metadata

license: other
base_model: nvidia/mit-b0
tags:
  - vision
  - image-segmentation
  - generated_from_trainer
model-index:
  - name: segformer-b0-miic-tl
    results: []

segformer-b0-miic-tl

This model is a fine-tuned version of nvidia/mit-b0 on the yijisuk/ic-chip-sample dataset. It achieves the following results on the evaluation set:

Loss: 0.2252
Mean Iou: 0.4213
Mean Accuracy: 0.8427
Overall Accuracy: 0.8427
Accuracy Unlabeled: nan
Accuracy Circuit: 0.8427
Iou Unlabeled: 0.0
Iou Circuit: 0.8427
Dice Coefficient: 0.8060

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

learning_rate: 6e-05
train_batch_size: 2
eval_batch_size: 2
seed: 42
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: linear
num_epochs: 50

Training results

Training Loss	Epoch	Step	Validation Loss	Mean Iou	Mean Accuracy	Overall Accuracy	Accuracy Unlabeled	Accuracy Circuit	Iou Circuit	Dice Coefficient
0.564	1.25	100	0.3976	0.2158	0.4316	0.4316	nan	0.4316	0.4316	0.3032
0.523	2.5	200	0.3853	0.2051	0.4102	0.4102	nan	0.4102	0.4102	0.2797
0.5447	3.75	300	0.3570	0.1866	0.3731	0.3731	nan	0.3731	0.3731	0.2145
0.5087	5.0	400	0.3325	0.2632	0.5264	0.5264	nan	0.5264	0.5264	0.4352
0.5064	6.25	500	0.3596	0.3047	0.6094	0.6094	nan	0.6094	0.6094	0.5244
0.4947	7.5	600	0.3153	0.3062	0.6124	0.6124	nan	0.6124	0.6124	0.5797
0.4703	8.75	700	0.2752	0.4433	0.8866	0.8866	nan	0.8866	0.8866	0.8004
0.4679	10.0	800	0.2900	0.3833	0.7666	0.7666	nan	0.7666	0.7666	0.7333
0.4691	11.25	900	0.3102	0.4024	0.8048	0.8048	nan	0.8048	0.8048	0.7452
0.4648	12.5	1000	0.2768	0.3698	0.7396	0.7396	nan	0.7396	0.7396	0.7157
0.4459	13.75	1100	0.2575	0.4120	0.8239	0.8239	nan	0.8239	0.8239	0.7781
0.446	15.0	1200	0.2927	0.4653	0.9306	0.9306	nan	0.9306	0.9306	0.8262
0.4299	16.25	1300	0.2682	0.3375	0.6749	0.6749	nan	0.6749	0.6749	0.6881
0.4464	17.5	1400	0.2379	0.4282	0.8563	0.8563	nan	0.8563	0.8563	0.8051
0.4241	18.75	1500	0.2479	0.3996	0.7993	0.7993	nan	0.7993	0.7993	0.7770
0.4154	20.0	1600	0.2441	0.4133	0.8265	0.8265	nan	0.8265	0.8265	0.7923
0.428	21.25	1700	0.2505	0.4258	0.8515	0.8515	nan	0.8515	0.8515	0.8082
0.4126	22.5	1800	0.2419	0.4549	0.9097	0.9097	nan	0.9097	0.9097	0.8370
0.3986	23.75	1900	0.2364	0.3863	0.7726	0.7726	nan	0.7726	0.7726	0.7577
0.4053	25.0	2000	0.2419	0.3752	0.7504	0.7504	nan	0.7504	0.7504	0.7367
0.4018	26.25	2100	0.2310	0.4299	0.8598	0.8598	nan	0.8598	0.8598	0.8078
0.4048	27.5	2200	0.2292	0.4288	0.8577	0.8577	nan	0.8577	0.8577	0.8095
0.3838	28.75	2300	0.2294	0.4185	0.8371	0.8371	nan	0.8371	0.8371	0.7979
0.389	30.0	2400	0.2255	0.4337	0.8675	0.8675	nan	0.8675	0.8675	0.8181
0.3889	31.25	2500	0.2247	0.4307	0.8613	0.8613	nan	0.8613	0.8613	0.8180
0.4058	32.5	2600	0.2290	0.3806	0.7611	0.7611	nan	0.7611	0.7611	0.7493
0.3822	33.75	2700	0.2301	0.4023	0.8046	0.8046	nan	0.8046	0.8046	0.7794
0.3807	35.0	2800	0.2261	0.3952	0.7904	0.7904	nan	0.7904	0.7904	0.7691
0.3993	36.25	2900	0.2199	0.4163	0.8326	0.8326	nan	0.8326	0.8326	0.7997
0.3586	37.5	3000	0.2238	0.4098	0.8195	0.8195	nan	0.8195	0.8195	0.7897
0.3894	38.75	3100	0.2334	0.3539	0.7077	0.7077	nan	0.7077	0.7077	0.7093
0.3627	40.0	3200	0.2311	0.3646	0.7292	0.7292	nan	0.7292	0.7292	0.7249
0.3704	41.25	3300	0.2266	0.3876	0.7751	0.7751	nan	0.7751	0.7751	0.7621
0.3808	42.5	3400	0.2227	0.3996	0.7993	0.7993	nan	0.7993	0.7993	0.7793
0.3631	43.75	3500	0.2222	0.3910	0.7820	0.7820	nan	0.7820	0.7820	0.7655
0.367	45.0	3600	0.2253	0.4118	0.8237	0.8237	nan	0.8237	0.8237	0.7939
0.3609	46.25	3700	0.2225	0.4082	0.8165	0.8165	nan	0.8165	0.8165	0.7897
0.3515	47.5	3800	0.2226	0.4210	0.8420	0.8420	nan	0.8420	0.8420	0.8064
0.3888	48.75	3900	0.2283	0.3815	0.7630	0.7630	nan	0.7630	0.7630	0.7509
0.3503	50.0	4000	0.2252	0.4213	0.8427	0.8427	nan	0.8427	0.8427	0.8060

Framework versions

Transformers 4.36.2
Pytorch 1.11.0+cu115
Datasets 2.15.0
Tokenizers 0.15.0