repository for [NASNet on GitHub](https://github.com/quic/ai-hub-models/blob/main/qai_hub_models/models/nasnet) for usage instructions. ## Model Details **Model Type:** Model_use_case.image_classification **Model Stats:** - Model checkpoint: nasnetalarge.tf_in1k - Input resolution: 224x224 - GMACs: 5.9 - Activations (M): 19.4 - Number of parameters: 88.7M - Model size (float): 338 MB ## Performance Summary | Model | Runtime | Precision | Chipset | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit |---|---|---|---|---|---|--- | NASNet | ONNX | float | Snapdragon® X Elite | 17.835 ms | 188 - 188 MB | NPU | NASNet | ONNX | float | Snapdragon® 8 Gen 3 Mobile | 12.849 ms | 1 - 839 MB | NPU | NASNet | ONNX | float | Qualcomm® QCS8550 (Proxy) | 17.435 ms | 0 - 196 MB | NPU | NASNet | ONNX | float | Qualcomm® QCS9075 | 28.348 ms | 0 - 4 MB | NPU | NASNet | ONNX | float | Snapdragon® 8 Elite For Galaxy Mobile | 10.583 ms | 1 - 676 MB | NPU | NASNet | ONNX | float | Snapdragon® 8 Elite Gen 5 Mobile | 8.382 ms | 1 - 673 MB | NPU | NASNet | ONNX | float | Snapdragon® X2 Elite | 8.272 ms | 189 - 189 MB | NPU | NASNet | ONNX | w8a8_mixed_fp16 | Snapdragon® X Elite | 11.095 ms | 98 - 98 MB | NPU | NASNet | ONNX | w8a8_mixed_fp16 | Snapdragon® 8 Gen 3 Mobile | 6.945 ms | 6 - 479 MB | NPU | NASNet | ONNX | w8a8_mixed_fp16 | Qualcomm® QCS8550 (Proxy) | 9.919 ms | 5 - 10 MB | NPU | NASNet | ONNX | w8a8_mixed_fp16 | Qualcomm® QCS9075 | 11.66 ms | 5 - 8 MB | NPU | NASNet | ONNX | w8a8_mixed_fp16 | Snapdragon® 8 Elite For Galaxy Mobile | 5.651 ms | 6 - 362 MB | NPU | NASNet | ONNX | w8a8_mixed_fp16 | Snapdragon® 8 Elite Gen 5 Mobile | 4.589 ms | 5 - 358 MB | NPU | NASNet | ONNX | w8a8_mixed_fp16 | Snapdragon® X2 Elite | 4.504 ms | 100 - 100 MB | NPU | NASNet | QNN_DLC | float | Snapdragon® X Elite | 19.34 ms | 1 - 1 MB | NPU | NASNet | QNN_DLC | float | Snapdragon® 8 Gen 3 Mobile | 12.455 ms | 0 - 814 MB | NPU | NASNet | QNN_DLC | float | Qualcomm® QCS8275 (Proxy) | 53.959 ms | 1 - 659 MB | NPU | NASNet | QNN_DLC | float | Qualcomm® QCS8550 (Proxy) | 19.142 ms | 1 - 2 MB | NPU | NASNet | QNN_DLC | float | Qualcomm® QCS9075 | 29.033 ms | 1 - 3 MB | NPU | NASNet | QNN_DLC | float | Qualcomm® QCS8450 (Proxy) | 34.65 ms | 0 - 794 MB | NPU | NASNet | QNN_DLC | float | Snapdragon® 8 Elite For Galaxy Mobile | 11.042 ms | 0 - 655 MB | NPU | NASNet | QNN_DLC | float | Snapdragon® 8 Elite Gen 5 Mobile | 8.769 ms | 1 - 657 MB | NPU | NASNet | QNN_DLC | float | Snapdragon® X2 Elite | 9.047 ms | 1 - 1 MB | NPU | NASNet | QNN_DLC | w8a8_mixed_fp16 | Snapdragon® X Elite | 9.174 ms | 0 - 0 MB | NPU | NASNet | QNN_DLC | w8a8_mixed_fp16 | Snapdragon® 8 Gen 3 Mobile | 6.034 ms | 0 - 494 MB | NPU | NASNet | QNN_DLC | w8a8_mixed_fp16 | Qualcomm® QCS8275 (Proxy) | 16.411 ms | 0 - 378 MB | NPU | NASNet | QNN_DLC | w8a8_mixed_fp16 | Qualcomm® QCS8550 (Proxy) | 8.89 ms | 0 - 4 MB | NPU | NASNet | QNN_DLC | w8a8_mixed_fp16 | Qualcomm® QCS9075 | 9.427 ms | 0 - 2 MB | NPU | NASNet | QNN_DLC | w8a8_mixed_fp16 | Qualcomm® QCS8450 (Proxy) | 11.055 ms | 0 - 506 MB | NPU | NASNet | QNN_DLC | w8a8_mixed_fp16 | Snapdragon® 8 Elite For Galaxy Mobile | 5.004 ms | 0 - 378 MB | NPU | NASNet | QNN_DLC | w8a8_mixed_fp16 | Snapdragon® 8 Elite Gen 5 Mobile | 3.867 ms | 0 - 379 MB | NPU | NASNet | QNN_DLC | w8a8_mixed_fp16 | Snapdragon® X2 Elite | 4.092 ms | 0 - 0 MB | NPU | NASNet | TFLITE | float | Snapdragon® 8 Gen 3 Mobile | 8.775 ms | 0 - 783 MB | NPU | NASNet | TFLITE | float | Qualcomm® QCS8275 (Proxy) | 44.328 ms | 0 - 629 MB | NPU | NASNet | TFLITE | float | Qualcomm® QCS8550 (Proxy) | 12.574 ms | 0 - 4 MB | NPU | NASNet | TFLITE | float | Qualcomm® QCS9075 | 15.543 ms | 0 - 192 MB | NPU | NASNet | TFLITE | float | Qualcomm® QCS8450 (Proxy) | 28.997 ms | 0 - 766 MB | NPU | NASNet | TFLITE | float | Snapdragon® 8 Elite For Galaxy Mobile | 6.954 ms | 0 - 634 MB | NPU | NASNet | TFLITE | float | Snapdragon® 8 Elite Gen 5 Mobile | 5.666 ms | 0 - 616 MB | NPU ## License * The license for the original implementation of NASNet can be found [here](https://github.com/huggingface/pytorch-image-models?tab=Apache-2.0-1-ov-file). ## References * [Learning Transferable Architectures for Scalable Image Recognition](https://arxiv.org/abs/1707.07012) * [Source Model Implementation](https://github.com/huggingface/pytorch-image-models/tree/main) ## Community * Join [our AI Hub Slack community](https://aihub.qualcomm.com/community/slack) to collaborate, post questions and learn more about on-device AI. * For questions or feedback please [reach out to us](mailto:ai-hub-support@qti.qualcomm.com).