train_3

This model was trained from scratch on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.1158

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: 5e-06
• train_batch_size: 56
• eval_batch_size: 56
• seed: 42
• distributed_type: multi-GPU
• num_devices: 2
• total_train_batch_size: 112
• total_eval_batch_size: 112
• optimizer: Use OptimizerNames.ADAMW_APEX_FUSED with betas=(0.826646043090655,0.991636944120939) and epsilon=3.4341677539323e-07 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 10000
• num_epochs: 200
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
0.0261	1.0	37441	0.1162
0.0252	2.0	74882	0.1142
0.025	3.0	112323	0.1154
0.0244	4.0	149764	0.1166
0.0244	5.0	187205	0.1150
0.0238	6.0	224646	0.1171
0.0238	7.0	262087	0.1179
0.0251	8.0	299528	0.1145
0.0251	9.0	336969	0.1147
0.0247	10.0	374410	0.1147
0.0247	11.0	411851	0.1142
0.0243	12.0	449292	0.1151
0.0243	13.0	486733	0.1150
0.0245	14.0	524174	0.1144
0.0241	15.0	561615	0.1155
0.0235	16.0	599056	0.1155
0.024	17.0	636497	0.1158

Framework versions

• Transformers 4.49.0.dev0
• Pytorch 2.6.0+cu126
• Datasets 3.3.0
• Tokenizers 0.21.0