bert-base-multilingual-cased-finetuned-yiddish-experiment-2

This model is a fine-tuned version of bert-base-multilingual-cased on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.7107

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: 1e-05
• train_batch_size: 4
• eval_batch_size: 4
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 8
• optimizer: Use 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_steps: 200
• num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss
9.1219	0.4728	100	1.3974
1.0378	0.9456	200	0.8499
0.773	1.4161	300	0.7918
0.729	1.8889	400	0.7863
0.6948	2.3593	500	0.7451
0.6802	2.8322	600	0.7354
0.6525	3.3026	700	0.7210
0.6603	3.7754	800	0.7329
0.6293	4.2459	900	0.7321
0.6292	4.7187	1000	0.7441
0.6155	5.1891	1100	0.7107
0.6025	5.6619	1200	0.7421
0.5976	6.1324	1300	0.7268
0.5855	6.6052	1400	0.7327
0.5873	7.0757	1500	0.7172
0.5826	7.5485	1600	0.7205
0.5594	8.0189	1700	0.7591
0.5671	8.4917	1800	0.7320
0.569	8.9645	1900	0.7419
0.5558	9.4350	2000	0.7290
0.5696	9.9078	2100	0.7328

Framework versions

• Transformers 4.47.0
• Pytorch 2.5.1+cu121
• Datasets 3.1.0
• Tokenizers 0.21.0