I am trying to create an audio classification model to specifically identify a “door chime” sound effect and keep false positives as close to zero as possible. The sound effect is a very typical/standard door chime sound and it lasts about 5 seconds total duration. Sounds kind of like a small bell ringing twice in a row (ex: ding pitch1, ding pitch2). Currently my model falsely identifies all kinds of background sounds as the door chime even though they aren’t even close to sounding like it. I have tried all kinds of variations on the training and variations on the training data samples, but I’m not making any real progress. How can I efficiently eliminate false positives (falsely identifying the door chime within random background noise)? Slower learning rate? More Epochs? MFCC instead of MFE?
How did you collect your data?
Laptop microphone from the data collection menu
Okay, I see, did you try to mix the data with a different noise from the background? or you have only registered the door chime sound?
2 background noise samples with no chimes and one sample of chimes only
The issue is mainly with the data collection step, if the door chime sounds look similar after the spectro-analysis to the other sounds then the training step will fail. Could you share the results of the processing block that you are using?
Also it can be helpful if you can share the results of the training and the validation steps.
Did the model fails on the testing data too?
Thanks,
Omar
Will do as soon as i get a chance
MFE output: Not sure why the 3D Visualization drastically changed on the 2nd time I “generated features”. I uploaded both the original 3D visualization and the visualization from the 2nd “feature generation” for comparison.
MFE:
Feature generation output
Creating job… OK (ID: 2021261)
Scheduling job in cluster…
Job started
Creating windows from 3 files…
[1/3] Creating windows from files…
[2/3] Creating windows from files…
[2/3] Creating windows from files…
[3/3] Creating windows from files…
[ 1/912] Resampling windows…
[912/912] Resampling windows…
Resampled 912 windows
Created 912 windows: 1min-chimes: 97, 5min-bgrandom-cleanish: 417, bgnoise-clean2min: 398
Creating features
[ 1/912] Creating features…
[141/912] Creating features…
[293/912] Creating features…
[455/912] Creating features…
[606/912] Creating features…
[762/912] Creating features…
[899/912] Creating features…
[912/912] Creating features…
Created features
Scheduling job in cluster…
Job started
Attached to job 2021261…
Attached to job 2021261…
Reducing dimensions for visualizations…
UMAP(n_components=3, verbose=True)
Construct fuzzy simplicial set
Fri Jan 21 03:37:23 2022 Finding Nearest Neighbors
Fri Jan 21 03:37:25 2022 Finished Nearest Neighbor Search
Still running…
Fri Jan 21 03:37:28 2022 Construct embedding
completed 0 / 500 epochs
completed 50 / 500 epochs
completed 100 / 500 epochs
completed 150 / 500 epochs
completed 200 / 500 epochs
Still running…
completed 250 / 500 epochs
completed 300 / 500 epochs
completed 350 / 500 epochs
completed 400 / 500 epochs
completed 450 / 500 epochs
Fri Jan 21 03:37:33 2022 Finished embedding
Reducing dimensions for visualizations OK
Job completed
NN Classifier Training Output:
Creating job… OK (ID: 2021298)
Job started
Splitting data into training and validation sets…
Splitting data into training and validation sets OK
Training model…
Training on 729 inputs, validating on 183 inputs
Epoch 1/300
23/23 - 2s - loss: 1.1384 - accuracy: 0.4170 - val_loss: 1.0914 - val_accuracy: 0.4317
Epoch 2/300
23/23 - 1s - loss: 1.0698 - accuracy: 0.4033 - val_loss: 1.0789 - val_accuracy: 0.4863
Epoch 3/300
23/23 - 1s - loss: 1.0511 - accuracy: 0.4444 - val_loss: 1.0629 - val_accuracy: 0.5628
Epoch 4/300
23/23 - 1s - loss: 0.9857 - accuracy: 0.5514 - val_loss: 1.0347 - val_accuracy: 0.6339
Epoch 5/300
23/23 - 1s - loss: 0.9431 - accuracy: 0.5885 - val_loss: 0.9967 - val_accuracy: 0.8197
Epoch 6/300
23/23 - 1s - loss: 0.8420 - accuracy: 0.6749 - val_loss: 0.9374 - val_accuracy: 0.9016
Epoch 7/300
23/23 - 1s - loss: 0.7880 - accuracy: 0.7531 - val_loss: 0.8699 - val_accuracy: 0.9781
Epoch 8/300
23/23 - 1s - loss: 0.6931 - accuracy: 0.8313 - val_loss: 0.7942 - val_accuracy: 0.9781
Epoch 9/300
23/23 - 1s - loss: 0.6259 - accuracy: 0.8560 - val_loss: 0.7052 - val_accuracy: 0.9781
Epoch 10/300
23/23 - 1s - loss: 0.5444 - accuracy: 0.8848 - val_loss: 0.6170 - val_accuracy: 0.9781
Epoch 11/300
23/23 - 1s - loss: 0.5179 - accuracy: 0.9081 - val_loss: 0.5432 - val_accuracy: 0.9781
Epoch 12/300
23/23 - 1s - loss: 0.4513 - accuracy: 0.9259 - val_loss: 0.4781 - val_accuracy: 0.9836
Epoch 13/300
23/23 - 1s - loss: 0.3944 - accuracy: 0.9355 - val_loss: 0.4161 - val_accuracy: 0.9836
Epoch 14/300
23/23 - 1s - loss: 0.3651 - accuracy: 0.9328 - val_loss: 0.3670 - val_accuracy: 0.9836
Epoch 15/300
23/23 - 1s - loss: 0.3269 - accuracy: 0.9643 - val_loss: 0.3224 - val_accuracy: 0.9836
Epoch 16/300
23/23 - 1s - loss: 0.3094 - accuracy: 0.9588 - val_loss: 0.2950 - val_accuracy: 0.9891
Epoch 17/300
23/23 - 1s - loss: 0.2886 - accuracy: 0.9588 - val_loss: 0.2561 - val_accuracy: 0.9891
Epoch 18/300
23/23 - 1s - loss: 0.2481 - accuracy: 0.9671 - val_loss: 0.2385 - val_accuracy: 0.9891
Epoch 19/300
23/23 - 1s - loss: 0.2508 - accuracy: 0.9712 - val_loss: 0.2210 - val_accuracy: 0.9891
Epoch 20/300
23/23 - 1s - loss: 0.2581 - accuracy: 0.9671 - val_loss: 0.2096 - val_accuracy: 0.9891
Epoch 21/300
23/23 - 1s - loss: 0.2402 - accuracy: 0.9657 - val_loss: 0.1900 - val_accuracy: 0.9891
Epoch 22/300
23/23 - 1s - loss: 0.2155 - accuracy: 0.9684 - val_loss: 0.1862 - val_accuracy: 0.9891
Epoch 23/300
23/23 - 1s - loss: 0.2140 - accuracy: 0.9726 - val_loss: 0.1773 - val_accuracy: 0.9891
Epoch 24/300
23/23 - 1s - loss: 0.2095 - accuracy: 0.9712 - val_loss: 0.1595 - val_accuracy: 0.9891
Epoch 25/300
23/23 - 1s - loss: 0.1967 - accuracy: 0.9808 - val_loss: 0.1614 - val_accuracy: 0.9891
Epoch 26/300
23/23 - 1s - loss: 0.1880 - accuracy: 0.9753 - val_loss: 0.1537 - val_accuracy: 0.9891
Epoch 27/300
23/23 - 1s - loss: 0.1948 - accuracy: 0.9671 - val_loss: 0.1498 - val_accuracy: 0.9891
Epoch 28/300
23/23 - 1s - loss: 0.2005 - accuracy: 0.9643 - val_loss: 0.1457 - val_accuracy: 0.9891
Epoch 29/300
23/23 - 1s - loss: 0.1868 - accuracy: 0.9781 - val_loss: 0.1457 - val_accuracy: 0.9891
Epoch 30/300
23/23 - 1s - loss: 0.1759 - accuracy: 0.9726 - val_loss: 0.1402 - val_accuracy: 0.9891
Epoch 31/300
23/23 - 1s - loss: 0.1888 - accuracy: 0.9781 - val_loss: 0.1318 - val_accuracy: 0.9891
Epoch 32/300
23/23 - 1s - loss: 0.1683 - accuracy: 0.9739 - val_loss: 0.1265 - val_accuracy: 0.9891
Epoch 33/300
23/23 - 1s - loss: 0.1503 - accuracy: 0.9739 - val_loss: 0.1200 - val_accuracy: 0.9891
Epoch 34/300
23/23 - 1s - loss: 0.1686 - accuracy: 0.9794 - val_loss: 0.1274 - val_accuracy: 0.9891
Epoch 35/300
23/23 - 1s - loss: 0.1767 - accuracy: 0.9739 - val_loss: 0.1297 - val_accuracy: 0.9891
Epoch 36/300
23/23 - 1s - loss: 0.1685 - accuracy: 0.9753 - val_loss: 0.1242 - val_accuracy: 0.9891
Epoch 37/300
23/23 - 1s - loss: 0.1382 - accuracy: 0.9808 - val_loss: 0.1228 - val_accuracy: 0.9891
Epoch 38/300
23/23 - 1s - loss: 0.1663 - accuracy: 0.9753 - val_loss: 0.1220 - val_accuracy: 0.9945
Epoch 39/300
23/23 - 1s - loss: 0.1424 - accuracy: 0.9849 - val_loss: 0.1076 - val_accuracy: 0.9945
Epoch 40/300
23/23 - 1s - loss: 0.1411 - accuracy: 0.9822 - val_loss: 0.1167 - val_accuracy: 0.9945
Epoch 41/300
23/23 - 1s - loss: 0.1566 - accuracy: 0.9726 - val_loss: 0.1177 - val_accuracy: 0.9945
Epoch 42/300
23/23 - 1s - loss: 0.1629 - accuracy: 0.9753 - val_loss: 0.1084 - val_accuracy: 0.9945
Epoch 43/300
23/23 - 1s - loss: 0.1430 - accuracy: 0.9767 - val_loss: 0.1092 - val_accuracy: 0.9945
Epoch 44/300
23/23 - 1s - loss: 0.1422 - accuracy: 0.9794 - val_loss: 0.1087 - val_accuracy: 0.9945
Epoch 45/300
23/23 - 1s - loss: 0.1454 - accuracy: 0.9781 - val_loss: 0.1083 - val_accuracy: 0.9945
Epoch 46/300
23/23 - 1s - loss: 0.1661 - accuracy: 0.9808 - val_loss: 0.1075 - val_accuracy: 0.9945
Epoch 47/300
23/23 - 1s - loss: 0.1492 - accuracy: 0.9794 - val_loss: 0.1151 - val_accuracy: 0.9945
Epoch 48/300
23/23 - 1s - loss: 0.1451 - accuracy: 0.9767 - val_loss: 0.1090 - val_accuracy: 0.9945
Epoch 49/300
23/23 - 1s - loss: 0.1453 - accuracy: 0.9739 - val_loss: 0.1065 - val_accuracy: 0.9945
Epoch 50/300
23/23 - 1s - loss: 0.1374 - accuracy: 0.9835 - val_loss: 0.1084 - val_accuracy: 0.9945
Epoch 51/300
23/23 - 1s - loss: 0.1479 - accuracy: 0.9808 - val_loss: 0.1091 - val_accuracy: 0.9945
Epoch 52/300
23/23 - 1s - loss: 0.1521 - accuracy: 0.9781 - val_loss: 0.1028 - val_accuracy: 0.9945
Epoch 53/300
23/23 - 1s - loss: 0.1400 - accuracy: 0.9835 - val_loss: 0.1078 - val_accuracy: 0.9945
Epoch 54/300
23/23 - 1s - loss: 0.1291 - accuracy: 0.9822 - val_loss: 0.1034 - val_accuracy: 0.9945
Epoch 55/300
23/23 - 1s - loss: 0.1674 - accuracy: 0.9767 - val_loss: 0.1105 - val_accuracy: 0.9945
Epoch 56/300
23/23 - 1s - loss: 0.1358 - accuracy: 0.9794 - val_loss: 0.1073 - val_accuracy: 0.9945
Epoch 57/300
23/23 - 1s - loss: 0.1407 - accuracy: 0.9726 - val_loss: 0.1067 - val_accuracy: 0.9945
Epoch 58/300
23/23 - 1s - loss: 0.1368 - accuracy: 0.9781 - val_loss: 0.1092 - val_accuracy: 0.9945
Epoch 59/300
23/23 - 1s - loss: 0.1219 - accuracy: 0.9822 - val_loss: 0.1066 - val_accuracy: 0.9945
Epoch 60/300
23/23 - 1s - loss: 0.1434 - accuracy: 0.9726 - val_loss: 0.1018 - val_accuracy: 0.9945
Epoch 61/300
23/23 - 1s - loss: 0.1319 - accuracy: 0.9822 - val_loss: 0.1037 - val_accuracy: 0.9945
Epoch 62/300
23/23 - 1s - loss: 0.1400 - accuracy: 0.9835 - val_loss: 0.1081 - val_accuracy: 0.9945
Epoch 63/300
23/23 - 1s - loss: 0.1353 - accuracy: 0.9808 - val_loss: 0.1186 - val_accuracy: 0.9945
Epoch 64/300
23/23 - 1s - loss: 0.1448 - accuracy: 0.9781 - val_loss: 0.1118 - val_accuracy: 0.9945
Epoch 65/300
23/23 - 1s - loss: 0.1399 - accuracy: 0.9835 - val_loss: 0.1102 - val_accuracy: 0.9945
Epoch 66/300
23/23 - 1s - loss: 0.1420 - accuracy: 0.9822 - val_loss: 0.1222 - val_accuracy: 0.9945
Epoch 67/300
23/23 - 1s - loss: 0.1323 - accuracy: 0.9767 - val_loss: 0.1163 - val_accuracy: 0.9945
Epoch 68/300
23/23 - 1s - loss: 0.1209 - accuracy: 0.9877 - val_loss: 0.1113 - val_accuracy: 0.9945
Epoch 69/300
23/23 - 1s - loss: 0.1358 - accuracy: 0.9781 - val_loss: 0.1182 - val_accuracy: 0.9945
Epoch 70/300
23/23 - 1s - loss: 0.1369 - accuracy: 0.9767 - val_loss: 0.1196 - val_accuracy: 0.9945
Epoch 71/300
23/23 - 1s - loss: 0.1281 - accuracy: 0.9822 - val_loss: 0.1205 - val_accuracy: 0.9945
Epoch 72/300
23/23 - 1s - loss: 0.1459 - accuracy: 0.9781 - val_loss: 0.1200 - val_accuracy: 0.9945
Epoch 73/300
23/23 - 1s - loss: 0.1338 - accuracy: 0.9808 - val_loss: 0.1275 - val_accuracy: 0.9891
Epoch 74/300
23/23 - 1s - loss: 0.1381 - accuracy: 0.9808 - val_loss: 0.1239 - val_accuracy: 0.9891
Epoch 75/300
23/23 - 1s - loss: 0.1296 - accuracy: 0.9835 - val_loss: 0.1215 - val_accuracy: 0.9945
Epoch 76/300
23/23 - 1s - loss: 0.1104 - accuracy: 0.9849 - val_loss: 0.1224 - val_accuracy: 0.9891
Epoch 77/300
23/23 - 1s - loss: 0.1163 - accuracy: 0.9849 - val_loss: 0.1196 - val_accuracy: 0.9891
Epoch 78/300
23/23 - 1s - loss: 0.1203 - accuracy: 0.9849 - val_loss: 0.1261 - val_accuracy: 0.9945
Epoch 79/300
23/23 - 1s - loss: 0.1474 - accuracy: 0.9808 - val_loss: 0.1263 - val_accuracy: 0.9945
Epoch 80/300
23/23 - 1s - loss: 0.1253 - accuracy: 0.9863 - val_loss: 0.1221 - val_accuracy: 0.9945
Epoch 81/300
23/23 - 1s - loss: 0.1212 - accuracy: 0.9877 - val_loss: 0.1186 - val_accuracy: 0.9945
Epoch 82/300
23/23 - 1s - loss: 0.1432 - accuracy: 0.9767 - val_loss: 0.1221 - val_accuracy: 0.9945
Epoch 83/300
23/23 - 1s - loss: 0.1451 - accuracy: 0.9781 - val_loss: 0.1204 - val_accuracy: 0.9945
Epoch 84/300
23/23 - 1s - loss: 0.1254 - accuracy: 0.9753 - val_loss: 0.1130 - val_accuracy: 0.9945
Epoch 85/300
23/23 - 1s - loss: 0.1252 - accuracy: 0.9835 - val_loss: 0.1136 - val_accuracy: 0.9945
Epoch 86/300
23/23 - 1s - loss: 0.1265 - accuracy: 0.9849 - val_loss: 0.1115 - val_accuracy: 0.9945
Epoch 87/300
23/23 - 1s - loss: 0.1103 - accuracy: 0.9890 - val_loss: 0.1165 - val_accuracy: 0.9945
Epoch 88/300
23/23 - 1s - loss: 0.1365 - accuracy: 0.9863 - val_loss: 0.1235 - val_accuracy: 0.9945
Epoch 89/300
23/23 - 1s - loss: 0.1209 - accuracy: 0.9849 - val_loss: 0.1150 - val_accuracy: 0.9945
Epoch 90/300
23/23 - 1s - loss: 0.1198 - accuracy: 0.9808 - val_loss: 0.1118 - val_accuracy: 0.9945
Epoch 91/300
23/23 - 1s - loss: 0.1290 - accuracy: 0.9835 - val_loss: 0.1199 - val_accuracy: 0.9945
Epoch 92/300
23/23 - 1s - loss: 0.1142 - accuracy: 0.9849 - val_loss: 0.1162 - val_accuracy: 0.9945
Epoch 93/300
23/23 - 1s - loss: 0.1268 - accuracy: 0.9863 - val_loss: 0.1156 - val_accuracy: 0.9945
Epoch 94/300
23/23 - 1s - loss: 0.1320 - accuracy: 0.9877 - val_loss: 0.1150 - val_accuracy: 0.9945
Epoch 95/300
23/23 - 1s - loss: 0.1297 - accuracy: 0.9808 - val_loss: 0.1140 - val_accuracy: 0.9945
Epoch 96/300
23/23 - 1s - loss: 0.1220 - accuracy: 0.9822 - val_loss: 0.1180 - val_accuracy: 0.9945
Epoch 97/300
23/23 - 1s - loss: 0.1321 - accuracy: 0.9835 - val_loss: 0.1174 - val_accuracy: 0.9945
Epoch 98/300
23/23 - 1s - loss: 0.1250 - accuracy: 0.9822 - val_loss: 0.1187 - val_accuracy: 0.9945
Epoch 99/300
23/23 - 1s - loss: 0.1150 - accuracy: 0.9849 - val_loss: 0.1248 - val_accuracy: 0.9945
Epoch 100/300
23/23 - 1s - loss: 0.1104 - accuracy: 0.9808 - val_loss: 0.1238 - val_accuracy: 0.9945
Epoch 101/300
23/23 - 1s - loss: 0.1423 - accuracy: 0.9822 - val_loss: 0.1238 - val_accuracy: 0.9945
Epoch 102/300
23/23 - 1s - loss: 0.1195 - accuracy: 0.9849 - val_loss: 0.1174 - val_accuracy: 0.9945
Epoch 103/300
23/23 - 1s - loss: 0.1138 - accuracy: 0.9849 - val_loss: 0.1171 - val_accuracy: 0.9945
Epoch 104/300
23/23 - 1s - loss: 0.1442 - accuracy: 0.9835 - val_loss: 0.1156 - val_accuracy: 0.9945
Epoch 105/300
23/23 - 1s - loss: 0.1141 - accuracy: 0.9822 - val_loss: 0.1168 - val_accuracy: 0.9945
Epoch 106/300
23/23 - 1s - loss: 0.1111 - accuracy: 0.9863 - val_loss: 0.1194 - val_accuracy: 0.9945
Epoch 107/300
23/23 - 1s - loss: 0.1459 - accuracy: 0.9794 - val_loss: 0.1229 - val_accuracy: 0.9945
Epoch 108/300
23/23 - 1s - loss: 0.1279 - accuracy: 0.9808 - val_loss: 0.1316 - val_accuracy: 0.9945
Epoch 109/300
23/23 - 1s - loss: 0.1213 - accuracy: 0.9849 - val_loss: 0.1279 - val_accuracy: 0.9945
Epoch 110/300
23/23 - 1s - loss: 0.1207 - accuracy: 0.9849 - val_loss: 0.1250 - val_accuracy: 0.9945
Epoch 111/300
23/23 - 1s - loss: 0.1285 - accuracy: 0.9863 - val_loss: 0.1353 - val_accuracy: 0.9945
Epoch 112/300
23/23 - 1s - loss: 0.1264 - accuracy: 0.9835 - val_loss: 0.1294 - val_accuracy: 0.9945
Epoch 113/300
23/23 - 1s - loss: 0.1216 - accuracy: 0.9835 - val_loss: 0.1303 - val_accuracy: 0.9945
Epoch 114/300
23/23 - 1s - loss: 0.1154 - accuracy: 0.9863 - val_loss: 0.1286 - val_accuracy: 0.9945
Epoch 115/300
23/23 - 1s - loss: 0.1242 - accuracy: 0.9877 - val_loss: 0.1336 - val_accuracy: 0.9945
Epoch 116/300
23/23 - 1s - loss: 0.1069 - accuracy: 0.9835 - val_loss: 0.1270 - val_accuracy: 0.9945
Epoch 117/300
23/23 - 1s - loss: 0.1179 - accuracy: 0.9849 - val_loss: 0.1307 - val_accuracy: 0.9945
Epoch 118/300
23/23 - 1s - loss: 0.1247 - accuracy: 0.9822 - val_loss: 0.1324 - val_accuracy: 0.9945
Epoch 119/300
23/23 - 1s - loss: 0.1136 - accuracy: 0.9835 - val_loss: 0.1300 - val_accuracy: 0.9945
Epoch 120/300
23/23 - 1s - loss: 0.1185 - accuracy: 0.9877 - val_loss: 0.1311 - val_accuracy: 0.9945
Epoch 121/300
23/23 - 1s - loss: 0.1111 - accuracy: 0.9863 - val_loss: 0.1263 - val_accuracy: 0.9945
Epoch 122/300
23/23 - 1s - loss: 0.1103 - accuracy: 0.9877 - val_loss: 0.1229 - val_accuracy: 0.9945
Epoch 123/300
23/23 - 1s - loss: 0.0952 - accuracy: 0.9890 - val_loss: 0.1209 - val_accuracy: 1.0000
Epoch 124/300
23/23 - 1s - loss: 0.1183 - accuracy: 0.9877 - val_loss: 0.1260 - val_accuracy: 1.0000
Epoch 125/300
23/23 - 1s - loss: 0.0973 - accuracy: 0.9877 - val_loss: 0.1237 - val_accuracy: 0.9945
Epoch 126/300
23/23 - 1s - loss: 0.1173 - accuracy: 0.9849 - val_loss: 0.1270 - val_accuracy: 0.9945
Epoch 127/300
23/23 - 1s - loss: 0.0990 - accuracy: 0.9835 - val_loss: 0.1211 - val_accuracy: 1.0000
Epoch 128/300
23/23 - 1s - loss: 0.1026 - accuracy: 0.9849 - val_loss: 0.1191 - val_accuracy: 1.0000
Epoch 129/300
23/23 - 1s - loss: 0.1154 - accuracy: 0.9863 - val_loss: 0.1160 - val_accuracy: 1.0000
Epoch 130/300
23/23 - 1s - loss: 0.1183 - accuracy: 0.9849 - val_loss: 0.1215 - val_accuracy: 0.9945
Epoch 131/300
23/23 - 1s - loss: 0.1104 - accuracy: 0.9808 - val_loss: 0.1252 - val_accuracy: 0.9945
Epoch 132/300
23/23 - 1s - loss: 0.1304 - accuracy: 0.9835 - val_loss: 0.1270 - val_accuracy: 0.9945
Epoch 133/300
23/23 - 1s - loss: 0.1059 - accuracy: 0.9849 - val_loss: 0.1254 - val_accuracy: 0.9945
Epoch 134/300
23/23 - 1s - loss: 0.1162 - accuracy: 0.9822 - val_loss: 0.1165 - val_accuracy: 1.0000
Epoch 135/300
23/23 - 1s - loss: 0.1079 - accuracy: 0.9822 - val_loss: 0.1255 - val_accuracy: 0.9945
Epoch 136/300
23/23 - 1s - loss: 0.1128 - accuracy: 0.9835 - val_loss: 0.1267 - val_accuracy: 0.9945
Epoch 137/300
23/23 - 1s - loss: 0.1080 - accuracy: 0.9808 - val_loss: 0.1233 - val_accuracy: 1.0000
Epoch 138/300
23/23 - 1s - loss: 0.1049 - accuracy: 0.9835 - val_loss: 0.1168 - val_accuracy: 1.0000
Epoch 139/300
23/23 - 1s - loss: 0.1420 - accuracy: 0.9849 - val_loss: 0.1310 - val_accuracy: 0.9945
Epoch 140/300
23/23 - 1s - loss: 0.1032 - accuracy: 0.9863 - val_loss: 0.1324 - val_accuracy: 0.9945
Epoch 141/300
23/23 - 1s - loss: 0.1117 - accuracy: 0.9849 - val_loss: 0.1333 - val_accuracy: 0.9945
Epoch 142/300
23/23 - 1s - loss: 0.0913 - accuracy: 0.9863 - val_loss: 0.1311 - val_accuracy: 0.9945
Epoch 143/300
23/23 - 1s - loss: 0.1150 - accuracy: 0.9863 - val_loss: 0.1246 - val_accuracy: 1.0000
Epoch 144/300
23/23 - 1s - loss: 0.0992 - accuracy: 0.9890 - val_loss: 0.1276 - val_accuracy: 0.9945
Epoch 145/300
23/23 - 1s - loss: 0.1015 - accuracy: 0.9863 - val_loss: 0.1324 - val_accuracy: 0.9945
Epoch 146/300
23/23 - 1s - loss: 0.1044 - accuracy: 0.9849 - val_loss: 0.1301 - val_accuracy: 0.9945
Epoch 147/300
23/23 - 1s - loss: 0.1088 - accuracy: 0.9877 - val_loss: 0.1310 - val_accuracy: 1.0000
Epoch 148/300
23/23 - 1s - loss: 0.1108 - accuracy: 0.9877 - val_loss: 0.1290 - val_accuracy: 1.0000
Epoch 149/300
23/23 - 1s - loss: 0.1066 - accuracy: 0.9835 - val_loss: 0.1298 - val_accuracy: 1.0000
Epoch 150/300
23/23 - 1s - loss: 0.1120 - accuracy: 0.9877 - val_loss: 0.1377 - val_accuracy: 1.0000
Epoch 151/300
23/23 - 1s - loss: 0.0942 - accuracy: 0.9890 - val_loss: 0.1295 - val_accuracy: 0.9945
Epoch 152/300
23/23 - 1s - loss: 0.0923 - accuracy: 0.9890 - val_loss: 0.1290 - val_accuracy: 0.9945
Epoch 153/300
23/23 - 1s - loss: 0.1031 - accuracy: 0.9849 - val_loss: 0.1335 - val_accuracy: 0.9945
Epoch 154/300
23/23 - 1s - loss: 0.1185 - accuracy: 0.9863 - val_loss: 0.1333 - val_accuracy: 0.9945
Epoch 155/300
23/23 - 1s - loss: 0.1209 - accuracy: 0.9877 - val_loss: 0.1250 - val_accuracy: 1.0000
Epoch 156/300
23/23 - 1s - loss: 0.1015 - accuracy: 0.9849 - val_loss: 0.1245 - val_accuracy: 1.0000
Epoch 157/300
23/23 - 1s - loss: 0.1299 - accuracy: 0.9877 - val_loss: 0.1324 - val_accuracy: 0.9945
Epoch 158/300
23/23 - 1s - loss: 0.1008 - accuracy: 0.9835 - val_loss: 0.1296 - val_accuracy: 1.0000
Epoch 159/300
23/23 - 1s - loss: 0.1191 - accuracy: 0.9890 - val_loss: 0.1359 - val_accuracy: 1.0000
Epoch 160/300
23/23 - 1s - loss: 0.0921 - accuracy: 0.9890 - val_loss: 0.1367 - val_accuracy: 1.0000
Epoch 161/300
23/23 - 1s - loss: 0.1147 - accuracy: 0.9835 - val_loss: 0.1375 - val_accuracy: 0.9945
Epoch 162/300
23/23 - 1s - loss: 0.1015 - accuracy: 0.9877 - val_loss: 0.1283 - val_accuracy: 1.0000
Epoch 163/300
23/23 - 1s - loss: 0.1134 - accuracy: 0.9849 - val_loss: 0.1304 - val_accuracy: 1.0000
Epoch 164/300
23/23 - 1s - loss: 0.1142 - accuracy: 0.9890 - val_loss: 0.1316 - val_accuracy: 1.0000
Epoch 165/300
23/23 - 1s - loss: 0.1070 - accuracy: 0.9877 - val_loss: 0.1333 - val_accuracy: 0.9945
Epoch 166/300
23/23 - 1s - loss: 0.1234 - accuracy: 0.9863 - val_loss: 0.1399 - val_accuracy: 0.9945
Epoch 167/300
23/23 - 1s - loss: 0.1149 - accuracy: 0.9863 - val_loss: 0.1356 - val_accuracy: 0.9945
Epoch 168/300
23/23 - 1s - loss: 0.1189 - accuracy: 0.9904 - val_loss: 0.1433 - val_accuracy: 0.9945
Epoch 169/300
23/23 - 1s - loss: 0.1215 - accuracy: 0.9822 - val_loss: 0.1439 - val_accuracy: 0.9945
Epoch 170/300
23/23 - 1s - loss: 0.1072 - accuracy: 0.9890 - val_loss: 0.1395 - val_accuracy: 0.9945
Epoch 171/300
23/23 - 1s - loss: 0.0973 - accuracy: 0.9877 - val_loss: 0.1356 - val_accuracy: 0.9945
Epoch 172/300
23/23 - 1s - loss: 0.1072 - accuracy: 0.9904 - val_loss: 0.1396 - val_accuracy: 0.9945
Epoch 173/300
23/23 - 1s - loss: 0.1069 - accuracy: 0.9863 - val_loss: 0.1401 - val_accuracy: 1.0000
Epoch 174/300
23/23 - 1s - loss: 0.1019 - accuracy: 0.9849 - val_loss: 0.1399 - val_accuracy: 0.9945
Epoch 175/300
23/23 - 1s - loss: 0.1079 - accuracy: 0.9904 - val_loss: 0.1396 - val_accuracy: 0.9945
Epoch 176/300
23/23 - 1s - loss: 0.1007 - accuracy: 0.9863 - val_loss: 0.1366 - val_accuracy: 1.0000
Epoch 177/300
23/23 - 1s - loss: 0.1122 - accuracy: 0.9863 - val_loss: 0.1344 - val_accuracy: 1.0000
Epoch 178/300
23/23 - 1s - loss: 0.1069 - accuracy: 0.9890 - val_loss: 0.1354 - val_accuracy: 1.0000
Epoch 179/300
23/23 - 1s - loss: 0.1116 - accuracy: 0.9863 - val_loss: 0.1405 - val_accuracy: 1.0000
Epoch 180/300
23/23 - 1s - loss: 0.1027 - accuracy: 0.9863 - val_loss: 0.1341 - val_accuracy: 1.0000
Epoch 181/300
23/23 - 1s - loss: 0.1207 - accuracy: 0.9863 - val_loss: 0.1433 - val_accuracy: 0.9945
Epoch 182/300
23/23 - 1s - loss: 0.1095 - accuracy: 0.9863 - val_loss: 0.1379 - val_accuracy: 0.9945
Epoch 183/300
23/23 - 1s - loss: 0.1045 - accuracy: 0.9877 - val_loss: 0.1392 - val_accuracy: 0.9945
Epoch 184/300
23/23 - 1s - loss: 0.1065 - accuracy: 0.9890 - val_loss: 0.1471 - val_accuracy: 0.9945
Epoch 185/300
23/23 - 1s - loss: 0.1144 - accuracy: 0.9863 - val_loss: 0.1491 - val_accuracy: 0.9945
Epoch 186/300
23/23 - 1s - loss: 0.1036 - accuracy: 0.9877 - val_loss: 0.1424 - val_accuracy: 0.9945
Epoch 187/300
23/23 - 1s - loss: 0.1104 - accuracy: 0.9877 - val_loss: 0.1443 - val_accuracy: 0.9945
Epoch 188/300
23/23 - 1s - loss: 0.0867 - accuracy: 0.9877 - val_loss: 0.1434 - val_accuracy: 0.9945
Epoch 189/300
23/23 - 1s - loss: 0.0987 - accuracy: 0.9904 - val_loss: 0.1442 - val_accuracy: 0.9945
Epoch 190/300
23/23 - 1s - loss: 0.1112 - accuracy: 0.9877 - val_loss: 0.1469 - val_accuracy: 0.9945
Epoch 191/300
23/23 - 1s - loss: 0.1126 - accuracy: 0.9877 - val_loss: 0.1386 - val_accuracy: 1.0000
Epoch 192/300
23/23 - 1s - loss: 0.1134 - accuracy: 0.9890 - val_loss: 0.1461 - val_accuracy: 1.0000
Epoch 193/300
23/23 - 1s - loss: 0.1051 - accuracy: 0.9863 - val_loss: 0.1401 - val_accuracy: 1.0000
Epoch 194/300
23/23 - 1s - loss: 0.0945 - accuracy: 0.9849 - val_loss: 0.1372 - val_accuracy: 1.0000
Epoch 195/300
23/23 - 1s - loss: 0.1108 - accuracy: 0.9863 - val_loss: 0.1341 - val_accuracy: 1.0000
Epoch 196/300
23/23 - 1s - loss: 0.1063 - accuracy: 0.9904 - val_loss: 0.1404 - val_accuracy: 1.0000
Epoch 197/300
23/23 - 1s - loss: 0.1030 - accuracy: 0.9877 - val_loss: 0.1401 - val_accuracy: 1.0000
Epoch 198/300
23/23 - 1s - loss: 0.1039 - accuracy: 0.9849 - val_loss: 0.1410 - val_accuracy: 1.0000
Epoch 199/300
23/23 - 1s - loss: 0.1033 - accuracy: 0.9877 - val_loss: 0.1357 - val_accuracy: 1.0000
Epoch 200/300
23/23 - 1s - loss: 0.1015 - accuracy: 0.9877 - val_loss: 0.1498 - val_accuracy: 1.0000
Epoch 201/300
23/23 - 1s - loss: 0.0949 - accuracy: 0.9877 - val_loss: 0.1430 - val_accuracy: 1.0000
Epoch 202/300
23/23 - 1s - loss: 0.1065 - accuracy: 0.9849 - val_loss: 0.1410 - val_accuracy: 1.0000
Epoch 203/300
23/23 - 1s - loss: 0.1022 - accuracy: 0.9877 - val_loss: 0.1443 - val_accuracy: 0.9945
Epoch 204/300
23/23 - 1s - loss: 0.0908 - accuracy: 0.9904 - val_loss: 0.1434 - val_accuracy: 0.9945
Epoch 205/300
23/23 - 1s - loss: 0.1130 - accuracy: 0.9890 - val_loss: 0.1427 - val_accuracy: 1.0000
Epoch 206/300
23/23 - 1s - loss: 0.0954 - accuracy: 0.9890 - val_loss: 0.1448 - val_accuracy: 1.0000
Epoch 207/300
23/23 - 1s - loss: 0.1086 - accuracy: 0.9877 - val_loss: 0.1435 - val_accuracy: 1.0000
Epoch 208/300
23/23 - 1s - loss: 0.0963 - accuracy: 0.9890 - val_loss: 0.1461 - val_accuracy: 1.0000
Epoch 209/300
23/23 - 1s - loss: 0.1104 - accuracy: 0.9904 - val_loss: 0.1446 - val_accuracy: 1.0000
Epoch 210/300
23/23 - 1s - loss: 0.0985 - accuracy: 0.9890 - val_loss: 0.1468 - val_accuracy: 1.0000
Epoch 211/300
23/23 - 1s - loss: 0.1089 - accuracy: 0.9849 - val_loss: 0.1554 - val_accuracy: 1.0000
Epoch 212/300
23/23 - 1s - loss: 0.1060 - accuracy: 0.9877 - val_loss: 0.1553 - val_accuracy: 0.9945
Epoch 213/300
23/23 - 1s - loss: 0.1167 - accuracy: 0.9849 - val_loss: 0.1519 - val_accuracy: 1.0000
Epoch 214/300
23/23 - 1s - loss: 0.1002 - accuracy: 0.9890 - val_loss: 0.1491 - val_accuracy: 1.0000
Epoch 215/300
23/23 - 1s - loss: 0.1183 - accuracy: 0.9877 - val_loss: 0.1519 - val_accuracy: 1.0000
Epoch 216/300
23/23 - 1s - loss: 0.0983 - accuracy: 0.9849 - val_loss: 0.1559 - val_accuracy: 0.9945
Epoch 217/300
23/23 - 1s - loss: 0.1125 - accuracy: 0.9890 - val_loss: 0.1551 - val_accuracy: 1.0000
Epoch 218/300
23/23 - 1s - loss: 0.1088 - accuracy: 0.9863 - val_loss: 0.1507 - val_accuracy: 1.0000
Epoch 219/300
23/23 - 1s - loss: 0.0870 - accuracy: 0.9890 - val_loss: 0.1510 - val_accuracy: 0.9945
Epoch 220/300
23/23 - 1s - loss: 0.1078 - accuracy: 0.9877 - val_loss: 0.1477 - val_accuracy: 1.0000
Epoch 221/300
23/23 - 1s - loss: 0.1167 - accuracy: 0.9890 - val_loss: 0.1476 - val_accuracy: 1.0000
Epoch 222/300
23/23 - 1s - loss: 0.1071 - accuracy: 0.9877 - val_loss: 0.1534 - val_accuracy: 1.0000
Epoch 223/300
23/23 - 1s - loss: 0.1101 - accuracy: 0.9863 - val_loss: 0.1562 - val_accuracy: 1.0000
Epoch 224/300
23/23 - 1s - loss: 0.1075 - accuracy: 0.9877 - val_loss: 0.1532 - val_accuracy: 1.0000
Epoch 225/300
23/23 - 1s - loss: 0.0986 - accuracy: 0.9863 - val_loss: 0.1483 - val_accuracy: 1.0000
Epoch 226/300
23/23 - 1s - loss: 0.1140 - accuracy: 0.9890 - val_loss: 0.1502 - val_accuracy: 1.0000
Epoch 227/300
23/23 - 1s - loss: 0.1162 - accuracy: 0.9877 - val_loss: 0.1502 - val_accuracy: 1.0000
Epoch 228/300
23/23 - 1s - loss: 0.1134 - accuracy: 0.9890 - val_loss: 0.1482 - val_accuracy: 1.0000
Epoch 229/300
23/23 - 1s - loss: 0.1174 - accuracy: 0.9904 - val_loss: 0.1471 - val_accuracy: 1.0000
Epoch 230/300
23/23 - 1s - loss: 0.1112 - accuracy: 0.9849 - val_loss: 0.1565 - val_accuracy: 1.0000
Epoch 231/300
23/23 - 1s - loss: 0.1211 - accuracy: 0.9835 - val_loss: 0.1557 - val_accuracy: 1.0000
Epoch 232/300
23/23 - 1s - loss: 0.1016 - accuracy: 0.9918 - val_loss: 0.1506 - val_accuracy: 1.0000
Epoch 233/300
23/23 - 1s - loss: 0.1137 - accuracy: 0.9863 - val_loss: 0.1526 - val_accuracy: 1.0000
Epoch 234/300
23/23 - 1s - loss: 0.1086 - accuracy: 0.9890 - val_loss: 0.1582 - val_accuracy: 1.0000
Epoch 235/300
23/23 - 1s - loss: 0.1137 - accuracy: 0.9877 - val_loss: 0.1627 - val_accuracy: 1.0000
Epoch 236/300
23/23 - 1s - loss: 0.0911 - accuracy: 0.9877 - val_loss: 0.1553 - val_accuracy: 1.0000
Epoch 237/300
23/23 - 1s - loss: 0.1097 - accuracy: 0.9890 - val_loss: 0.1603 - val_accuracy: 1.0000
Epoch 238/300
23/23 - 1s - loss: 0.1106 - accuracy: 0.9890 - val_loss: 0.1555 - val_accuracy: 1.0000
Epoch 239/300
23/23 - 1s - loss: 0.1109 - accuracy: 0.9904 - val_loss: 0.1600 - val_accuracy: 1.0000
Epoch 240/300
23/23 - 1s - loss: 0.1066 - accuracy: 0.9904 - val_loss: 0.1633 - val_accuracy: 1.0000
Epoch 241/300
23/23 - 1s - loss: 0.1040 - accuracy: 0.9849 - val_loss: 0.1665 - val_accuracy: 1.0000
Epoch 242/300
23/23 - 1s - loss: 0.0916 - accuracy: 0.9890 - val_loss: 0.1559 - val_accuracy: 1.0000
Epoch 243/300
23/23 - 1s - loss: 0.0950 - accuracy: 0.9890 - val_loss: 0.1575 - val_accuracy: 1.0000
Epoch 244/300
23/23 - 1s - loss: 0.0899 - accuracy: 0.9890 - val_loss: 0.1593 - val_accuracy: 1.0000
Epoch 245/300
23/23 - 1s - loss: 0.0946 - accuracy: 0.9877 - val_loss: 0.1611 - val_accuracy: 1.0000
Epoch 246/300
23/23 - 1s - loss: 0.1142 - accuracy: 0.9877 - val_loss: 0.1633 - val_accuracy: 1.0000
Epoch 247/300
23/23 - 1s - loss: 0.1155 - accuracy: 0.9863 - val_loss: 0.1647 - val_accuracy: 1.0000
Epoch 248/300
23/23 - 1s - loss: 0.1090 - accuracy: 0.9890 - val_loss: 0.1697 - val_accuracy: 1.0000
Epoch 249/300
23/23 - 1s - loss: 0.1145 - accuracy: 0.9849 - val_loss: 0.1701 - val_accuracy: 1.0000
Epoch 250/300
23/23 - 1s - loss: 0.1110 - accuracy: 0.9877 - val_loss: 0.1731 - val_accuracy: 0.9945
Epoch 251/300
23/23 - 1s - loss: 0.0980 - accuracy: 0.9877 - val_loss: 0.1649 - val_accuracy: 0.9945
Epoch 252/300
23/23 - 1s - loss: 0.1098 - accuracy: 0.9890 - val_loss: 0.1655 - val_accuracy: 1.0000
Epoch 253/300
23/23 - 1s - loss: 0.1001 - accuracy: 0.9863 - val_loss: 0.1579 - val_accuracy: 1.0000
Epoch 254/300
23/23 - 1s - loss: 0.1030 - accuracy: 0.9890 - val_loss: 0.1597 - val_accuracy: 1.0000
Epoch 255/300
23/23 - 1s - loss: 0.0978 - accuracy: 0.9863 - val_loss: 0.1602 - val_accuracy: 1.0000
Epoch 256/300
23/23 - 1s - loss: 0.0913 - accuracy: 0.9904 - val_loss: 0.1610 - val_accuracy: 1.0000
Epoch 257/300
23/23 - 1s - loss: 0.1166 - accuracy: 0.9890 - val_loss: 0.1665 - val_accuracy: 1.0000
Epoch 258/300
23/23 - 1s - loss: 0.1178 - accuracy: 0.9877 - val_loss: 0.1668 - val_accuracy: 1.0000
Epoch 259/300
23/23 - 1s - loss: 0.1090 - accuracy: 0.9835 - val_loss: 0.1740 - val_accuracy: 1.0000
Epoch 260/300
23/23 - 1s - loss: 0.1038 - accuracy: 0.9890 - val_loss: 0.1734 - val_accuracy: 1.0000
Epoch 261/300
23/23 - 1s - loss: 0.1026 - accuracy: 0.9904 - val_loss: 0.1702 - val_accuracy: 1.0000
Epoch 262/300
23/23 - 1s - loss: 0.0945 - accuracy: 0.9890 - val_loss: 0.1659 - val_accuracy: 1.0000
Epoch 263/300
23/23 - 1s - loss: 0.1229 - accuracy: 0.9863 - val_loss: 0.1721 - val_accuracy: 0.9945
Epoch 264/300
23/23 - 1s - loss: 0.1002 - accuracy: 0.9890 - val_loss: 0.1629 - val_accuracy: 1.0000
Epoch 265/300
23/23 - 1s - loss: 0.1062 - accuracy: 0.9890 - val_loss: 0.1638 - val_accuracy: 1.0000
Epoch 266/300
23/23 - 1s - loss: 0.1035 - accuracy: 0.9877 - val_loss: 0.1671 - val_accuracy: 1.0000
Epoch 267/300
23/23 - 1s - loss: 0.0984 - accuracy: 0.9877 - val_loss: 0.1675 - val_accuracy: 1.0000
Epoch 268/300
23/23 - 1s - loss: 0.1111 - accuracy: 0.9835 - val_loss: 0.1751 - val_accuracy: 1.0000
Epoch 269/300
23/23 - 1s - loss: 0.1005 - accuracy: 0.9849 - val_loss: 0.1769 - val_accuracy: 1.0000
Epoch 270/300
23/23 - 1s - loss: 0.1111 - accuracy: 0.9877 - val_loss: 0.1825 - val_accuracy: 1.0000
Epoch 271/300
23/23 - 1s - loss: 0.0984 - accuracy: 0.9890 - val_loss: 0.1787 - val_accuracy: 1.0000
Epoch 272/300
23/23 - 1s - loss: 0.1128 - accuracy: 0.9904 - val_loss: 0.1744 - val_accuracy: 1.0000
Epoch 273/300
23/23 - 1s - loss: 0.1025 - accuracy: 0.9877 - val_loss: 0.1743 - val_accuracy: 1.0000
Epoch 274/300
23/23 - 1s - loss: 0.0965 - accuracy: 0.9904 - val_loss: 0.1712 - val_accuracy: 1.0000
Epoch 275/300
23/23 - 1s - loss: 0.1066 - accuracy: 0.9890 - val_loss: 0.1762 - val_accuracy: 1.0000
Epoch 276/300
23/23 - 1s - loss: 0.1035 - accuracy: 0.9863 - val_loss: 0.1763 - val_accuracy: 1.0000
Epoch 277/300
23/23 - 1s - loss: 0.1118 - accuracy: 0.9835 - val_loss: 0.1770 - val_accuracy: 1.0000
Epoch 278/300
23/23 - 1s - loss: 0.1264 - accuracy: 0.9863 - val_loss: 0.1762 - val_accuracy: 1.0000
Epoch 279/300
23/23 - 1s - loss: 0.0974 - accuracy: 0.9890 - val_loss: 0.1746 - val_accuracy: 1.0000
Epoch 280/300
23/23 - 1s - loss: 0.0995 - accuracy: 0.9835 - val_loss: 0.1706 - val_accuracy: 1.0000
Epoch 281/300
23/23 - 1s - loss: 0.1092 - accuracy: 0.9890 - val_loss: 0.1696 - val_accuracy: 1.0000
Epoch 282/300
23/23 - 1s - loss: 0.0991 - accuracy: 0.9863 - val_loss: 0.1704 - val_accuracy: 1.0000
Epoch 283/300
23/23 - 1s - loss: 0.1022 - accuracy: 0.9890 - val_loss: 0.1762 - val_accuracy: 1.0000
Epoch 284/300
23/23 - 1s - loss: 0.1049 - accuracy: 0.9877 - val_loss: 0.1776 - val_accuracy: 1.0000
Epoch 285/300
23/23 - 1s - loss: 0.0955 - accuracy: 0.9904 - val_loss: 0.1789 - val_accuracy: 1.0000
Epoch 286/300
23/23 - 1s - loss: 0.1111 - accuracy: 0.9822 - val_loss: 0.1801 - val_accuracy: 1.0000
Epoch 287/300
23/23 - 1s - loss: 0.0976 - accuracy: 0.9835 - val_loss: 0.1766 - val_accuracy: 1.0000
Epoch 288/300
23/23 - 1s - loss: 0.1106 - accuracy: 0.9835 - val_loss: 0.1733 - val_accuracy: 1.0000
Epoch 289/300
23/23 - 1s - loss: 0.0868 - accuracy: 0.9904 - val_loss: 0.1726 - val_accuracy: 1.0000
Epoch 290/300
23/23 - 1s - loss: 0.1025 - accuracy: 0.9877 - val_loss: 0.1771 - val_accuracy: 1.0000
Epoch 291/300
23/23 - 1s - loss: 0.0940 - accuracy: 0.9890 - val_loss: 0.1765 - val_accuracy: 1.0000
Epoch 292/300
23/23 - 1s - loss: 0.0885 - accuracy: 0.9890 - val_loss: 0.1720 - val_accuracy: 1.0000
Epoch 293/300
23/23 - 1s - loss: 0.0891 - accuracy: 0.9890 - val_loss: 0.1758 - val_accuracy: 1.0000
Epoch 294/300
23/23 - 1s - loss: 0.1085 - accuracy: 0.9863 - val_loss: 0.1804 - val_accuracy: 1.0000
Epoch 295/300
23/23 - 1s - loss: 0.0989 - accuracy: 0.9849 - val_loss: 0.1769 - val_accuracy: 1.0000
Epoch 296/300
23/23 - 1s - loss: 0.0966 - accuracy: 0.9890 - val_loss: 0.1761 - val_accuracy: 1.0000
Epoch 297/300
23/23 - 1s - loss: 0.0940 - accuracy: 0.9877 - val_loss: 0.1785 - val_accuracy: 1.0000
Epoch 298/300
23/23 - 1s - loss: 0.1068 - accuracy: 0.9890 - val_loss: 0.1822 - val_accuracy: 1.0000
Epoch 299/300
23/23 - 1s - loss: 0.1052 - accuracy: 0.9877 - val_loss: 0.1854 - val_accuracy: 1.0000
Epoch 300/300
23/23 - 1s - loss: 0.1122 - accuracy: 0.9904 - val_loss: 0.1781 - val_accuracy: 1.0000
Finished training
Saving best performing model…
Converting TensorFlow Lite float32 model…
Converting TensorFlow Lite int8 quantized model with int8 input and output…
Calculating performance metrics…
Profiling float32 model…
Profiling int8 model…
Model training complete
Job completed
I would say that the model is overfitting.
Would you add datasets to the model testing step and let us know about the results?
Thanks,
Omar
I’m using EON tuner to attempt to sort ML models by precision for the “chimes” audio class, but the sort function appears to not do anything. Regardless of what I click the list of models does not re-sort in any way.
I see what you mean, I am saying is that, the model is very trained so it is fitting too well on the training datasets only. Therefore it is overfitting.
I would highly recommend to share the model testing steps so we can help you further.
Thanks,
Omar
Hi @OmarShrit , I have similar question. I have been working with Edge Impulse to detect water faucet sounds and deployed it on a Esp32 Wroom board with Inmp441 external mic. It works great and does water flow predictions really well. However, it turns ON falsely for vacuum cleaner sounds or loud fan sounds or if you talked really loudly. Although I have tried to include these sounds as noise category in Edge Impulse, it somehow does not work in real-world i.e. at the board level. But Edge Impulse classifier works correctly when you test the model online with an accuracy of 97%. I have played around with MFCC, MFE and Spectrogram methods but they all are similar when it comes to real-world situations. Do you have any insights on this? Thanks and much appreciated.
Hi Akash,
You can add (use) an Anomaly detection block after the classification and train this block, it will detect all other sounds and classify them as anomalies.
Regards,
Omar