hiera-finetuned-stroke-multi-ultrasound-combined

This model was trained from scratch on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.3436
• Accuracy: 0.8661
• F1: 0.8650
• Precision: 0.8650
• Recall: 0.8661

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: 2e-05
• train_batch_size: 16
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 64
• optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: cosine_with_restarts
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 12
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1	Precision	Recall
0.5712	1.7964	100	0.4956	0.7923	0.7933	0.7973	0.7923
0.4495	3.5792	200	0.4193	0.8309	0.8315	0.8325	0.8309
0.4266	5.3620	300	0.4364	0.8150	0.8091	0.8221	0.8150
0.3375	7.1448	400	0.3775	0.8513	0.8473	0.8535	0.8513
0.2979	8.9412	500	0.3360	0.8627	0.8617	0.8614	0.8627
0.2703	10.7240	600	0.3436	0.8661	0.8650	0.8650	0.8661

Framework versions

• Transformers 4.53.1
• Pytorch 2.7.1+cu126
• Datasets 3.6.0
• Tokenizers 0.21.2