c14kevincardenas/mobilevit-small_alpha0.5_temp5.0_t2

mobilevit-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: 0.8370
• Accuracy: 0.6769

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
1.2117	1.0	90	1.4099	0.2451
1.1611	2.0	180	1.3374	0.2945
1.0726	3.0	270	1.2361	0.3794
0.9066	4.0	360	1.0439	0.5168
0.8097	5.0	450	0.9753	0.5800
0.7541	6.0	540	0.9093	0.6146
0.7054	7.0	630	0.8776	0.6344
0.6567	8.0	720	0.9050	0.6136
0.6028	9.0	810	0.8489	0.6670
0.585	10.0	900	0.8526	0.6453
0.5614	11.0	990	0.8427	0.6690
0.5282	12.0	1080	0.8506	0.6611
0.4993	13.0	1170	0.8682	0.6492
0.4814	14.0	1260	0.8370	0.6769
0.4798	15.0	1350	0.8531	0.6640
0.4554	16.0	1440	0.8401	0.6709
0.4514	17.0	1530	0.8405	0.6690
0.4313	18.0	1620	0.8409	0.6729
0.4352	19.0	1710	0.8418	0.6749
0.4362	20.0	1800	0.8477	0.6709

Framework versions

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