c14kevincardenas/beit-base-patch16-384_alpha0.7_temp5.0_t2

beit-base-patch16-384_alpha0.7_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: 0.5012
• Accuracy: 0.8399

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
1.098	1.0	90	1.3922	0.3063
0.8736	2.0	180	0.9673	0.6136
0.5416	3.0	270	0.6812	0.7599
0.3877	4.0	360	0.5916	0.7984
0.2839	5.0	450	0.5838	0.8063
0.2368	6.0	540	0.5499	0.8182
0.2149	7.0	630	0.6019	0.7984
0.1883	8.0	720	0.5797	0.8053
0.1724	9.0	810	0.5546	0.8251
0.1587	10.0	900	0.5658	0.8231
0.1592	11.0	990	0.5165	0.8231
0.1455	12.0	1080	0.5143	0.8300
0.1438	13.0	1170	0.5300	0.8330
0.1303	14.0	1260	0.5376	0.8271
0.1306	15.0	1350	0.5235	0.8281
0.1274	16.0	1440	0.5012	0.8399
0.121	17.0	1530	0.5207	0.8370
0.1197	18.0	1620	0.5180	0.8370
0.1189	19.0	1710	0.5124	0.8409
0.1157	20.0	1800	0.5116	0.8419

Framework versions

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