gpt2_bce_farshad_half_2_2

This model is a fine-tuned version of gpt2 on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.0760
• 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
1.3519	5.8501	50	0.7420	0.6497	0.7214
0.3022	11.7002	100	0.1212	0.9582	0.9602
0.0868	17.5503	150	0.0667	0.9782	0.9790
0.0509	23.4004	200	0.0564	0.9832	0.9837
0.0314	29.2505	250	0.0515	0.9861	0.9865
0.0165	35.1005	300	0.0460	0.9887	0.9891
0.009	40.9506	350	0.0563	0.9884	0.9888
0.0062	46.8007	400	0.0591	0.9884	0.9888
0.004	52.6508	450	0.0564	0.9893	0.9897
0.003	58.5009	500	0.0606	0.9887	0.9891
0.0033	64.3510	550	0.0558	0.9884	0.9888
0.0019	70.2011	600	0.0643	0.9898	0.9902
0.002	76.0512	650	0.0670	0.9884	0.9887
0.0015	81.9013	700	0.0603	0.9887	0.9891
0.0016	87.7514	750	0.0612	0.9898	0.9902
0.0014	93.6015	800	0.0760	0.9896	0.9899

Framework versions

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