intfloat-multilingual-e5-base-arabic-fp16

This model is a fine-tuned version of intfloat/multilingual-e5-base on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.4961
• Accuracy: 0.7986
• Precision: 0.7991
• Recall: 0.7986
• F1: 0.7988

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: 64
• eval_batch_size: 64
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 128
• optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.3
• num_epochs: 10
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision	Recall	F1
1.0686	0.3636	50	1.0146	0.5495	0.7252	0.5495	0.4582
0.9589	0.7273	100	0.8046	0.6777	0.7234	0.6777	0.6081
0.7431	1.0873	150	0.6238	0.7595	0.7565	0.7595	0.7530
0.6066	1.4509	200	0.5485	0.7945	0.7947	0.7945	0.7906
0.5558	1.8145	250	0.5530	0.7827	0.7860	0.7827	0.7837
0.5343	2.1745	300	0.5430	0.7973	0.8009	0.7973	0.7983
0.4965	2.5382	350	0.5178	0.7986	0.7993	0.7986	0.7988
0.5017	2.9018	400	0.4961	0.7986	0.7991	0.7986	0.7988
0.4525	3.2618	450	0.5441	0.7932	0.7991	0.7932	0.7950
0.4194	3.6255	500	0.5147	0.8027	0.8051	0.8027	0.8027
0.4353	3.9891	550	0.4918	0.8118	0.8109	0.8118	0.8110
0.3635	4.3491	600	0.5659	0.7977	0.8058	0.7977	0.7980
0.3529	4.7127	650	0.5493	0.8023	0.8066	0.8023	0.8029
0.3574	5.0727	700	0.5438	0.8023	0.8043	0.8023	0.8031

Framework versions

• Transformers 4.51.1
• Pytorch 2.6.0+cu124
• Datasets 3.5.0
• Tokenizers 0.21.1