t5_es_farshad_half_4_2

This model is a fine-tuned version of google-t5/t5-base on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.0615
• Accuracy: 0.9896
• F1: 0.9899

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: 0.0001
• train_batch_size: 64
• eval_batch_size: 64
• seed: 42
• gradient_accumulation_steps: 64
• total_train_batch_size: 4096
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 1000
• num_epochs: 100

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1
0.6971	5.8501	50	0.6649	0.6589	0.6963
0.6328	11.7002	100	0.4862	0.8385	0.8422
0.2936	17.5503	150	0.1150	0.9626	0.9632
0.0908	23.4004	200	0.0712	0.9771	0.9776
0.0517	29.2505	250	0.0537	0.9846	0.9851
0.0342	35.1005	300	0.0500	0.9864	0.9867
0.0234	40.9506	350	0.0483	0.9884	0.9887
0.0166	46.8007	400	0.0522	0.9864	0.9867
0.0128	52.6508	450	0.0553	0.9869	0.9873
0.0099	58.5009	500	0.0559	0.9884	0.9887
0.0077	64.3510	550	0.0450	0.9901	0.9905
0.0061	70.2011	600	0.0477	0.9904	0.9907
0.0054	76.0512	650	0.0628	0.9867	0.9870
0.004	81.9013	700	0.0533	0.9896	0.9899
0.0039	87.7514	750	0.0445	0.9919	0.9921
0.0027	93.6015	800	0.0615	0.9896	0.9899

Framework versions

• Transformers 4.40.0
• Pytorch 2.4.1+cu121
• Datasets 3.1.0
• Tokenizers 0.19.1