beit-base-patch16-224_alpha0.7_temp5.0_t3

This model is a fine-tuned version of c14kevincardenas/ClimBEiT-t3 on the c14kevincardenas/beta_caller_284_person_crop_seq_withlimb dataset. It achieves the following results on the evaluation set:

• Loss: 0.7231
• Accuracy: 0.8080

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: 32
• eval_batch_size: 32
• 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
0.5663	1.0	164	1.4230	0.2549
0.5186	2.0	328	1.2953	0.4024
0.3841	3.0	492	0.8618	0.7289
0.3209	4.0	656	0.8259	0.7505
0.2829	5.0	820	0.7344	0.8037
0.2556	6.0	984	0.7868	0.7939
0.2507	7.0	1148	0.7231	0.8080
0.2438	8.0	1312	0.7640	0.7842
0.2438	9.0	1476	0.7505	0.8048
0.2297	10.0	1640	0.7407	0.8178
0.227	11.0	1804	0.7734	0.7993
0.2303	12.0	1968	0.7463	0.7950
0.2221	13.0	2132	0.7577	0.7928
0.2256	14.0	2296	0.7403	0.8124
0.2267	15.0	2460	0.7455	0.8059
0.2222	16.0	2624	0.7406	0.8069
0.2268	17.0	2788	0.7435	0.8102
0.2168	18.0	2952	0.7401	0.8059
0.2224	19.0	3116	0.7402	0.8048
0.217	20.0	3280	0.7440	0.8037

Framework versions

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