Hugging Face's logo

selvaa
/
segformer-b1-finetuned-cityscapes-1024-1024-latestt

Transformers
TensorBoard
Safetensors
segformer
Generated from Trainer
Inference Endpoints
Model card Files Metrics Community
selvaa commited on
Commit
de92830
·
verified ·
1 Parent(s): 59548ce

End of training

Browse files
Files changed (5) hide show
  1. README.md +110 -0
  2. config.json +82 -0
  3. model.safetensors +3 -0
  4. runs/May23_20-55-30_antd/events.out.tfevents.1716490538.antd.8486.0 +3 -0
  5. training_args.bin +3 -0
README.md ADDED
@@ -0,0 +1,110 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: other
3
+ base_model: nvidia/segformer-b1-finetuned-cityscapes-1024-1024
4
+ tags:
5
+ - vision
6
+ - image-segmentation
7
+ - generated_from_trainer
8
+ model-index:
9
+ - name: segformer-b1-finetuned-cityscapes-1024-1024-latestt
10
+ results: []
11
+ ---
12
+
13
+ <!-- This model card has been generated automatically according to the information the Trainer had access to. You
14
+ should probably proofread and complete it, then remove this comment. -->
15
+
16
+ # segformer-b1-finetuned-cityscapes-1024-1024-latestt
17
+
18
+ This model is a fine-tuned version of [nvidia/segformer-b1-finetuned-cityscapes-1024-1024](https://huggingface.co/nvidia/segformer-b1-finetuned-cityscapes-1024-1024) on the selvaa/straighter-only-final dataset.
19
+ It achieves the following results on the evaluation set:
20
+ - Loss: 0.0311
21
+ - Mean Iou: 0.9368
22
+ - Mean Accuracy: 0.9605
23
+ - Overall Accuracy: 0.9893
24
+ - Accuracy Default: 1e-06
25
+ - Accuracy Pipe: 0.8961
26
+ - Accuracy Floor: 0.9893
27
+ - Accuracy Background: 0.9960
28
+ - Iou Default: 1e-06
29
+ - Iou Pipe: 0.8401
30
+ - Iou Floor: 0.9817
31
+ - Iou Background: 0.9886
32
+
33
+ ## Model description
34
+
35
+ More information needed
36
+
37
+ ## Intended uses & limitations
38
+
39
+ More information needed
40
+
41
+ ## Training and evaluation data
42
+
43
+ More information needed
44
+
45
+ ## Training procedure
46
+
47
+ ### Training hyperparameters
48
+
49
+ The following hyperparameters were used during training:
50
+ - learning_rate: 0.0002
51
+ - train_batch_size: 3
52
+ - eval_batch_size: 3
53
+ - seed: 42
54
+ - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
55
+ - lr_scheduler_type: linear
56
+ - num_epochs: 60
57
+
58
+ ### Training results
59
+
60
+ | Training Loss | Epoch | Step | Validation Loss | Mean Iou | Mean Accuracy | Overall Accuracy | Accuracy Default | Accuracy Pipe | Accuracy Floor | Accuracy Background | Iou Default | Iou Pipe | Iou Floor | Iou Background |
61
+ |:-------------:|:-----:|:----:|:---------------:|:--------:|:-------------:|:----------------:|:----------------:|:-------------:|:--------------:|:-------------------:|:-----------:|:--------:|:---------:|:--------------:|
62
+ | 0.5401 | 1.0 | 36 | 0.2151 | 0.7616 | 0.8215 | 0.9549 | 1e-06 | 0.5060 | 0.9791 | 0.9793 | 1e-06 | 0.4188 | 0.9090 | 0.9571 |
63
+ | 0.1576 | 2.0 | 72 | 0.1166 | 0.8481 | 0.8906 | 0.9737 | 1e-06 | 0.7051 | 0.9739 | 0.9929 | 1e-06 | 0.6121 | 0.9598 | 0.9724 |
64
+ | 0.0941 | 3.0 | 108 | 0.0739 | 0.8869 | 0.9301 | 0.9802 | 1e-06 | 0.8180 | 0.9807 | 0.9916 | 1e-06 | 0.7164 | 0.9646 | 0.9798 |
65
+ | 0.0678 | 4.0 | 144 | 0.0600 | 0.9001 | 0.9439 | 0.9824 | 1e-06 | 0.8588 | 0.9814 | 0.9916 | 1e-06 | 0.7499 | 0.9685 | 0.9820 |
66
+ | 0.0566 | 5.0 | 180 | 0.0493 | 0.9087 | 0.9531 | 0.9841 | 1e-06 | 0.8887 | 0.9778 | 0.9929 | 1e-06 | 0.7711 | 0.9711 | 0.9840 |
67
+ | 0.0463 | 6.0 | 216 | 0.0454 | 0.9115 | 0.9455 | 0.9847 | 1e-06 | 0.8608 | 0.9809 | 0.9949 | 1e-06 | 0.7769 | 0.9735 | 0.9840 |
68
+ | 0.0412 | 7.0 | 252 | 0.0425 | 0.9138 | 0.9469 | 0.9854 | 1e-06 | 0.8611 | 0.9851 | 0.9945 | 1e-06 | 0.7823 | 0.9739 | 0.9852 |
69
+ | 0.038 | 8.0 | 288 | 0.0379 | 0.9226 | 0.9550 | 0.9868 | 1e-06 | 0.8864 | 0.9836 | 0.9951 | 1e-06 | 0.8049 | 0.9766 | 0.9864 |
70
+ | 0.0337 | 9.0 | 324 | 0.0392 | 0.9184 | 0.9468 | 0.9862 | 1e-06 | 0.8626 | 0.9812 | 0.9966 | 1e-06 | 0.7935 | 0.9765 | 0.9853 |
71
+ | 0.0323 | 10.0 | 360 | 0.0350 | 0.9246 | 0.9544 | 0.9872 | 1e-06 | 0.8827 | 0.9852 | 0.9953 | 1e-06 | 0.8095 | 0.9779 | 0.9864 |
72
+ | 0.0294 | 11.0 | 396 | 0.0350 | 0.9253 | 0.9523 | 0.9873 | 1e-06 | 0.8725 | 0.9898 | 0.9947 | 1e-06 | 0.8114 | 0.9783 | 0.9863 |
73
+ | 0.0275 | 12.0 | 432 | 0.0379 | 0.9185 | 0.9461 | 0.9862 | 1e-06 | 0.8606 | 0.9810 | 0.9968 | 1e-06 | 0.7950 | 0.9748 | 0.9857 |
74
+ | 0.0279 | 13.0 | 468 | 0.0333 | 0.9267 | 0.9572 | 0.9875 | 1e-06 | 0.8914 | 0.9853 | 0.9951 | 1e-06 | 0.8160 | 0.9770 | 0.9871 |
75
+ | 0.0269 | 14.0 | 504 | 0.0323 | 0.9267 | 0.9495 | 0.9878 | 1e-06 | 0.8640 | 0.9878 | 0.9967 | 1e-06 | 0.8141 | 0.9790 | 0.9872 |
76
+ | 0.0239 | 15.0 | 540 | 0.0300 | 0.9324 | 0.9570 | 0.9886 | 1e-06 | 0.8864 | 0.9887 | 0.9959 | 1e-06 | 0.8290 | 0.9802 | 0.9880 |
77
+ | 0.0229 | 16.0 | 576 | 0.0303 | 0.9343 | 0.9610 | 0.9888 | 1e-06 | 0.9005 | 0.9867 | 0.9959 | 1e-06 | 0.8344 | 0.9801 | 0.9882 |
78
+ | 0.0217 | 17.0 | 612 | 0.0318 | 0.9290 | 0.9531 | 0.9882 | 1e-06 | 0.8743 | 0.9889 | 0.9962 | 1e-06 | 0.8197 | 0.9797 | 0.9877 |
79
+ | 0.021 | 18.0 | 648 | 0.0305 | 0.9314 | 0.9540 | 0.9886 | 1e-06 | 0.8756 | 0.9904 | 0.9961 | 1e-06 | 0.8256 | 0.9806 | 0.9879 |
80
+ | 0.0209 | 19.0 | 684 | 0.0296 | 0.9344 | 0.9637 | 0.9887 | 1e-06 | 0.9059 | 0.9915 | 0.9937 | 1e-06 | 0.8357 | 0.9795 | 0.9880 |
81
+ | 0.0199 | 20.0 | 720 | 0.0306 | 0.9335 | 0.9585 | 0.9888 | 1e-06 | 0.8896 | 0.9902 | 0.9955 | 1e-06 | 0.8319 | 0.9802 | 0.9883 |
82
+ | 0.0187 | 21.0 | 756 | 0.0305 | 0.9331 | 0.9576 | 0.9887 | 1e-06 | 0.8877 | 0.9891 | 0.9959 | 1e-06 | 0.8308 | 0.9803 | 0.9881 |
83
+ | 0.0184 | 22.0 | 792 | 0.0298 | 0.9353 | 0.9594 | 0.9891 | 1e-06 | 0.8926 | 0.9896 | 0.9959 | 1e-06 | 0.8364 | 0.9810 | 0.9884 |
84
+ | 0.0175 | 23.0 | 828 | 0.0301 | 0.9340 | 0.9576 | 0.9889 | 1e-06 | 0.8866 | 0.9906 | 0.9957 | 1e-06 | 0.8332 | 0.9806 | 0.9882 |
85
+ | 0.0174 | 24.0 | 864 | 0.0288 | 0.9360 | 0.9633 | 0.9891 | 1e-06 | 0.9062 | 0.9883 | 0.9953 | 1e-06 | 0.8389 | 0.9805 | 0.9885 |
86
+ | 0.0178 | 25.0 | 900 | 0.0306 | 0.9343 | 0.9603 | 0.9889 | 1e-06 | 0.8956 | 0.9902 | 0.9951 | 1e-06 | 0.8342 | 0.9804 | 0.9882 |
87
+ | 0.0165 | 26.0 | 936 | 0.0304 | 0.9349 | 0.9579 | 0.9891 | 1e-06 | 0.8884 | 0.9889 | 0.9963 | 1e-06 | 0.8353 | 0.9810 | 0.9883 |
88
+ | 0.016 | 27.0 | 972 | 0.0300 | 0.9352 | 0.9597 | 0.9891 | 1e-06 | 0.8934 | 0.9902 | 0.9956 | 1e-06 | 0.8362 | 0.9810 | 0.9884 |
89
+ | 0.0159 | 28.0 | 1008 | 0.0311 | 0.9343 | 0.9575 | 0.9890 | 1e-06 | 0.8872 | 0.9891 | 0.9962 | 1e-06 | 0.8340 | 0.9804 | 0.9884 |
90
+ | 0.0157 | 29.0 | 1044 | 0.0302 | 0.9362 | 0.9631 | 0.9891 | 1e-06 | 0.9050 | 0.9894 | 0.9951 | 1e-06 | 0.8389 | 0.9813 | 0.9883 |
91
+ | 0.015 | 30.0 | 1080 | 0.0312 | 0.9344 | 0.9601 | 0.9890 | 1e-06 | 0.8959 | 0.9887 | 0.9957 | 1e-06 | 0.8340 | 0.9809 | 0.9883 |
92
+ | 0.0162 | 31.0 | 1116 | 0.0334 | 0.9321 | 0.9558 | 0.9886 | 1e-06 | 0.8833 | 0.9876 | 0.9965 | 1e-06 | 0.8276 | 0.9807 | 0.9879 |
93
+ | 0.0144 | 32.0 | 1152 | 0.0312 | 0.9352 | 0.9610 | 0.9890 | 1e-06 | 0.8976 | 0.9900 | 0.9952 | 1e-06 | 0.8366 | 0.9805 | 0.9883 |
94
+ | 0.0147 | 33.0 | 1188 | 0.0299 | 0.9375 | 0.9607 | 0.9895 | 1e-06 | 0.8960 | 0.9902 | 0.9959 | 1e-06 | 0.8419 | 0.9817 | 0.9888 |
95
+ | 0.0144 | 34.0 | 1224 | 0.0323 | 0.9342 | 0.9592 | 0.9889 | 1e-06 | 0.8937 | 0.9879 | 0.9961 | 1e-06 | 0.8341 | 0.9802 | 0.9884 |
96
+ | 0.0144 | 35.0 | 1260 | 0.0303 | 0.9359 | 0.9608 | 0.9892 | 1e-06 | 0.8977 | 0.9890 | 0.9959 | 1e-06 | 0.8378 | 0.9812 | 0.9886 |
97
+ | 0.014 | 36.0 | 1296 | 0.0314 | 0.9359 | 0.9577 | 0.9892 | 1e-06 | 0.8878 | 0.9886 | 0.9967 | 1e-06 | 0.8378 | 0.9814 | 0.9885 |
98
+ | 0.0136 | 37.0 | 1332 | 0.0316 | 0.9365 | 0.9600 | 0.9893 | 1e-06 | 0.8954 | 0.9883 | 0.9963 | 1e-06 | 0.8397 | 0.9813 | 0.9885 |
99
+ | 0.0138 | 38.0 | 1368 | 0.0325 | 0.9352 | 0.9577 | 0.9891 | 1e-06 | 0.8869 | 0.9899 | 0.9962 | 1e-06 | 0.8361 | 0.9812 | 0.9884 |
100
+ | 0.0137 | 39.0 | 1404 | 0.0316 | 0.9363 | 0.9597 | 0.9893 | 1e-06 | 0.8933 | 0.9896 | 0.9960 | 1e-06 | 0.8391 | 0.9811 | 0.9886 |
101
+ | 0.0132 | 40.0 | 1440 | 0.0320 | 0.9353 | 0.9590 | 0.9891 | 1e-06 | 0.8930 | 0.9876 | 0.9965 | 1e-06 | 0.8368 | 0.9807 | 0.9884 |
102
+ | 0.0129 | 41.0 | 1476 | 0.0311 | 0.9368 | 0.9605 | 0.9893 | 1e-06 | 0.8961 | 0.9893 | 0.9960 | 1e-06 | 0.8401 | 0.9817 | 0.9886 |
103
+
104
+
105
+ ### Framework versions
106
+
107
+ - Transformers 4.35.2
108
+ - Pytorch 2.0.1
109
+ - Datasets 2.15.0
110
+ - Tokenizers 0.15.0
config.json ADDED
@@ -0,0 +1,82 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "nvidia/segformer-b1-finetuned-cityscapes-1024-1024",
3
+ "architectures": [
4
+ "SegformerForSemanticSegmentation"
5
+ ],
6
+ "attention_probs_dropout_prob": 0.0,
7
+ "classifier_dropout_prob": 0.1,
8
+ "decoder_hidden_size": 256,
9
+ "depths": [
10
+ 2,
11
+ 2,
12
+ 2,
13
+ 2
14
+ ],
15
+ "downsampling_rates": [
16
+ 1,
17
+ 4,
18
+ 8,
19
+ 16
20
+ ],
21
+ "drop_path_rate": 0.1,
22
+ "hidden_act": "gelu",
23
+ "hidden_dropout_prob": 0.0,
24
+ "hidden_sizes": [
25
+ 64,
26
+ 128,
27
+ 320,
28
+ 512
29
+ ],
30
+ "id2label": {
31
+ "0": "default",
32
+ "1": "pipe",
33
+ "2": "floor",
34
+ "3": "background"
35
+ },
36
+ "image_size": 224,
37
+ "initializer_range": 0.02,
38
+ "label2id": {
39
+ "background": 3,
40
+ "default": 0,
41
+ "floor": 2,
42
+ "pipe": 1
43
+ },
44
+ "layer_norm_eps": 1e-06,
45
+ "mlp_ratios": [
46
+ 4,
47
+ 4,
48
+ 4,
49
+ 4
50
+ ],
51
+ "model_type": "segformer",
52
+ "num_attention_heads": [
53
+ 1,
54
+ 2,
55
+ 5,
56
+ 8
57
+ ],
58
+ "num_channels": 3,
59
+ "num_encoder_blocks": 4,
60
+ "patch_sizes": [
61
+ 7,
62
+ 3,
63
+ 3,
64
+ 3
65
+ ],
66
+ "reshape_last_stage": true,
67
+ "semantic_loss_ignore_index": 255,
68
+ "sr_ratios": [
69
+ 8,
70
+ 4,
71
+ 2,
72
+ 1
73
+ ],
74
+ "strides": [
75
+ 4,
76
+ 2,
77
+ 2,
78
+ 2
79
+ ],
80
+ "torch_dtype": "float32",
81
+ "transformers_version": "4.35.2"
82
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f05ab0320141a26e95b78d131ef63b083d7dd79fed34de8c30033abf9a869565
3
+ size 54739432
runs/May23_20-55-30_antd/events.out.tfevents.1716490538.antd.8486.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:54f64d903254879be2a34fff14107da0b7d94d0c43434d57dab53f66c615390d
3
+ size 48412
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e621e314b4d5997b9c11b85ff40a1100606a02d22e149572b69eba374b92640b
3
+ size 4283