bertin-roberta-base-spanish-finetuned-recores3

This model is a fine-tuned version of bertin-project/bertin-roberta-base-spanish on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 6.0975
• Accuracy: 0.3884

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: 2
• eval_batch_size: 2
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 3000
• num_epochs: 25

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
1.6095	1.0	524	1.6094	0.2342
1.607	2.0	1048	1.5612	0.3058
1.4059	3.0	1572	1.6292	0.3361
0.7047	4.0	2096	2.5111	0.4132
0.2671	5.0	2620	3.2399	0.3499
0.1065	6.0	3144	5.1217	0.3444
0.0397	7.0	3668	4.3270	0.3691
0.0162	8.0	4192	5.1796	0.3719
0.0096	9.0	4716	5.2161	0.3994
0.0118	10.0	5240	4.9225	0.3719
0.0015	11.0	5764	5.0544	0.3829
0.0091	12.0	6288	5.7731	0.3884
0.0052	13.0	6812	4.1606	0.3939
0.0138	14.0	7336	6.2725	0.3857
0.0027	15.0	7860	6.2274	0.3857
0.0003	16.0	8384	6.0935	0.4022
0.0002	17.0	8908	5.7650	0.3994
0.0	18.0	9432	6.3595	0.4215
0.0	19.0	9956	5.8934	0.3747
0.0001	20.0	10480	6.0571	0.3884
0.0	21.0	11004	6.0718	0.3884
0.0	22.0	11528	6.0844	0.3884
0.0	23.0	12052	6.0930	0.3884
0.0	24.0	12576	6.0966	0.3884
0.0	25.0	13100	6.0975	0.3884

Framework versions

• Transformers 4.19.2
• Pytorch 1.11.0+cu113
• Datasets 2.2.2
• Tokenizers 0.12.1