xlmr-en-zh-no_shuffled-orig-test1000

This model is a fine-tuned version of xlm-roberta-base on the wmt20_mlqe_task1 dataset. It achieves the following results on the evaluation set:

• Loss: 0.5862
• R Squared: -0.1664
• Mae: 0.5991
• Pearson R: 0.4148

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: 16
• eval_batch_size: 16
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 5

Training results

Training Loss	Epoch	Step	Validation Loss	R Squared	Mae	Pearson R
No log	1.0	438	0.4458	0.1130	0.5377	0.3977
0.7823	2.0	876	0.4252	0.1539	0.5123	0.4517
0.6317	3.0	1314	0.4948	0.0155	0.5423	0.4304
0.4449	4.0	1752	0.5149	-0.0245	0.5577	0.4247
0.3139	5.0	2190	0.5862	-0.1664	0.5991	0.4148

Framework versions

• Transformers 4.34.1
• Pytorch 2.0.1+cu117
• Datasets 2.14.6
• Tokenizers 0.14.1