mobilevit-x-small_alpha0.7_temp3.0

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: 0.8319
• Accuracy: 0.6650

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
1.1391	1.0	90	1.3902	0.2895
1.0994	2.0	180	1.3219	0.3765
0.9868	3.0	270	1.1753	0.4664
0.8583	4.0	360	1.0554	0.5277
0.752	5.0	450	0.9602	0.5840
0.7175	6.0	540	0.9255	0.5761
0.6641	7.0	630	0.8845	0.6324
0.6509	8.0	720	0.9110	0.6028
0.6031	9.0	810	0.8721	0.6374
0.6017	10.0	900	0.8564	0.6522
0.5642	11.0	990	0.8444	0.6482
0.5584	12.0	1080	0.8647	0.6364
0.5325	13.0	1170	0.8772	0.6294
0.5035	14.0	1260	0.8566	0.6354
0.5084	15.0	1350	0.8670	0.6383
0.4972	16.0	1440	0.8465	0.6551
0.4891	17.0	1530	0.8490	0.6443
0.4841	18.0	1620	0.8504	0.6551
0.4713	19.0	1710	0.8319	0.6650
0.4845	20.0	1800	0.8481	0.6542

Framework versions

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