Training Report - test_synthetic - version_16 - Segformer - resnet18

[github-actions]

Segmentation Model Training Report

Project: test_synthetic
Run Time: 2025-05-15T13:03:54.754896

Run Info Summary

Key	Value
project_name	test_synthetic
timestamp	2025-05-15T13:03:30.978142
version_dir	projects/test_synthetic/train/version_16
model_name	Segformer
encoder	resnet18
encoder_weights	imagenet
loss_function	smpDiceLoss
base_lr	0.001
num_GPUs	2
num_epochs	50
batch_size	4
image_size	2048x2048
training_images	225
validation_images	29
valid_dataset_iou	0.727329432964325
valid_per_image_iou	0.7283428311347961

---

Last 10 Training Steps

epoch	step	train_loss	train_class_0_iou	train_class_1_iou	train_class_2_iou	train_dataset_iou	train_per_image_iou	valid_loss	valid_class_0_iou	valid_class_1_iou	valid_class_2_iou	valid_dataset_iou	valid_per_image_iou
41	1175	0.1449	0.8099	0.6888	0.7789	0.7606	0.7619	0.1635	0.8122	0.6695	0.6977	0.7158	0.7169
42	1203	0.1448	0.8095	0.6875	0.7919	0.7677	0.7692	0.158	0.8086	0.7047	0.6903	0.7243	0.7252
43	1231	0.1443	0.811	0.6821	0.7856	0.7631	0.7644	0.1562	0.8112	0.698	0.7027	0.7266	0.7276
44	1259	0.1415	0.8069	0.6953	0.7884	0.7667	0.7681	0.1588	0.8067	0.6954	0.6967	0.7235	0.7244
45	1287	0.1406	0.818	0.693	0.781	0.7651	0.7663	0.1556	0.8119	0.6974	0.7044	0.7276	0.7286
46	1315	0.1429	0.8019	0.6932	0.7967	0.7696	0.7708	0.1575	0.8095	0.6952	0.6994	0.7251	0.7261
47	1343	0.1398	0.8135	0.6958	0.7853	0.7683	0.7696	0.1589	0.8009	0.693	0.7043	0.7227	0.7237
48	1371	0.1411	0.8051	0.6986	0.7922	0.7699	0.771	0.1587	0.8103	0.6858	0.7015	0.723	0.724
49	1399	0.1437	0.8036	0.6907	0.7833	0.7624	0.7637	0.1555	0.8108	0.6969	0.7048	0.7273	0.7283
50	1400	-	-	-	-	-	-	0.1555	0.8108	0.6969	0.7048	0.7273	0.7283

---

Loss and IoU

Test Set Prediction

Real-World Predictions

---

Triggered by push to develop branch.