Edit model card

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

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

  • Loss: 0.4463
  • Accuracy: 0.8551
  • Precision: 0.8509
  • Recall: 0.8551
  • F1: 0.8514

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
1.7326 0.99 62 1.6140 0.4758 0.4530 0.4758 0.4312
1.1706 2.0 125 1.0827 0.6218 0.6294 0.6218 0.5983
0.9046 2.99 187 0.7418 0.7542 0.7566 0.7542 0.7351
0.7305 4.0 250 0.6540 0.7880 0.7823 0.7880 0.7789
0.6378 4.99 312 0.5903 0.8089 0.8054 0.8089 0.8047
0.6447 6.0 375 0.5915 0.7954 0.8041 0.7954 0.7865
0.6228 6.99 437 0.5513 0.8162 0.8201 0.8162 0.8164
0.5758 8.0 500 0.5553 0.8078 0.8094 0.8078 0.8033
0.5831 8.99 562 0.5207 0.8191 0.8246 0.8191 0.8162
0.537 10.0 625 0.4981 0.8286 0.8233 0.8286 0.8222
0.5322 10.99 687 0.4830 0.8337 0.8340 0.8337 0.8332
0.5171 12.0 750 0.4931 0.8253 0.8258 0.8253 0.8233
0.5092 12.99 812 0.4891 0.8360 0.8360 0.8360 0.8325
0.5245 14.0 875 0.4585 0.8450 0.8452 0.8450 0.8431
0.4585 14.99 937 0.4682 0.8422 0.8407 0.8422 0.8407
0.455 16.0 1000 0.4659 0.8388 0.8370 0.8388 0.8357
0.4175 16.99 1062 0.4633 0.8382 0.8363 0.8382 0.8351
0.4574 18.0 1125 0.4479 0.8450 0.8435 0.8450 0.8428
0.4593 18.99 1187 0.4577 0.8439 0.8446 0.8439 0.8430
0.4423 20.0 1250 0.4589 0.8461 0.8426 0.8461 0.8413
0.4141 20.99 1312 0.4732 0.8326 0.8339 0.8326 0.8299
0.4534 22.0 1375 0.4477 0.8461 0.8422 0.8461 0.8433
0.4011 22.99 1437 0.4614 0.8399 0.8403 0.8399 0.8390
0.4162 24.0 1500 0.4576 0.8450 0.8443 0.8450 0.8437
0.4291 24.99 1562 0.4609 0.8472 0.8441 0.8472 0.8439
0.3698 26.0 1625 0.4469 0.8506 0.8484 0.8506 0.8482
0.3957 26.99 1687 0.4488 0.8478 0.8464 0.8478 0.8464
0.4053 28.0 1750 0.4463 0.8551 0.8509 0.8551 0.8514
0.377 28.99 1812 0.4429 0.8540 0.8504 0.8540 0.8508
0.381 29.76 1860 0.4423 0.8517 0.8483 0.8517 0.8489

Framework versions

  • Transformers 4.37.2
  • Pytorch 2.3.0
  • Datasets 2.19.1
  • Tokenizers 0.15.1
Downloads last month
4
Safetensors
Model size
4.85M 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-femto-1k-224-finetuned-galaxy10-decals

Finetuned
(3)
this model