convnextv2-atto-1k-224-finetuned-galaxy10-decals

This model is a fine-tuned version of facebook/convnextv2-atto-1k-224 on the matthieulel/galaxy10_decals dataset. It achieves the following results on the evaluation set:

  • Loss: 0.4668
  • Accuracy: 0.8461
  • Precision: 0.8444
  • Recall: 0.8461
  • F1: 0.8442

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: 5e-05
  • train_batch_size: 64
  • eval_batch_size: 64
  • seed: 42
  • gradient_accumulation_steps: 4
  • total_train_batch_size: 256
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • lr_scheduler_warmup_ratio: 0.1
  • num_epochs: 30

Training results

Training Loss Epoch Step Validation Loss Accuracy Precision Recall F1
2.0062 0.99 62 1.8928 0.3450 0.3432 0.3450 0.2956
1.1323 2.0 125 1.0026 0.6590 0.6634 0.6590 0.6399
0.8977 2.99 187 0.7348 0.7486 0.7415 0.7486 0.7399
0.7119 4.0 250 0.6395 0.7892 0.7878 0.7892 0.7770
0.6393 4.99 312 0.5801 0.7971 0.7916 0.7971 0.7915
0.6463 6.0 375 0.5958 0.7976 0.8147 0.7976 0.7909
0.6197 6.99 437 0.5363 0.8151 0.8119 0.8151 0.8112
0.5779 8.0 500 0.5276 0.8207 0.8205 0.8207 0.8185
0.5841 8.99 562 0.5197 0.8185 0.8203 0.8185 0.8157
0.5597 10.0 625 0.5025 0.8253 0.8192 0.8253 0.8193
0.5437 10.99 687 0.4912 0.8309 0.8295 0.8309 0.8296
0.5242 12.0 750 0.5001 0.8275 0.8303 0.8275 0.8245
0.5029 12.99 812 0.5075 0.8241 0.8228 0.8241 0.8208
0.5396 14.0 875 0.4784 0.8393 0.8395 0.8393 0.8371
0.4746 14.99 937 0.4727 0.8331 0.8318 0.8331 0.8317
0.4786 16.0 1000 0.4856 0.8331 0.8308 0.8331 0.8300
0.4338 16.99 1062 0.4884 0.8337 0.8333 0.8337 0.8309
0.4772 18.0 1125 0.4618 0.8405 0.8370 0.8405 0.8377
0.4733 18.99 1187 0.4740 0.8393 0.8394 0.8393 0.8381
0.4475 20.0 1250 0.4678 0.8388 0.8349 0.8388 0.8345
0.4229 20.99 1312 0.4881 0.8331 0.8317 0.8331 0.8303
0.46 22.0 1375 0.4728 0.8410 0.8382 0.8410 0.8371
0.4298 22.99 1437 0.4642 0.8360 0.8348 0.8360 0.8345
0.4225 24.0 1500 0.4706 0.8371 0.8368 0.8371 0.8359
0.426 24.99 1562 0.4733 0.8399 0.8367 0.8399 0.8371
0.3839 26.0 1625 0.4682 0.8444 0.8423 0.8444 0.8422
0.4007 26.99 1687 0.4665 0.8382 0.8371 0.8382 0.8367
0.4245 28.0 1750 0.4695 0.8388 0.8357 0.8388 0.8358
0.3868 28.99 1812 0.4668 0.8461 0.8444 0.8461 0.8442
0.3933 29.76 1860 0.4657 0.8461 0.8442 0.8461 0.8440

Framework versions

  • Transformers 4.37.2
  • Pytorch 2.3.0
  • Datasets 2.19.1
  • Tokenizers 0.15.1
Downloads last month
12
Safetensors
Model size
3.39M params
Tensor type
F32
·
Inference Examples
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social visibility and check back later, or deploy to Inference Endpoints (dedicated) instead.

Model tree for matthieulel/convnextv2-atto-1k-224-finetuned-galaxy10-decals

Finetuned
(3)
this model