Add efficientnet_v2_s and efficientnet_convnext_tiny

efficientnet_convnext_tiny/benchmark_results.json

@@ -0,0 +1,46 @@
+{
+  "benchmarked_devices": [
+    "cuda"
+  ],
+  "parameters": {
+    "total_params": 27833970,
+    "trainable_params": 14304786,
+    "total_params_m": 27.83397,
+    "trainable_params_m": 14.304786
+  },
+  "cuda": {
+    "memory": {
+      "baseline_memory_mb": 288.01806640625,
+      "peak_memory_mb": 302.52978515625,
+      "current_memory_mb": 288.5927734375,
+      "inference_memory_mb": 14.51171875,
+      "device": "cuda"
+    },
+    "inference_single": {
+      "batch_size": 1,
+      "mean_latency_ms": 3.333154039864894,
+      "std_latency_ms": 0.017885221990080365,
+      "median_latency_ms": 3.3323254974675365,
+      "p95_latency_ms": 3.342447954128147,
+      "p99_latency_ms": 3.4173157204349995,
+      "min_latency_ms": 3.309674000774976,
+      "max_latency_ms": 3.4711439948296174,
+      "latency_per_image_ms": 3.333154039864894,
+      "throughput_img_per_sec": 300.01613728015224,
+      "device": "cuda"
+    },
+    "inference_batch32": {
+      "batch_size": 32,
+      "mean_latency_ms": 43.753516640135786,
+      "std_latency_ms": 0.14937650685647402,
+      "median_latency_ms": 43.77655049756868,
+      "p95_latency_ms": 43.96002150133427,
+      "p99_latency_ms": 44.032208782446105,
+      "min_latency_ms": 43.447225994896144,
+      "max_latency_ms": 44.07967900624499,
+      "latency_per_image_ms": 1.3672973950042433,
+      "throughput_img_per_sec": 731.3697836723345,
+      "device": "cuda"
+    }
+  }
+}

efficientnet_convnext_tiny/class_mapping.txt

@@ -0,0 +1,18 @@
+battery:0
+car_battery:1
+cardboard:2
+food_organics:3
+glass:4
+light_bulb:5
+metal:6
+mirror:7
+miscellaneous_trash:8
+neon:9
+paper:10
+pharmacy:11
+plastic:12
+printer_cartridge:13
+textile_trash:14
+tire:15
+vegetation:16
+wood:17

efficientnet_convnext_tiny/classification_report.txt

@@ -0,0 +1,27 @@
+Classification Report
+================================================================================
+
+                     precision    recall  f1-score   support
+
+            battery     1.0000    0.9928    0.9964       276
+        car_battery     0.9959    1.0000    0.9979       243
+          cardboard     0.9504    0.9091    0.9293       253
+      food_organics     0.9489    0.9701    0.9594       268
+              glass     0.9451    0.9627    0.9538       268
+         light_bulb     0.9961    0.9922    0.9941       256
+              metal     0.8737    0.9468    0.9088       263
+             mirror     0.9913    1.0000    0.9957       229
+miscellaneous_trash     0.8972    0.9116    0.9044       249
+               neon     0.9912    0.9912    0.9912       227
+              paper     0.9192    0.8852    0.9019       270
+           pharmacy     0.9921    1.0000    0.9960       252
+            plastic     0.9120    0.8383    0.8736       235
+  printer_cartridge     0.9710    1.0000    0.9853       234
+      textile_trash     0.9655    0.9333    0.9492       240
+               tire     1.0000    1.0000    1.0000       248
+         vegetation     0.9884    0.9884    0.9884       259
+               wood     0.9914    1.0000    0.9957       230
+
+           accuracy                         0.9620      4500
+          macro avg     0.9627    0.9623    0.9623      4500
+       weighted avg     0.9622    0.9620    0.9618      4500

efficientnet_convnext_tiny/model.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4dcc47ea9d3969aead7decb248ca389361e98117447e65f3a4ba362f1486d70a
+size 111390619

efficientnet_convnext_tiny/training_history.txt

@@ -0,0 +1,192 @@
+Training History
+================================================================================
+
+Epoch 1:
+  Train Loss: 0.782064
+  Train Accuracy: 0.796889
+  Val Accuracy: 0.914889
+  Learning Rate: 0.00010000
+  Epoch Time: 37.31s
+
+Epoch 2:
+  Train Loss: 0.207995
+  Train Accuracy: 0.940963
+  Val Accuracy: 0.941111
+  Learning Rate: 0.00010000
+  Epoch Time: 33.06s
+
+Epoch 3:
+  Train Loss: 0.117121
+  Train Accuracy: 0.966963
+  Val Accuracy: 0.949778
+  Learning Rate: 0.00010000
+  Epoch Time: 33.10s
+
+Epoch 4:
+  Train Loss: 0.072624
+  Train Accuracy: 0.981481
+  Val Accuracy: 0.950444
+  Learning Rate: 0.00010000
+  Epoch Time: 33.08s
+
+Epoch 5:
+  Train Loss: 0.057926
+  Train Accuracy: 0.984148
+  Val Accuracy: 0.949333
+  Learning Rate: 0.00010000
+  Epoch Time: 33.11s
+
+Epoch 6:
+  Train Loss: 0.047323
+  Train Accuracy: 0.987630
+  Val Accuracy: 0.952444
+  Learning Rate: 0.00010000
+  Epoch Time: 33.07s
+
+Epoch 7:
+  Train Loss: 0.042567
+  Train Accuracy: 0.987630
+  Val Accuracy: 0.955333
+  Learning Rate: 0.00010000
+  Epoch Time: 33.12s
+
+Epoch 8:
+  Train Loss: 0.036085
+  Train Accuracy: 0.990074
+  Val Accuracy: 0.956444
+  Learning Rate: 0.00010000
+  Epoch Time: 33.10s
+
+Epoch 9:
+  Train Loss: 0.030210
+  Train Accuracy: 0.991556
+  Val Accuracy: 0.953111
+  Learning Rate: 0.00010000
+  Epoch Time: 33.11s
+
+Epoch 10:
+  Train Loss: 0.033506
+  Train Accuracy: 0.990593
+  Val Accuracy: 0.956222
+  Learning Rate: 0.00010000
+  Epoch Time: 33.09s
+
+Epoch 11:
+  Train Loss: 0.027432
+  Train Accuracy: 0.992519
+  Val Accuracy: 0.957556
+  Learning Rate: 0.00010000
+  Epoch Time: 33.11s
+
+Epoch 12:
+  Train Loss: 0.027744
+  Train Accuracy: 0.991926
+  Val Accuracy: 0.956000
+  Learning Rate: 0.00010000
+  Epoch Time: 33.09s
+
+Epoch 13:
+  Train Loss: 0.026804
+  Train Accuracy: 0.991778
+  Val Accuracy: 0.959111
+  Learning Rate: 0.00010000
+  Epoch Time: 33.11s
+
+Epoch 14:
+  Train Loss: 0.024633
+  Train Accuracy: 0.992444
+  Val Accuracy: 0.958889
+  Learning Rate: 0.00010000
+  Epoch Time: 33.08s
+
+Epoch 15:
+  Train Loss: 0.025374
+  Train Accuracy: 0.992815
+  Val Accuracy: 0.959111
+  Learning Rate: 0.00010000
+  Epoch Time: 33.10s
+
+Epoch 16:
+  Train Loss: 0.021128
+  Train Accuracy: 0.993630
+  Val Accuracy: 0.957111
+  Learning Rate: 0.00010000
+  Epoch Time: 33.08s
+
+Epoch 17:
+  Train Loss: 0.020442
+  Train Accuracy: 0.994148
+  Val Accuracy: 0.962000
+  Learning Rate: 0.00010000
+  Epoch Time: 33.10s
+
+Epoch 18:
+  Train Loss: 0.016865
+  Train Accuracy: 0.994963
+  Val Accuracy: 0.959333
+  Learning Rate: 0.00010000
+  Epoch Time: 33.11s
+
+Epoch 19:
+  Train Loss: 0.019852
+  Train Accuracy: 0.994148
+  Val Accuracy: 0.958667
+  Learning Rate: 0.00010000
+  Epoch Time: 33.10s
+
+Epoch 20:
+  Train Loss: 0.021633
+  Train Accuracy: 0.992296
+  Val Accuracy: 0.958444
+  Learning Rate: 0.00010000
+  Epoch Time: 33.10s
+
+Epoch 21:
+  Train Loss: 0.024195
+  Train Accuracy: 0.992519
+  Val Accuracy: 0.959778
+  Learning Rate: 0.00010000
+  Epoch Time: 33.09s
+
+Epoch 22:
+  Train Loss: 0.018204
+  Train Accuracy: 0.995037
+  Val Accuracy: 0.959111
+  Learning Rate: 0.00010000
+  Epoch Time: 33.08s
+
+Epoch 23:
+  Train Loss: 0.018052
+  Train Accuracy: 0.994074
+  Val Accuracy: 0.957778
+  Learning Rate: 0.00010000
+  Epoch Time: 33.08s
+
+Epoch 24:
+  Train Loss: 0.014583
+  Train Accuracy: 0.995704
+  Val Accuracy: 0.960000
+  Learning Rate: 0.00005000
+  Epoch Time: 33.10s
+
+Epoch 25:
+  Train Loss: 0.009234
+  Train Accuracy: 0.997333
+  Val Accuracy: 0.959778
+  Learning Rate: 0.00005000
+  Epoch Time: 33.10s
+
+Epoch 26:
+  Train Loss: 0.010239
+  Train Accuracy: 0.996667
+  Val Accuracy: 0.958667
+  Learning Rate: 0.00005000
+  Epoch Time: 33.10s
+
+Epoch 27:
+  Train Loss: 0.011204
+  Train Accuracy: 0.996667
+  Val Accuracy: 0.960000
+  Learning Rate: 0.00005000
+  Epoch Time: 33.08s
+

efficientnet_v2_s_custom_v2/benchmark_results.json

@@ -0,0 +1,46 @@
+{
+  "benchmarked_devices": [
+    "cuda"
+  ],
+  "parameters": {
+    "total_params": 20200546,
+    "trainable_params": 18378970,
+    "total_params_m": 20.200546,
+    "trainable_params_m": 18.37897
+  },
+  "cuda": {
+    "memory": {
+      "baseline_memory_mb": 314.5068359375,
+      "peak_memory_mb": 321.7763671875,
+      "current_memory_mb": 315.08154296875,
+      "inference_memory_mb": 7.26953125,
+      "device": "cuda"
+    },
+    "inference_single": {
+      "batch_size": 1,
+      "mean_latency_ms": 11.141735980272642,
+      "std_latency_ms": 0.04346183998720415,
+      "median_latency_ms": 11.13254650044837,
+      "p95_latency_ms": 11.162551251254627,
+      "p99_latency_ms": 11.36283751031442,
+      "min_latency_ms": 11.105517995019909,
+      "max_latency_ms": 11.442682000051718,
+      "latency_per_image_ms": 11.141735980272642,
+      "throughput_img_per_sec": 89.7526203969096,
+      "device": "cuda"
+    },
+    "inference_batch32": {
+      "batch_size": 32,
+      "mean_latency_ms": 41.95240021988866,
+      "std_latency_ms": 0.12925345881619055,
+      "median_latency_ms": 41.9302680020337,
+      "p95_latency_ms": 41.98958034867246,
+      "p99_latency_ms": 42.48272620359785,
+      "min_latency_ms": 41.89117100031581,
+      "max_latency_ms": 42.82661800243659,
+      "latency_per_image_ms": 1.3110125068715206,
+      "throughput_img_per_sec": 762.7692297049917,
+      "device": "cuda"
+    }
+  }
+}

efficientnet_v2_s_custom_v2/class_mapping.txt

@@ -0,0 +1,18 @@
+battery:0
+car_battery:1
+cardboard:2
+food_organics:3
+glass:4
+light_bulb:5
+metal:6
+mirror:7
+miscellaneous_trash:8
+neon:9
+paper:10
+pharmacy:11
+plastic:12
+printer_cartridge:13
+textile_trash:14
+tire:15
+vegetation:16
+wood:17

efficientnet_v2_s_custom_v2/classification_report.txt

@@ -0,0 +1,27 @@
+Classification Report
+================================================================================
+
+                     precision    recall  f1-score   support
+
+            battery     0.9928    0.9964    0.9946       276
+        car_battery     1.0000    1.0000    1.0000       243
+          cardboard     0.9631    0.9289    0.9457       253
+      food_organics     0.9660    0.9552    0.9606       268
+              glass     0.9731    0.9440    0.9583       268
+         light_bulb     1.0000    0.9883    0.9941       256
+              metal     0.8921    0.9430    0.9168       263
+             mirror     0.9957    1.0000    0.9978       229
+miscellaneous_trash     0.9328    0.8916    0.9117       249
+               neon     0.9870    1.0000    0.9934       227
+              paper     0.9101    0.9000    0.9050       270
+           pharmacy     0.9882    0.9960    0.9921       252
+            plastic     0.8631    0.8851    0.8739       235
+  printer_cartridge     0.9915    0.9915    0.9915       234
+      textile_trash     0.9218    0.9333    0.9275       240
+               tire     0.9920    1.0000    0.9960       248
+         vegetation     0.9848    1.0000    0.9923       259
+               wood     1.0000    1.0000    1.0000       230
+
+           accuracy                         0.9636      4500
+          macro avg     0.9641    0.9641    0.9640      4500
+       weighted avg     0.9638    0.9636    0.9636      4500

efficientnet_v2_s_custom_v2/model.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1cf90bb35d87e00208fb69923f1098690455403374ce2739dd77cdab440ea16d
+size 81658491

efficientnet_v2_s_custom_v2/training_history.txt

@@ -0,0 +1,325 @@
+Training History
+================================================================================
+
+Epoch 1:
+  Train Loss: 0.790740
+  Train Accuracy: 0.791704
+  Val Accuracy: 0.920667
+  Learning Rate: 0.00010000
+  Epoch Time: 52.34s
+
+Epoch 2:
+  Train Loss: 0.211552
+  Train Accuracy: 0.937407
+  Val Accuracy: 0.942444
+  Learning Rate: 0.00010000
+  Epoch Time: 51.90s
+
+Epoch 3:
+  Train Loss: 0.122989
+  Train Accuracy: 0.961926
+  Val Accuracy: 0.948889
+  Learning Rate: 0.00010000
+  Epoch Time: 51.88s
+
+Epoch 4:
+  Train Loss: 0.082435
+  Train Accuracy: 0.975037
+  Val Accuracy: 0.952000
+  Learning Rate: 0.00010000
+  Epoch Time: 51.91s
+
+Epoch 5:
+  Train Loss: 0.060021
+  Train Accuracy: 0.982000
+  Val Accuracy: 0.947778
+  Learning Rate: 0.00010000
+  Epoch Time: 51.88s
+
+Epoch 6:
+  Train Loss: 0.057305
+  Train Accuracy: 0.981926
+  Val Accuracy: 0.951778
+  Learning Rate: 0.00010000
+  Epoch Time: 51.87s
+
+Epoch 7:
+  Train Loss: 0.045054
+  Train Accuracy: 0.985556
+  Val Accuracy: 0.958000
+  Learning Rate: 0.00010000
+  Epoch Time: 51.88s
+
+Epoch 8:
+  Train Loss: 0.041575
+  Train Accuracy: 0.988222
+  Val Accuracy: 0.957111
+  Learning Rate: 0.00010000
+  Epoch Time: 51.86s
+
+Epoch 9:
+  Train Loss: 0.033491
+  Train Accuracy: 0.989926
+  Val Accuracy: 0.957333
+  Learning Rate: 0.00010000
+  Epoch Time: 51.90s
+
+Epoch 10:
+  Train Loss: 0.038237
+  Train Accuracy: 0.987556
+  Val Accuracy: 0.957778
+  Learning Rate: 0.00010000
+  Epoch Time: 51.86s
+
+Epoch 11:
+  Train Loss: 0.030728
+  Train Accuracy: 0.990148
+  Val Accuracy: 0.952222
+  Learning Rate: 0.00010000
+  Epoch Time: 51.88s
+
+Epoch 12:
+  Train Loss: 0.033249
+  Train Accuracy: 0.989407
+  Val Accuracy: 0.956222
+  Learning Rate: 0.00010000
+  Epoch Time: 51.88s
+
+Epoch 13:
+  Train Loss: 0.034960
+  Train Accuracy: 0.988593
+  Val Accuracy: 0.951333
+  Learning Rate: 0.00010000
+  Epoch Time: 51.87s
+
+Epoch 14:
+  Train Loss: 0.022761
+  Train Accuracy: 0.992444
+  Val Accuracy: 0.958444
+  Learning Rate: 0.00005000
+  Epoch Time: 51.93s
+
+Epoch 15:
+  Train Loss: 0.012618
+  Train Accuracy: 0.996000
+  Val Accuracy: 0.958222
+  Learning Rate: 0.00005000
+  Epoch Time: 51.90s
+
+Epoch 16:
+  Train Loss: 0.012236
+  Train Accuracy: 0.996741
+  Val Accuracy: 0.956667
+  Learning Rate: 0.00005000
+  Epoch Time: 51.86s
+
+Epoch 17:
+  Train Loss: 0.011240
+  Train Accuracy: 0.996593
+  Val Accuracy: 0.958000
+  Learning Rate: 0.00005000
+  Epoch Time: 51.82s
+
+Epoch 18:
+  Train Loss: 0.010189
+  Train Accuracy: 0.996222
+  Val Accuracy: 0.956222
+  Learning Rate: 0.00005000
+  Epoch Time: 51.93s
+
+Epoch 19:
+  Train Loss: 0.012398
+  Train Accuracy: 0.995481
+  Val Accuracy: 0.957111
+  Learning Rate: 0.00005000
+  Epoch Time: 51.84s
+
+Epoch 20:
+  Train Loss: 0.009429
+  Train Accuracy: 0.996815
+  Val Accuracy: 0.955111
+  Learning Rate: 0.00005000
+  Epoch Time: 51.90s
+
+Epoch 21:
+  Train Loss: 0.008474
+  Train Accuracy: 0.997630
+  Val Accuracy: 0.960444
+  Learning Rate: 0.00002500
+  Epoch Time: 51.87s
+
+Epoch 22:
+  Train Loss: 0.007404
+  Train Accuracy: 0.997778
+  Val Accuracy: 0.962000
+  Learning Rate: 0.00002500
+  Epoch Time: 51.90s
+
+Epoch 23:
+  Train Loss: 0.006255
+  Train Accuracy: 0.998000
+  Val Accuracy: 0.960222
+  Learning Rate: 0.00002500
+  Epoch Time: 51.89s
+
+Epoch 24:
+  Train Loss: 0.007118
+  Train Accuracy: 0.998000
+  Val Accuracy: 0.960000
+  Learning Rate: 0.00002500
+  Epoch Time: 51.94s
+
+Epoch 25:
+  Train Loss: 0.006164
+  Train Accuracy: 0.998000
+  Val Accuracy: 0.961111
+  Learning Rate: 0.00002500
+  Epoch Time: 51.89s
+
+Epoch 26:
+  Train Loss: 0.006518
+  Train Accuracy: 0.997037
+  Val Accuracy: 0.958444
+  Learning Rate: 0.00002500
+  Epoch Time: 51.88s
+
+Epoch 27:
+  Train Loss: 0.005568
+  Train Accuracy: 0.997852
+  Val Accuracy: 0.962667
+  Learning Rate: 0.00002500
+  Epoch Time: 51.88s
+
+Epoch 28:
+  Train Loss: 0.007298
+  Train Accuracy: 0.997333
+  Val Accuracy: 0.960444
+  Learning Rate: 0.00002500
+  Epoch Time: 51.88s
+
+Epoch 29:
+  Train Loss: 0.007539
+  Train Accuracy: 0.997481
+  Val Accuracy: 0.958667
+  Learning Rate: 0.00002500
+  Epoch Time: 51.95s
+
+Epoch 30:
+  Train Loss: 0.005392
+  Train Accuracy: 0.997926
+  Val Accuracy: 0.958444
+  Learning Rate: 0.00002500
+  Epoch Time: 51.90s
+
+Epoch 31:
+  Train Loss: 0.004572
+  Train Accuracy: 0.998296
+  Val Accuracy: 0.958889
+  Learning Rate: 0.00002500
+  Epoch Time: 51.90s
+
+Epoch 32:
+  Train Loss: 0.006626
+  Train Accuracy: 0.997852
+  Val Accuracy: 0.956667
+  Learning Rate: 0.00002500
+  Epoch Time: 51.90s
+
+Epoch 33:
+  Train Loss: 0.006203
+  Train Accuracy: 0.997778
+  Val Accuracy: 0.960000
+  Learning Rate: 0.00002500
+  Epoch Time: 51.90s
+
+Epoch 34:
+  Train Loss: 0.006261
+  Train Accuracy: 0.997852
+  Val Accuracy: 0.960444
+  Learning Rate: 0.00001250
+  Epoch Time: 51.87s
+
+Epoch 35:
+  Train Loss: 0.004751
+  Train Accuracy: 0.998148
+  Val Accuracy: 0.960444
+  Learning Rate: 0.00001250
+  Epoch Time: 51.90s
+
+Epoch 36:
+  Train Loss: 0.004397
+  Train Accuracy: 0.997926
+  Val Accuracy: 0.963556
+  Learning Rate: 0.00001250
+  Epoch Time: 51.87s
+
+Epoch 37:
+  Train Loss: 0.003549
+  Train Accuracy: 0.998815
+  Val Accuracy: 0.961333
+  Learning Rate: 0.00001250
+  Epoch Time: 51.87s
+
+Epoch 38:
+  Train Loss: 0.003816
+  Train Accuracy: 0.998889
+  Val Accuracy: 0.960444
+  Learning Rate: 0.00001250
+  Epoch Time: 51.90s
+
+Epoch 39:
+  Train Loss: 0.004508
+  Train Accuracy: 0.998222
+  Val Accuracy: 0.960222
+  Learning Rate: 0.00001250
+  Epoch Time: 51.92s
+
+Epoch 40:
+  Train Loss: 0.004999
+  Train Accuracy: 0.998222
+  Val Accuracy: 0.962444
+  Learning Rate: 0.00001250
+  Epoch Time: 51.89s
+
+Epoch 41:
+  Train Loss: 0.005224
+  Train Accuracy: 0.998296
+  Val Accuracy: 0.962667
+  Learning Rate: 0.00001250
+  Epoch Time: 51.87s
+
+Epoch 42:
+  Train Loss: 0.004965
+  Train Accuracy: 0.998296
+  Val Accuracy: 0.963333
+  Learning Rate: 0.00001250
+  Epoch Time: 51.88s
+
+Epoch 43:
+  Train Loss: 0.003975
+  Train Accuracy: 0.998667
+  Val Accuracy: 0.963333
+  Learning Rate: 0.00000625
+  Epoch Time: 52.01s
+
+Epoch 44:
+  Train Loss: 0.003588
+  Train Accuracy: 0.998741
+  Val Accuracy: 0.961333
+  Learning Rate: 0.00000625
+  Epoch Time: 51.92s
+
+Epoch 45:
+  Train Loss: 0.003801
+  Train Accuracy: 0.998667
+  Val Accuracy: 0.962000
+  Learning Rate: 0.00000625
+  Epoch Time: 51.85s
+
+Epoch 46:
+  Train Loss: 0.003075
+  Train Accuracy: 0.998815
+  Val Accuracy: 0.961778
+  Learning Rate: 0.00000625
+  Epoch Time: 51.86s
+