Added hardware support matrix

README.md

@@ -2,8 +2,214 @@
 ![version](https://img.shields.io/badge/version-20.12-0091BD)
 > A collection of machine learning models optimized for Arm IP.
 
-## Key Links
-- [Model List](MODEL_LIST.md)
+
+## Image Classification
+
+<table>
+    <tr>
+        <th width="250" style="text-align:left">Network</th>
+        <th width="100" style="text-align:left">Type</th>
+        <th width="160" style="text-align:left">Framework</th>
+        <th width="100" style="text-align:left">Cortex-A</th>
+        <th width="100" style="text-align:left">Cortex-M</th>
+        <th width="100" style="text-align:left">Mali GPU</th>
+        <th width="100" style="text-align:left">Ethos U</th>
+    </tr>
+    <tr>
+        <td>[MobileNet v2 1.0 224 UINT8](models/image_classification/mobilenet_v2_1.0_224/tflite_uint8)</td>
+        <td>UINT8</td>
+        <td>TensorFlow Lite</td>
+        <td align="center">:heavy_check_mark:</td>
+        <td align="center">:heavy_multiplication_x:</td>
+        <td align="center">:heavy_check_mark:</td>
+        <td align="center">:heavy_check_mark:</td>
+    </tr>
+<table/>
+
+## Keyword Spotting
+
+<table>
+    <tr>
+        <th width="250" style="text-align:left">Network</th>
+        <th width="100" style="text-align:left">Type</th>
+        <th width="160" style="text-align:left">Framework</th>
+        <th width="100" style="text-align:left">Cortex-A</th>
+        <th width="100" style="text-align:left">Cortex-M</th>
+        <th width="100" style="text-align:left">Mali GPU</th>
+        <th width="100" style="text-align:left">Ethos U</th>
+    </tr>
+    <tr>
+        <td>[CNN Large INT8 *](models/keyword_spotting/cnn_large/tflite_int8)</td>
+        <td>INT8</td>
+        <td>TensorFlow Lite</td>
+        <td align="center">:heavy_check_mark:</td>
+        <td align="center">:heavy_check_mark:</td>
+        <td align="center">:heavy_check_mark:</td>
+        <td align="center">:heavy_check_mark:</td>
+    </tr>
+    <tr>
+        <td>[CNN Medium INT8 *](models/keyword_spotting/cnn_medium/tflite_int8)</td>
+        <td>INT8</td>
+        <td>TensorFlow Lite</td>
+        <td align="center">:heavy_check_mark:</td>
+        <td align="center">:heavy_check_mark:</td>
+        <td align="center">:heavy_check_mark:</td>
+        <td align="center">:heavy_check_mark:</td>
+    </tr>
+    <tr>
+        <td>[CNN Small INT8 *](models/keyword_spotting/cnn_small/tflite_int8)</td>
+        <td>INT8</td>
+        <td>TensorFlow Lite</td>
+        <td align="center">:heavy_check_mark:</td>
+        <td align="center">:heavy_check_mark:</td>
+        <td align="center">:heavy_check_mark:</td>
+        <td align="center">:heavy_check_mark:</td>
+    </tr>
+    <tr>
+        <td>[DNN Large INT8 *](models/keyword_spotting/dnn_large/tflite_int8)</td>
+        <td>INT8</td>
+        <td>TensorFlow Lite</td>
+        <td align="center">:heavy_check_mark:</td>
+        <td align="center">:heavy_check_mark:</td>
+        <td align="center">:heavy_check_mark:</td>
+        <td align="center">:heavy_check_mark:</td>
+    </tr>
+    <tr>
+        <td>[DNN Medium INT8 *](models/keyword_spotting/dnn_medium/tflite_int8)</td>
+        <td>INT8</td>
+        <td>TensorFlow Lite</td>
+        <td align="center">:heavy_check_mark:</td>
+        <td align="center">:heavy_check_mark:</td>
+        <td align="center">:heavy_check_mark:</td>
+        <td align="center">:heavy_check_mark:</td>
+    </tr>
+    <tr>
+        <td>[DNN Small INT8 *](models/keyword_spotting/dnn_small/tflite_int8)</td>
+        <td>INT8</td>
+        <td>TensorFlow Lite</td>
+        <td align="center">:heavy_check_mark:</td>
+        <td align="center">:heavy_check_mark:</td>
+        <td align="center">:heavy_check_mark:</td>
+        <td align="center">:heavy_check_mark:</td>
+    </tr>
+    <tr>
+        <td>[DS-CNN Clustered FP32](models/keyword_spotting/ds_cnn_large/tflite_clustered_fp32)</td>
+        <td>FP32</td>
+        <td>TensorFlow Lite</td>
+        <td align="center">:heavy_check_mark:</td>
+        <td align="center">:heavy_multiplication_x:</td>
+        <td align="center">:heavy_check_mark:</td>
+        <td align="center">:heavy_multiplication_x:</td>
+    </tr>
+    <tr>
+        <td>[DS-CNN Clustered INT8](models/keyword_spotting/ds_cnn_large/tflite_clustered_int8)</td>
+        <td>INT8</td>
+        <td>TensorFlow Lite</td>
+        <td align="center">:heavy_check_mark:</td>
+        <td align="center">:heavy_check_mark:</td>
+        <td align="center">:heavy_check_mark:</td>
+        <td align="center">:heavy_check_mark:</td>
+    </tr>
+    <tr>
+        <td>[DS-CNN Large INT8 *](models/keyword_spotting/ds_cnn_large/tflite_int8)</td>
+        <td>INT8</td>
+        <td>TensorFlow Lite</td>
+        <td align="center">:heavy_check_mark:</td>
+        <td align="center">:heavy_check_mark:</td>
+        <td align="center">:heavy_check_mark:</td>
+        <td align="center">:heavy_check_mark:</td>
+    </tr>
+    <tr>
+        <td>[DS-CNN Medium INT8 *](models/keyword_spotting/ds_cnn_medium/tflite_int8)</td>
+        <td>INT8</td>
+        <td>TensorFlow Lite</td>
+        <td align="center">:heavy_check_mark:</td>
+        <td align="center">:heavy_check_mark:</td>
+        <td align="center">:heavy_check_mark:</td>
+        <td align="center">:heavy_check_mark:</td>
+    </tr>
+    <tr>
+        <td>[DS-CNN Small INT8 *](models/keyword_spotting/ds_cnn_small/tflite_int8)</td>
+        <td>INT8</td>
+        <td>TensorFlow Lite</td>
+        <td align="center">:heavy_check_mark:</td>
+        <td align="center">:heavy_check_mark:</td>
+        <td align="center">:heavy_check_mark:</td>
+        <td align="center">:heavy_check_mark:</td>
+    </tr>
+<table/>
+
+## Object Detection
+
+<table>
+    <tr>
+        <th width="250" style="text-align:left">Network</th>
+        <th width="100" style="text-align:left">Type</th>
+        <th width="160" style="text-align:left">Framework</th>
+        <th width="100" style="text-align:left">Cortex-A</th>
+        <th width="100" style="text-align:left">Cortex-M</th>
+        <th width="100" style="text-align:left">Mali GPU</th>
+        <th width="100" style="text-align:left">Ethos U</th>
+    </tr>
+    <tr>
+        <td>[SSD MobileNet v1 FP32 *](models/object_detection/ssd_mobilenet_v1/tflite_fp32)</td>
+        <td>FP32</td>
+        <td>TensorFlow Lite</td>
+        <td align="center">:heavy_check_mark:</td>
+        <td align="center">:heavy_multiplication_x:</td>
+        <td align="center">:heavy_check_mark:</td>
+        <td align="center">:heavy_multiplication_x:</td>
+    </tr>
+    <tr>
+        <td>[SSD MobileNet v1 UINT8 *](models/object_detection/ssd_mobilenet_v1/tflite_uint8)</td>
+        <td>UINT8</td>
+        <td>TensorFlow Lite</td>
+        <td align="center">:heavy_check_mark:</td>
+        <td align="center">:heavy_multiplication_x:</td>
+        <td align="center">:heavy_check_mark:</td>
+        <td align="center">:heavy_multiplication_x:</td>
+    </tr>
+    <tr>
+        <td>[YOLO v3 Tiny FP32 *](models/object_detection/yolo_v3_tiny/tflite_fp32)</td>
+        <td>FP32</td>
+        <td>TensorFlow Lite</td>
+        <td align="center">:heavy_check_mark:</td>
+        <td align="center">:heavy_multiplication_x:</td>
+        <td align="center">:heavy_check_mark:</td>
+        <td align="center">:heavy_multiplication_x:</td>
+    </tr>
+<table/>
+
+## Speech Recognition
+
+<table>
+    <tr>
+        <th width="250" style="text-align:left">Network</th>
+        <th width="100" style="text-align:left">Type</th>
+        <th width="160" style="text-align:left">Framework</th>
+        <th width="100" style="text-align:left">Cortex-A</th>
+        <th width="100" style="text-align:left">Cortex-M</th>
+        <th width="100" style="text-align:left">Mali GPU</th>
+        <th width="100" style="text-align:left">Ethos U</th>
+    </tr>
+    <tr>
+        <td>[Wav2letter INT8](models/speech_recognition/wav2letter/tflite_int8)</td>
+        <td>INT8</td>
+        <td>TensorFlow Lite</td>
+        <td align="center">:heavy_check_mark:</td>
+        <td align="center">:heavy_check_mark:</td>
+        <td align="center">:heavy_check_mark:</td>
+        <td align="center">:heavy_check_mark:</td>
+    </tr>
+<table/>
+
+### Key
+* :heavy_check_mark: - Will run on this platform.
+* :heavy_multiplication_x: - Will not run on this platform.
+* `*` - Code to recreate model available.
+
 
 ## License
 [Apache-2.0](https://spdx.org/licenses/Apache-2.0.html) unless otherwise explicitly stated.
+
+

models/image_classification/mobilenet_v2_1.0_224/tflite_uint8/README.md

@@ -17,14 +17,16 @@ MobileNet v2 is an efficient image classification neural network, targeted for m
 
 ## Performance
 | Platform | Optimized |
-| -------- | ---------- |
-|   CPU    |      :heavy_check_mark:      |
-|   GPU    |      :heavy_check_mark:      |
+|----------|:---------:|
+| Cortex-A |:heavy_check_mark:         |
+| Cortex-M |:heavy_multiplication_x:         |
+| Mali GPU |:heavy_check_mark:         |
+| Ethos U  |:heavy_check_mark:         |
 
 ### Key
- - :heavy_check_mark: - Optimized for the platform.
- - :heavy_minus_sign: - Not optimized, but will run on the platform.
- - :heavy_multiplication_x: - Not optimized and will not run on the platform.g
+* :heavy_check_mark: - Will run on this platform.
+* :heavy_multiplication_x: - Will not run on this platform.
+
 
 ## Accuracy
 Dataset: Ilsvrc 2012

models/keyword_spotting/cnn_large/tflite_int8/README.md

@@ -24,14 +24,17 @@ Dataset: Google Speech Commands Test Set
 
 ## Performance
 | Platform | Optimized |
-| -------- | ---------- |
-|   CPU    |      :heavy_check_mark:      |
-|   GPU    |      :heavy_check_mark:      |
+|----------|:---------:|
+| Cortex-A |:heavy_check_mark:         |
+| Cortex-M |:heavy_check_mark:         |
+| Mali GPU |:heavy_check_mark:         |
+| Ethos U  |:heavy_check_mark:         |
 
 ### Key
- - :heavy_check_mark: - Optimized for the platform.
- - :heavy_minus_sign: - Not optimized, but will run on the platform.
- - :heavy_multiplication_x: - Not optimized and will not run on the platform.
+* :heavy_check_mark: - Will run on this platform.
+* :heavy_multiplication_x: - Will not run on this platform.
+
+
 
 ## Optimizations
 | Optimization |  Value  |

models/keyword_spotting/cnn_medium/tflite_int8/README.md

@@ -24,14 +24,17 @@ Dataset: Google Speech Commands Test Set
 
 ## Performance
 | Platform | Optimized |
-| -------- | ---------- |
-|   CPU    |      :heavy_check_mark:      |
-|   GPU    |      :heavy_check_mark:      |
+|----------|:---------:|
+| Cortex-A |:heavy_check_mark:         |
+| Cortex-M |:heavy_check_mark:         |
+| Mali GPU |:heavy_check_mark:         |
+| Ethos U  |:heavy_check_mark:         |
 
 ### Key
- - :heavy_check_mark: - Optimized for the platform.
- - :heavy_minus_sign: - Not optimized, but will run on the platform.
- - :heavy_multiplication_x: - Not optimized and will not run on the platform.
+* :heavy_check_mark: - Will run on this platform.
+* :heavy_multiplication_x: - Will not run on this platform.
+
+
 
 ## Optimizations
 | Optimization |  Value  |

models/keyword_spotting/cnn_small/tflite_int8/README.md

@@ -24,14 +24,17 @@ Dataset: Google Speech Commands Test Set
 
 ## Performance
 | Platform | Optimized |
-| -------- | ---------- |
-|   CPU    |      :heavy_check_mark:      |
-|   GPU    |      :heavy_check_mark:      |
+|----------|:---------:|
+| Cortex-A |:heavy_check_mark:         |
+| Cortex-M |:heavy_check_mark:         |
+| Mali GPU |:heavy_check_mark:         |
+| Ethos U  |:heavy_check_mark:         |
 
 ### Key
- - :heavy_check_mark: - Optimized for the platform.
- - :heavy_minus_sign: - Not optimized, but will run on the platform.
- - :heavy_multiplication_x: - Not optimized and will not run on the platform.
+* :heavy_check_mark: - Will run on this platform.
+* :heavy_multiplication_x: - Will not run on this platform.
+
+
 
 ## Optimizations
 | Optimization |  Value  |

models/keyword_spotting/dnn_large/tflite_int8/README.md

@@ -24,14 +24,17 @@ Dataset: Google Speech Commands Test Set
 
 ## Performance
 | Platform | Optimized |
-| -------- | ---------- |
-|   CPU    |      :heavy_check_mark:      |
-|   GPU    |      :heavy_check_mark:      |
+|----------|:---------:|
+| Cortex-A |:heavy_check_mark:         |
+| Cortex-M |:heavy_check_mark:         |
+| Mali GPU |:heavy_check_mark:         |
+| Ethos U  |:heavy_check_mark:         |
 
 ### Key
- - :heavy_check_mark: - Optimized for the platform.
- - :heavy_minus_sign: - Not optimized, but will run on the platform.
- - :heavy_multiplication_x: - Not optimized and will not run on the platform.
+* :heavy_check_mark: - Will run on this platform.
+* :heavy_multiplication_x: - Will not run on this platform.
+
+
 
 ## Optimizations
 | Optimization |  Value  |

models/keyword_spotting/dnn_medium/tflite_int8/README.md

@@ -24,14 +24,17 @@ Dataset: Google Speech Commands Test Set
 
 ## Performance
 | Platform | Optimized |
-| -------- | ---------- |
-|   CPU    |      :heavy_check_mark:      |
-|   GPU    |      :heavy_check_mark:      |
+|----------|:---------:|
+| Cortex-A |:heavy_check_mark:         |
+| Cortex-M |:heavy_check_mark:         |
+| Mali GPU |:heavy_check_mark:         |
+| Ethos U  |:heavy_check_mark:         |
 
 ### Key
- - :heavy_check_mark: - Optimized for the platform.
- - :heavy_minus_sign: - Not optimized, but will run on the platform.
- - :heavy_multiplication_x: - Not optimized and will not run on the platform.
+* :heavy_check_mark: - Will run on this platform.
+* :heavy_multiplication_x: - Will not run on this platform.
+
+
 
 ## Optimizations
 | Optimization |  Value  |