Add README to CNN examples

Examples/MAX78000/CNN/cats-vs-dogs-chw/README.md

@@ -1,12 +1,30 @@
 ## Description
 
-TBD<!--TBD-->
+This demo decides whether an image is a cat or a dog.
+
+The code is auto-generated by the ai8x-synthesis tool and runs a known-answer
+test with a pre-defined input sample. This example will be enhance with live capture and TFT
+output in the near future.
 
 ## Required Connections
 
-TBD<!--TBD-->
+-   Connect a USB cable between the PC and the CN1 (USB/PWR) connector.
+-   Connect pins 1 and 2 (P0_1) of the JH1 (UART 0 EN) header.
+-   Open an terminal application on the PC and connect to the EV kit's console UART at 115200, 8-N-1.
 
 ## Expected Output
 
-TBD<!--TBD-->
+The Console UART of the device will output these messages:
+
+```
+*** CNN Test ***
+
+*** PASS ***
+
+Time for CNN: 2526 us
+
+Classification results:
+[-13146] -> Class 0: 19.0%
+[ 13227] -> Class 1: 79.0%
+```
 

Examples/MAX78000/CNN/cifar-10/README.md

@@ -1,12 +1,26 @@
 ## Description
 
-TBD<!--TBD-->
+CIFAR-10 is a simple object classification test, see https://en.wikipedia.org/wiki/CIFAR-10. The input size is 3x32x32.
+
+The code is auto-generated by the ai8x-synthesis tool and runs a known-answer
+test with a pre-defined input sample. This example will be enhance with live capture and TFT
+output in the near future.
 
 ## Required Connections
 
-TBD<!--TBD-->
+-   Connect a USB cable between the PC and the CN1 (USB/PWR) connector.
+-   Connect pins 1 and 2 (P0_1) of the JH1 (UART 0 EN) header.
+-   Open an terminal application on the PC and connect to the EV kit's console UART at 115200, 8-N-1.
 
 ## Expected Output
 
-TBD<!--TBD-->
+The Console UART of the device will output these messages:
+
+```
+*** CNN Test ***
+
+*** PASS ***
+
+Time for CNN: 1685 us
+```
 

Examples/MAX78000/CNN/cifar-100/README.md

@@ -1,12 +1,128 @@
 ## Description
 
-TBD<!--TBD-->
+CIFAR-100 is a simple object classification test, using 100 different classes. The input size is 3x32x32.
+
+The code is auto-generated by the ai8x-synthesis tool and runs a known-answer
+test with a pre-defined input sample. This example will be enhance with live capture and TFT
+output in the near future.
 
 ## Required Connections
 
-TBD<!--TBD-->
+-   Connect a USB cable between the PC and the CN1 (USB/PWR) connector.
+-   Connect pins 1 and 2 (P0_1) of the JH1 (UART 0 EN) header.
+-   Open an terminal application on the PC and connect to the EV kit's console UART at 115200, 8-N-1.
 
 ## Expected Output
 
-TBD<!--TBD-->
+The Console UART of the device will output these messages:
+
+```
+*** CNN Test ***
+
+*** PASS ***
+
+Time for CNN: 2570 us
+
+Classification results:
+[  35997] -> Class 0: 28.7%
+[ -64813] -> Class 1: 0.4%
+[-121777] -> Class 2: 0.0%
+[-131153] -> Class 3: 0.0%
+[-138095] -> Class 4: 0.0%
+[ -69347] -> Class 5: 0.2%
+[-115346] -> Class 6: 0.1%
+[ -76513] -> Class 7: 0.2%
+[-157312] -> Class 8: 0.0%
+[ -77299] -> Class 9: 0.2%
+[ -25727] -> Class 10: 1.8%
+[-150924] -> Class 11: 0.0%
+[-162090] -> Class 12: 0.0%
+[-121492] -> Class 13: 0.0%
+[-104610] -> Class 14: 0.1%
+[ -97926] -> Class 15: 0.1%
+[ -73847] -> Class 16: 0.2%
+[-153786] -> Class 17: 0.0%
+[ -67095] -> Class 18: 0.4%
+[-106266] -> Class 19: 0.1%
+[ -94020] -> Class 20: 0.1%
+[-173073] -> Class 21: 0.0%
+[-107865] -> Class 22: 0.1%
+[ -89592] -> Class 23: 0.1%
+[-114038] -> Class 24: 0.1%
+[ -24653] -> Class 25: 1.8%
+[-116263] -> Class 26: 0.1%
+[-157031] -> Class 27: 0.0%
+[ -54430] -> Class 28: 0.4%
+[-106754] -> Class 29: 0.1%
+[-176985] -> Class 30: 0.0%
+[-134921] -> Class 31: 0.0%
+[ -31922] -> Class 32: 1.8%
+[ -71717] -> Class 33: 0.2%
+[-153564] -> Class 34: 0.0%
+[-173929] -> Class 35: 0.0%
+[-114326] -> Class 36: 0.1%
+[-143840] -> Class 37: 0.0%
+[-203443] -> Class 38: 0.0%
+[-101671] -> Class 39: 0.1%
+[-120907] -> Class 40: 0.0%
+[-160199] -> Class 41: 0.0%
+[-198842] -> Class 42: 0.0%
+[-140705] -> Class 43: 0.0%
+[-106246] -> Class 44: 0.1%
+[ -67912] -> Class 45: 0.4%
+[-155456] -> Class 46: 0.0%
+[ -41800] -> Class 47: 0.9%
+[-228728] -> Class 48: 0.0%
+[-119607] -> Class 49: 0.0%
+[-106100] -> Class 50: 0.1%
+[ -63703] -> Class 51: 0.4%
+[-139052] -> Class 52: 0.0%
+[   2502] -> Class 53: 7.2%
+[-122198] -> Class 54: 0.0%
+[-132491] -> Class 55: 0.0%
+[-114259] -> Class 56: 0.1%
+[  15489] -> Class 57: 14.3%
+[-164336] -> Class 58: 0.0%
+[-112811] -> Class 59: 0.1%
+[ -40528] -> Class 60: 0.9%
+[ -96538] -> Class 61: 0.1%
+[ -53379] -> Class 62: 0.4%
+[-125755] -> Class 63: 0.0%
+[-196914] -> Class 64: 0.0%
+[-125558] -> Class 65: 0.0%
+[-219963] -> Class 66: 0.0%
+[ -63646] -> Class 67: 0.4%
+[ -32604] -> Class 68: 1.8%
+[-114963] -> Class 69: 0.1%
+[ -52485] -> Class 70: 0.9%
+[ -61693] -> Class 71: 0.4%
+[-127933] -> Class 72: 0.0%
+[-119397] -> Class 73: 0.0%
+[-150996] -> Class 74: 0.0%
+[-194726] -> Class 75: 0.0%
+[-125050] -> Class 76: 0.0%
+[ -74555] -> Class 77: 0.2%
+[-114428] -> Class 78: 0.1%
+[-103334] -> Class 79: 0.1%
+[-143704] -> Class 80: 0.0%
+[-135457] -> Class 81: 0.0%
+[ -31013] -> Class 82: 1.8%
+[  37491] -> Class 83: 28.7%
+[ -86443] -> Class 84: 0.1%
+[-189897] -> Class 85: 0.0%
+[-121106] -> Class 86: 0.0%
+[ -90377] -> Class 87: 0.1%
+[-143171] -> Class 88: 0.0%
+[-120577] -> Class 89: 0.0%
+[-135003] -> Class 90: 0.0%
+[ -71930] -> Class 91: 0.2%
+[ -39165] -> Class 92: 0.9%
+[-107021] -> Class 93: 0.1%
+[ -60423] -> Class 94: 0.4%
+[-147585] -> Class 95: 0.0%
+[ -78406] -> Class 96: 0.2%
+[-232354] -> Class 97: 0.0%
+[-157146] -> Class 98: 0.0%
+[ -63513] -> Class 99: 0.4%
+```
 

Examples/MAX78000/CNN/kws20/README.md

@@ -1,12 +1,50 @@
 ## Description
 
-TBD<!--TBD-->
+This demo classifies audio into one of 20 different keywords. It is trained with the 2nd version of Google's
+speech commands dataset. This dataset consists of 35 keywords and more than 100K utterances.
+
+The code is auto-generated by the ai8x-synthesis tool and runs a known-answer
+test with a pre-defined input sample. This example will be enhance with live capture and TFT
+output in the near future.
 
 ## Required Connections
 
-TBD<!--TBD-->
+-   Connect a USB cable between the PC and the CN1 (USB/PWR) connector.
+-   Connect pins 1 and 2 (P0_1) of the JH1 (UART 0 EN) header.
+-   Open an terminal application on the PC and connect to the EV kit's console UART at 115200, 8-N-1.
 
 ## Expected Output
 
-TBD<!--TBD-->
+The Console UART of the device will output these messages:
+
+```
+*** CNN Test ***
+
+*** PASS ***
+
+Time for CNN: 1792 us
+
+Classification results:
+[-227748] -> Class 0: 0.0%
+[-200184] -> Class 1: 0.0%
+[-154269] -> Class 2: 0.1%
+[ -94510] -> Class 3: 1.3%
+[-223096] -> Class 4: 0.0%
+[-177922] -> Class 5: 0.0%
+[-253925] -> Class 6: 0.0%
+[-214902] -> Class 7: 0.0%
+[-129075] -> Class 8: 0.3%
+[-240221] -> Class 9: 0.0%
+[   9379] -> Class 10: 82.6%
+[-391897] -> Class 11: 0.0%
+[-182153] -> Class 12: 0.0%
+[ -52206] -> Class 13: 5.2%
+[ -44969] -> Class 14: 10.3%
+[-328764] -> Class 15: 0.0%
+[-251813] -> Class 16: 0.0%
+[-203954] -> Class 17: 0.0%
+[-176528] -> Class 18: 0.0%
+[-267346] -> Class 19: 0.0%
+[-146511] -> Class 20: 0.1%
+```
 

Examples/MAX78000/CNN/mnist-riscv/README.md

@@ -1,12 +1,40 @@
 ## Description
 
-TBD<!--TBD-->
+This test is identical to the MNIST test, but runs using the RISC-V core instead of the ARM core.
+
+This is the "Hello World" of machine learning. It uses the MNIST dataset of handwritten digits. See http://yann.lecun.com/exdb/mnist/. The input size is 28x28 pixels monochrome (i.e., 1x28x28 in CHW notation).
+
+The code is auto-generated by the ai8x-synthesis tool and runs a known-answer
+test with a pre-defined input sample. This example will be enhance with live capture and TFT
+output in the near future.
 
 ## Required Connections
 
-TBD<!--TBD-->
+-   Connect a USB cable between the PC and the CN1 (USB/PWR) connector.
+-   Connect pins 1 and 2 (P0_1) of the JH1 (UART 0 EN) header.
+-   Open an terminal application on the PC and connect to the EV kit's console UART at 115200, 8-N-1.
 
 ## Expected Output
 
-TBD<!--TBD-->
+The Console UART of the device will output these messages:
+
+```
+*** CNN Test ***
+
+*** PASS ***
+
+Time for CNN: 1333 us
+
+Classification results:
+[ -29649] -> Class 0: 4.2%
+[ -25548] -> Class 1: 4.2%
+[ -27429] -> Class 2: 4.2%
+[ -25779] -> Class 3: 4.2%
+[ -25914] -> Class 4: 4.2%
+[ -38948] -> Class 5: 2.1%
+[ -68083] -> Class 6: 0.5%
+[  39323] -> Class 7: 67.7%
+[ -31039] -> Class 8: 4.2%
+[ -28505] -> Class 9: 4.2%
+```
 

Examples/MAX78000/CNN/mnist/README.md

@@ -1,12 +1,38 @@
 ## Description
 
-TBD<!--TBD-->
+This is the "Hello World" of machine learning. It uses the MNIST dataset of handwritten digits. See http://yann.lecun.com/exdb/mnist/. The input size is 28x28 pixels monochrome (i.e., 1x28x28 in CHW notation).
+
+The code is auto-generated by the ai8x-synthesis tool and runs a known-answer
+test with a pre-defined input sample. This example will be enhance with live capture and TFT
+output in the near future.
 
 ## Required Connections
 
-TBD<!--TBD-->
+-   Connect a USB cable between the PC and the CN1 (USB/PWR) connector.
+-   Connect pins 1 and 2 (P0_1) of the JH1 (UART 0 EN) header.
+-   Open an terminal application on the PC and connect to the EV kit's console UART at 115200, 8-N-1.
 
 ## Expected Output
 
-TBD<!--TBD-->
+The Console UART of the device will output these messages:
+
+```
+*** CNN Test ***
+
+*** PASS ***
+
+Time for CNN: 1386 us
+
+Classification results:
+[ -29649] -> Class 0: 4.2%
+[ -25548] -> Class 1: 4.2%
+[ -27429] -> Class 2: 4.2%
+[ -25779] -> Class 3: 4.2%
+[ -25914] -> Class 4: 4.2%
+[ -38948] -> Class 5: 2.1%
+[ -68083] -> Class 6: 0.5%
+[  39323] -> Class 7: 67.7%
+[ -31039] -> Class 8: 4.2%
+[ -28505] -> Class 9: 4.2%
+```