xlm-roberta-base_original_esp-kin-eng_train

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

• Loss: 0.0301
• Spearman Corr: 0.7160

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: 32
• eval_batch_size: 128
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 64
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 30
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Spearman Corr
No log	1.63	200	0.0292	0.6299
0.0375	3.27	400	0.0290	0.6845
0.0212	4.9	600	0.0259	0.6877
0.0164	6.53	800	0.0287	0.7094
0.0127	8.16	1000	0.0285	0.7219
0.0127	9.8	1200	0.0318	0.7103
0.0102	11.43	1400	0.0299	0.7125
0.008	13.06	1600	0.0357	0.7183
0.0067	14.69	1800	0.0282	0.7207
0.0056	16.33	2000	0.0314	0.7154
0.0056	17.96	2200	0.0317	0.7145
0.0049	19.59	2400	0.0302	0.7111
0.0044	21.22	2600	0.0301	0.7160

Framework versions

• Transformers 4.37.2
• Pytorch 2.1.0+cu121
• Datasets 2.17.0
• Tokenizers 0.15.2