DmitryYarov/aristotle-graph-logic-nlayer

aristotle_new_layer_md

This model is a fine-tuned version of ai-forever/rugpt3small_based_on_gpt2 on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 4.5495

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-05
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 32
• optimizer: Use OptimizerNames.ADAFACTOR and the args are: No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 500
• num_epochs: 30
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
9.9676	1.0	95	7.7187
7.3648	2.0	190	6.7524
6.3191	3.0	285	5.9377
5.4765	4.0	380	5.5640
5.0688	5.0	475	5.3898
4.9737	6.0	570	5.2278
4.8498	7.0	665	5.1026
4.7189	8.0	760	5.0769
4.635	9.0	855	4.8809
4.5526	10.0	950	4.9227
4.4848	11.0	1045	4.8871
4.4918	12.0	1140	4.7096
4.3303	13.0	1235	4.7373
4.3412	14.0	1330	4.7861
4.2659	15.0	1425	4.6738
4.2575	16.0	1520	4.6526
4.2401	17.0	1615	4.7348
4.1957	18.0	1710	4.5350
4.1738	19.0	1805	4.5723
4.1671	20.0	1900	4.5821
4.0669	21.0	1995	4.5495

Framework versions

• Transformers 4.48.3
• Pytorch 2.5.1+cu124
• Tokenizers 0.21.0