mobilevit-x-small_alpha0.5_temp5.0

This model is a fine-tuned version of c14kevincardenas/ClimBEiTv2 on the c14kevincardenas/beta_caller_284_person_crop_seq_withlimb_2 dataset. It achieves the following results on the evaluation set:

• Loss: 0.8056
• Accuracy: 0.6670

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
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
1.1992	1.0	90	1.3813	0.3053
1.1608	2.0	180	1.3167	0.3775
1.0525	3.0	270	1.1716	0.4704
0.9202	4.0	360	1.0438	0.5296
0.8105	5.0	450	0.9463	0.5860
0.7675	6.0	540	0.9055	0.5949
0.7178	7.0	630	0.8651	0.6294
0.6966	8.0	720	0.8816	0.6206
0.6529	9.0	810	0.8390	0.6403
0.6468	10.0	900	0.8408	0.6393
0.6115	11.0	990	0.8196	0.6561
0.6059	12.0	1080	0.8353	0.6551
0.5808	13.0	1170	0.8468	0.6374
0.5506	14.0	1260	0.8270	0.6453
0.5535	15.0	1350	0.8425	0.6403
0.5444	16.0	1440	0.8188	0.6542
0.5409	17.0	1530	0.8202	0.6502
0.5296	18.0	1620	0.8193	0.6591
0.5153	19.0	1710	0.8056	0.6670
0.5239	20.0	1800	0.8194	0.6591

Framework versions

• Transformers 4.45.2
• Pytorch 2.5.0+cu124
• Datasets 3.0.1
• Tokenizers 0.20.1