xlm-roberta-base_hau_corr_2e-05

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: 0.0218
• Spearman Corr: 0.7816

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	0.95	200	0.0215	0.7799
No log	1.91	400	0.0226	0.7758
0.0008	2.86	600	0.0218	0.7788
0.0008	3.82	800	0.0213	0.7793
0.0008	4.77	1000	0.0213	0.7786
0.0008	5.73	1200	0.0224	0.7798
0.0007	6.68	1400	0.0215	0.7808
0.0007	7.64	1600	0.0213	0.7792
0.0007	8.59	1800	0.0208	0.7802
0.0007	9.55	2000	0.0221	0.7777
0.0006	10.5	2200	0.0210	0.7797
0.0006	11.46	2400	0.0207	0.7799
0.0006	12.41	2600	0.0213	0.7805
0.0006	13.37	2800	0.0219	0.7825
0.0006	14.32	3000	0.0209	0.7799
0.0006	15.27	3200	0.0214	0.7818
0.0005	16.23	3400	0.0210	0.7811
0.0005	17.18	3600	0.0216	0.7803
0.0005	18.14	3800	0.0215	0.7807
0.0005	19.09	4000	0.0210	0.7793
0.0005	20.05	4200	0.0208	0.7816
0.0005	21.0	4400	0.0218	0.7816

Framework versions

• Transformers 4.37.2
• Pytorch 2.2.0+cu121
• Datasets 2.17.0
• Tokenizers 0.15.2