t5_es_farshad_half_4_4

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: 0.0424
• Accuracy: 0.9922
• F1: 0.9924

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
0.7459	5.8501	50	0.6868	0.5426	0.6423
0.6483	11.7002	100	0.5144	0.8518	0.8540
0.3069	17.5503	150	0.1038	0.9675	0.9681
0.0869	23.4004	200	0.0563	0.9820	0.9825
0.0496	29.2505	250	0.0440	0.9864	0.9868
0.0327	35.1005	300	0.0365	0.9887	0.9891
0.0226	40.9506	350	0.0333	0.9916	0.9919
0.0161	46.8007	400	0.0316	0.9925	0.9927
0.0125	52.6508	450	0.0311	0.9936	0.9938
0.0097	58.5009	500	0.0322	0.9933	0.9935
0.0076	64.3510	550	0.0366	0.9927	0.9930
0.0069	70.2011	600	0.0407	0.9919	0.9921
0.0055	76.0512	650	0.0342	0.9927	0.9930
0.0041	81.9013	700	0.0364	0.9936	0.9938
0.003	87.7514	750	0.0411	0.9933	0.9936
0.0026	93.6015	800	0.0424	0.9922	0.9924

Framework versions

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