xlm-roberta-base_lr5e-06_seed42_basic_original_kin-amh-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: 0.0312
• Spearman Corr: 0.7479

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-06
• 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	1.75	200	0.0278	0.6150
0.0637	3.51	400	0.0254	0.6786
0.0312	5.26	600	0.0284	0.7074
0.0249	7.02	800	0.0320	0.7195
0.0199	8.77	1000	0.0344	0.7285
0.0189	10.53	1200	0.0303	0.7318
0.0167	12.28	1400	0.0299	0.7356
0.0154	14.04	1600	0.0319	0.7411
0.0154	15.79	1800	0.0304	0.7382
0.0141	17.54	2000	0.0327	0.7460
0.0134	19.3	2200	0.0305	0.7507
0.0123	21.05	2400	0.0312	0.7479

Framework versions

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