Upload TFSegformerForSemanticSegmentation
Browse files- README.md +73 -0
- config.json +78 -0
- tf_model.h5 +3 -0
README.md
ADDED
@@ -0,0 +1,73 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
license: other
|
3 |
+
base_model: nvidia/mit-b0
|
4 |
+
tags:
|
5 |
+
- generated_from_keras_callback
|
6 |
+
model-index:
|
7 |
+
- name: sf_20240219_v1
|
8 |
+
results: []
|
9 |
+
---
|
10 |
+
|
11 |
+
<!-- This model card has been generated automatically according to the information Keras had access to. You should
|
12 |
+
probably proofread and complete it, then remove this comment. -->
|
13 |
+
|
14 |
+
# sf_20240219_v1
|
15 |
+
|
16 |
+
This model is a fine-tuned version of [nvidia/mit-b0](https://huggingface.co/nvidia/mit-b0) on an unknown dataset.
|
17 |
+
It achieves the following results on the evaluation set:
|
18 |
+
- Train Loss: 0.0640
|
19 |
+
- Validation Loss: 0.0852
|
20 |
+
- Epoch: 19
|
21 |
+
|
22 |
+
## Model description
|
23 |
+
|
24 |
+
More information needed
|
25 |
+
|
26 |
+
## Intended uses & limitations
|
27 |
+
|
28 |
+
More information needed
|
29 |
+
|
30 |
+
## Training and evaluation data
|
31 |
+
|
32 |
+
More information needed
|
33 |
+
|
34 |
+
## Training procedure
|
35 |
+
|
36 |
+
### Training hyperparameters
|
37 |
+
|
38 |
+
The following hyperparameters were used during training:
|
39 |
+
- optimizer: {'name': 'Adam', 'weight_decay': None, 'clipnorm': None, 'global_clipnorm': None, 'clipvalue': None, 'use_ema': False, 'ema_momentum': 0.99, 'ema_overwrite_frequency': None, 'jit_compile': True, 'is_legacy_optimizer': False, 'learning_rate': 6e-05, 'beta_1': 0.9, 'beta_2': 0.999, 'epsilon': 1e-07, 'amsgrad': False}
|
40 |
+
- training_precision: float32
|
41 |
+
|
42 |
+
### Training results
|
43 |
+
|
44 |
+
| Train Loss | Validation Loss | Epoch |
|
45 |
+
|:----------:|:---------------:|:-----:|
|
46 |
+
| 0.3507 | 0.2338 | 0 |
|
47 |
+
| 0.1528 | 0.1519 | 1 |
|
48 |
+
| 0.1186 | 0.1249 | 2 |
|
49 |
+
| 0.1038 | 0.1116 | 3 |
|
50 |
+
| 0.1008 | 0.0938 | 4 |
|
51 |
+
| 0.0907 | 0.0920 | 5 |
|
52 |
+
| 0.0936 | 0.1053 | 6 |
|
53 |
+
| 0.0898 | 0.0914 | 7 |
|
54 |
+
| 0.0843 | 0.0833 | 8 |
|
55 |
+
| 0.0795 | 0.0828 | 9 |
|
56 |
+
| 0.0779 | 0.0875 | 10 |
|
57 |
+
| 0.0750 | 0.0818 | 11 |
|
58 |
+
| 0.0729 | 0.0877 | 12 |
|
59 |
+
| 0.0723 | 0.0769 | 13 |
|
60 |
+
| 0.0709 | 0.0939 | 14 |
|
61 |
+
| 0.0722 | 0.0787 | 15 |
|
62 |
+
| 0.0692 | 0.0796 | 16 |
|
63 |
+
| 0.0663 | 0.0736 | 17 |
|
64 |
+
| 0.0651 | 0.0727 | 18 |
|
65 |
+
| 0.0640 | 0.0852 | 19 |
|
66 |
+
|
67 |
+
|
68 |
+
### Framework versions
|
69 |
+
|
70 |
+
- Transformers 4.37.0
|
71 |
+
- TensorFlow 2.15.0
|
72 |
+
- Datasets 2.1.0
|
73 |
+
- Tokenizers 0.15.1
|
config.json
ADDED
@@ -0,0 +1,78 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_name_or_path": "nvidia/mit-b0",
|
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 |
+
32,
|
26 |
+
64,
|
27 |
+
160,
|
28 |
+
256
|
29 |
+
],
|
30 |
+
"id2label": {
|
31 |
+
"0": "background",
|
32 |
+
"1": "water"
|
33 |
+
},
|
34 |
+
"image_size": 224,
|
35 |
+
"initializer_range": 0.02,
|
36 |
+
"label2id": {
|
37 |
+
"background": 0,
|
38 |
+
"water": 1
|
39 |
+
},
|
40 |
+
"layer_norm_eps": 1e-06,
|
41 |
+
"mlp_ratios": [
|
42 |
+
4,
|
43 |
+
4,
|
44 |
+
4,
|
45 |
+
4
|
46 |
+
],
|
47 |
+
"model_type": "segformer",
|
48 |
+
"num_attention_heads": [
|
49 |
+
1,
|
50 |
+
2,
|
51 |
+
5,
|
52 |
+
8
|
53 |
+
],
|
54 |
+
"num_channels": 6,
|
55 |
+
"num_encoder_blocks": 4,
|
56 |
+
"patch_sizes": [
|
57 |
+
7,
|
58 |
+
3,
|
59 |
+
3,
|
60 |
+
3
|
61 |
+
],
|
62 |
+
"reshape_last_stage": true,
|
63 |
+
"semantic_loss_ignore_index": 255,
|
64 |
+
"sr_ratios": [
|
65 |
+
8,
|
66 |
+
4,
|
67 |
+
2,
|
68 |
+
1
|
69 |
+
],
|
70 |
+
"strides": [
|
71 |
+
4,
|
72 |
+
2,
|
73 |
+
2,
|
74 |
+
2
|
75 |
+
],
|
76 |
+
"torch_dtype": "float32",
|
77 |
+
"transformers_version": "4.37.0"
|
78 |
+
}
|
tf_model.h5
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:31c7706374181c9ed31229c27a5741d253dcf764dcc35ff9ac2d560e3ae5f311
|
3 |
+
size 15152368
|