Hugging Face's logo

mouadenna
/
segformer-b1-finetuned-segments-pv_v1_normalized_t4_16batch

Image Segmentation
Transformers
TensorBoard
Safetensors
segformer
vision
Generated from Trainer
Inference Endpoints
Model card Files Metrics Community
segformer-b1-finetuned-segments-pv_v1_normalized_t4_16batch / README.md
mouadenna's picture
mouadenna
Upload processor
e5285a5 verified
preview code
|
raw
history blame
5.59 kB
metadata
base_model: nvidia/segformer-b1-finetuned-ade-512-512
license: other
metrics:
  - precision
tags:
  - vision
  - image-segmentation
  - generated_from_trainer
model-index:
  - name: segformer-b1-finetuned-segments-pv_v1_normalized_t4_16batch
    results: []

Visualize in Weights & Biases

segformer-b1-finetuned-segments-pv_v1_normalized_t4_16batch

This model is a fine-tuned version of nvidia/segformer-b1-finetuned-ade-512-512 on the mouadenna/satellite_PV_dataset_train_test_v1 dataset. It achieves the following results on the evaluation set:

  • Loss: 0.0082
  • Mean Iou: 0.8638
  • Precision: 0.9180

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: 0.0016
  • train_batch_size: 32
  • eval_batch_size: 32
  • seed: 42
  • gradient_accumulation_steps: 4
  • total_train_batch_size: 128
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • num_epochs: 50
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Validation Loss Mean Iou Precision
0.0089 0.9739 28 0.0081 0.7635 0.8665
0.0087 1.9826 57 0.0098 0.7340 0.7592
0.0053 2.9913 86 0.0056 0.8284 0.9168
0.0073 4.0 115 0.0069 0.7988 0.8608
0.005 4.9739 143 0.0059 0.8186 0.8704
0.0042 5.9826 172 0.0059 0.8262 0.8836
0.0047 6.9913 201 0.0058 0.8225 0.8749
0.0044 8.0 230 0.0060 0.8168 0.8677
0.0041 8.9739 258 0.0070 0.8057 0.8494
0.0031 9.9826 287 0.0057 0.8260 0.8832
0.0057 10.9913 316 0.0058 0.8289 0.8744
0.0039 12.0 345 0.0065 0.8190 0.8625
0.0033 12.9739 373 0.0061 0.8316 0.9113
0.0031 13.9826 402 0.0066 0.8329 0.8947
0.0036 14.9913 431 0.0059 0.8392 0.8996
0.0034 16.0 460 0.0060 0.8379 0.8977
0.004 16.9739 488 0.0064 0.8434 0.9075
0.0029 17.9826 517 0.0060 0.8450 0.9010
0.0027 18.9913 546 0.0059 0.8431 0.9038
0.0035 20.0 575 0.0060 0.8445 0.9049
0.0028 20.9739 603 0.0062 0.8474 0.9320
0.0024 21.9826 632 0.0063 0.8442 0.8984
0.0021 22.9913 661 0.0059 0.8526 0.9030
0.0021 24.0 690 0.0059 0.8546 0.9126
0.0026 24.9739 718 0.0065 0.8542 0.9110
0.0024 25.9826 747 0.0068 0.8451 0.8921
0.0019 26.9913 776 0.0069 0.8462 0.9029
0.002 28.0 805 0.0071 0.8522 0.9145
0.0022 28.9739 833 0.0077 0.8395 0.9304
0.0019 29.9826 862 0.0069 0.8567 0.9167
0.0027 30.9913 891 0.0073 0.8478 0.8957
0.0026 32.0 920 0.0069 0.8575 0.8994
0.0017 32.9739 948 0.0065 0.8602 0.9098
0.0023 33.9826 977 0.0071 0.8517 0.8915
0.0019 34.9913 1006 0.0069 0.8629 0.9200
0.0016 36.0 1035 0.0064 0.8684 0.9213
0.0017 36.9739 1063 0.0068 0.8612 0.9147
0.002 37.9826 1092 0.0074 0.8608 0.9295
0.0014 38.9913 1121 0.0069 0.8660 0.9222
0.0014 40.0 1150 0.0075 0.8624 0.9173
0.0013 40.9739 1178 0.0076 0.8560 0.9105
0.0013 41.9826 1207 0.0076 0.8640 0.9193
0.0013 42.9913 1236 0.0076 0.8625 0.9078
0.0012 44.0 1265 0.0078 0.8647 0.9152
0.0013 44.9739 1293 0.0078 0.8652 0.9176
0.0012 45.9826 1322 0.0078 0.8638 0.9209
0.0012 46.9913 1351 0.0081 0.8639 0.9186
0.001 48.0 1380 0.0079 0.8641 0.9181
0.0011 48.6957 1400 0.0082 0.8638 0.9180

Framework versions

  • Transformers 4.42.3
  • Pytorch 2.1.2
  • Datasets 2.20.0
  • Tokenizers 0.19.1