xlmr-large-vinli

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

• Loss: 1.7569
• Accuracy: 0.8185
• F1: 0.8184
• Precision: 0.8188
• Recall: 0.8185

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: 1e-05
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1	Precision	Recall
0.6648	0.4374	1000	0.6488	0.7509	0.7506	0.7524	0.7509
0.6183	0.8749	2000	0.6322	0.7699	0.7699	0.7867	0.7699
0.528	1.3123	3000	0.6445	0.7805	0.7796	0.7831	0.7805
0.4812	1.7498	4000	0.6283	0.7995	0.7995	0.7995	0.7995
0.3723	2.1872	5000	0.8542	0.7968	0.7954	0.8031	0.7968
0.4176	2.6247	6000	0.8994	0.7990	0.7991	0.8007	0.7990
0.281	3.0621	7000	1.0532	0.8034	0.8031	0.8049	0.8034
0.3376	3.4996	8000	1.0045	0.8079	0.8079	0.8083	0.8079
0.318	3.9370	9000	1.1812	0.7946	0.7948	0.7973	0.7946
0.2388	4.3745	10000	1.2944	0.8043	0.8045	0.8056	0.8043
0.193	4.8119	11000	1.2180	0.8105	0.8103	0.8108	0.8105
0.1639	5.2493	12000	1.3971	0.8065	0.8063	0.8073	0.8065
0.1557	5.6868	13000	1.3454	0.8021	0.8023	0.8033	0.8021
0.1229	6.1242	14000	1.4378	0.8123	0.8123	0.8128	0.8123
0.0371	6.5617	15000	1.5330	0.8092	0.8094	0.8123	0.8092
0.1189	6.9991	16000	1.4320	0.8105	0.8100	0.8110	0.8105
0.047	7.4366	17000	1.6426	0.8127	0.8126	0.8128	0.8127
0.0511	7.8740	18000	1.7100	0.8123	0.8122	0.8124	0.8123
0.0566	8.3115	19000	1.6801	0.8052	0.8055	0.8083	0.8052
0.0004	8.7489	20000	1.6785	0.8180	0.8182	0.8191	0.8180
0.0213	9.1864	21000	1.7024	0.8198	0.8200	0.8211	0.8198
0.0107	9.6238	22000	1.7569	0.8185	0.8184	0.8188	0.8185

Framework versions

• Transformers 4.44.2
• Pytorch 2.2.1
• Datasets 2.21.0
• Tokenizers 0.19.1