Add channelwise INT8 ResNet50 artifact

README.md

@@ -41,12 +41,49 @@ The ResNet-50 architecture is a convolutional neural network pre-trained on the
 
 ## Model description
 
-The model was converted from a checkpoint from PyTorch Vision. 
+The model was converted from a checkpoint from PyTorch Vision.
 
-The original model has:    
-acc@1 (on ImageNet-1K): 76.13%    
-acc@5 (on ImageNet-1K): 92.862%    
-num_params: 25,557,032 
+The original model has:
+acc@1 (on ImageNet-1K): 76.13%
+acc@5 (on ImageNet-1K): 92.862%
+num_params: 25,557,032
+
+## Available Model Files
+
+| File | Description | Quantization |
+|---|---|---|
+| `resnet50.tflite` | Floating-point LiteRT/TFLite model. | Floating-point weights and activations. |
+| `resnet50_dynamic_wi8_afp32.tflite` | Dynamic weight-quantized LiteRT/TFLite model. | INT8 weights with floating-point activations. |
+| `resnet50_int8_channelwise.tflite` | Static INT8 LiteRT/TFLite model. | INT8 weights and INT8 activations, with channelwise weight quantization. |
+
+## Quantization Schema
+
+`resnet50_int8_channelwise.tflite` was quantized with AI Edge Quantizer's
+static W8A8 recipe (`STATIC_WI8_AI8`).
+
+The schema is:
+
+| Tensor group | Quantization |
+|---|---|
+| Weights | INT8, symmetric, channelwise quantization. |
+| Activations | INT8, asymmetric, tensorwise quantization. |
+| Model input | INT8, tensorwise quantized NCHW image tensor with shape `[1, 3, 224, 224]`. |
+| Model output | INT8, tensorwise quantized logits tensor with shape `[1, 1000]`. |
+
+Calibration used real ImageNet validation images with the TorchVision ResNet
+preprocessing flow. When using APIs that expose raw tensor buffers, prepare the
+input and output using the quantization parameters stored in the model.
+
+## Runtime Compatibility
+
+These artifacts are intended for LiteRT CPU and GPU execution. The static INT8
+channelwise artifact is also suitable for Qualcomm NPU deployment through the
+LiteRT Qualcomm compiler plugin and QNN AOT compilation on compatible Qualcomm
+devices.
+
+MediaTek NPU enablement for the static INT8 channelwise artifact is still under
+validation, so this repository does not mark that artifact as MediaTek-NPU-ready
+yet. Use LiteRT CPU/GPU or a compatible Qualcomm NPU path for that file today.
 
 ## Intended uses & limitations
 
@@ -54,9 +91,9 @@ The model files were converted from pretrained weights from PyTorch Vision. The
 
 ## How to Use
 
-​​**1. Install Dependencies** Ensure your Python environment is set up with the required libraries. Run the following command in your terminal:   
+​​**1. Install Dependencies** Ensure your Python environment is set up with the required libraries. Run the following command in your terminal:
 
-```bash 
+```bash
 pip install numpy Pillow huggingface_hub ai-edge-litert
 ```
 
@@ -127,16 +164,17 @@ if __name__ == "__main__":
 
 **4. Execute the Python Script**  Run the below command:
 
-```bash 
+```bash
 python classify.py --image cat.jpg
 ```
 
 ### BibTeX entry and citation info
 
 ```bibtex
-@inproceedings{he2016deep, 
-title={Deep residual learning for image recognition}, 
-author={He, Kaiming and Zhang, Xiangyu and Ren, Shaoqing and Sun, Jian}, 
-booktitle={Proceedings of the IEEE conference on computer vision and pattern recognition}, pages={770--778}, 
-year={2016} 
-}
+@inproceedings{he2016deep,
+title={Deep residual learning for image recognition},
+author={He, Kaiming and Zhang, Xiangyu and Ren, Shaoqing and Sun, Jian},
+booktitle={Proceedings of the IEEE conference on computer vision and pattern recognition}, pages={770--778},
+year={2016}
+}
+```

resnet50_int8_channelwise.tflite

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4b47a9100b1ccd66c1dd5ad59c2ef6bc1ae43dbda240cd9354e1ea874a6f0ebc
+size 26307408