turkish-gpt2-medium-finetuned-1gb-cX-corpus

This model is a fine-tuned version of ytu-ce-cosmos/turkish-gpt2-medium on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 2.5434

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: 4
• eval_batch_size: 8
• seed: 42
• distributed_type: multi-GPU
• num_devices: 4
• gradient_accumulation_steps: 16
• total_train_batch_size: 256
• total_eval_batch_size: 32
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: cosine
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 1.0

Training results

Training Loss	Epoch	Step	Validation Loss
2.7299	0.0573	200	2.6433
2.6926	0.1147	400	2.6090
2.6536	0.1720	600	2.5891
2.6492	0.2293	800	2.5791
2.627	0.2866	1000	2.5694
2.6162	0.3440	1200	2.5619
2.6192	0.4013	1400	2.5568
2.6585	0.4586	1600	2.5528
2.6232	0.5159	1800	2.5499
2.6252	0.5733	2000	2.5472
2.6295	0.6306	2200	2.5454
2.5898	0.6879	2400	2.5444
2.6553	0.7453	2600	2.5439
2.611	0.8026	2800	2.5435
2.6149	0.8599	3000	2.5435
2.5969	0.9172	3200	2.5434
2.6132	0.9746	3400	2.5434

Framework versions

• Transformers 4.50.0
• Pytorch 2.5.1+cu124
• Datasets 3.6.0
• Tokenizers 0.21.1