final-ft__roberta-base-bne__70k-ultrasounds

This model is a fine-tuned version of PlanTL-GOB-ES/roberta-base-bne on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.6164

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: 64
• eval_batch_size: 64
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 256
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 20
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
No log	1.0	259	1.1165
No log	2.0	518	0.9244
No log	3.0	777	0.8153
1.0516	4.0	1036	0.7842
1.0516	5.0	1295	0.7262
1.0516	6.0	1554	0.7235
1.0516	7.0	1813	0.6960
0.7109	8.0	2072	0.6787
0.7109	9.0	2331	0.6799
0.7109	10.0	2590	0.6718
0.7109	11.0	2849	0.6488
0.6385	12.0	3108	0.6426
0.6385	13.0	3367	0.6415
0.6385	14.0	3626	0.6232
0.6385	15.0	3885	0.6329
0.6042	16.0	4144	0.6036
0.6042	17.0	4403	0.6161
0.6042	18.0	4662	0.6193
0.6042	19.0	4921	0.6183
0.587	20.0	5180	0.6164

Framework versions

• Transformers 4.40.1
• Pytorch 2.2.1+cu121
• Datasets 2.19.0
• Tokenizers 0.19.1