rhythm00/finetune_t5_small_gazeta_full

finetune_t5_small_gusev_full

This model is a fine-tuned version of cointegrated/rut5-small on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 1.7795

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: 0.0004
• train_batch_size: 24
• eval_batch_size: 24
• seed: 42
• gradient_accumulation_steps: 12
• total_train_batch_size: 288
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 100
• num_epochs: 8

Training results

Training Loss	Epoch	Step	Validation Loss
2.7388	0.61	150	2.2027
2.5562	1.22	300	2.0326
2.4982	1.83	450	1.9607
2.4324	2.44	600	1.9077
2.4015	3.05	750	1.8711
2.3623	3.65	900	1.8451
2.3282	4.26	1050	1.8304
2.3072	4.87	1200	1.8120
2.2878	5.48	1350	1.8007
2.2689	6.09	1500	1.7919
2.2814	6.7	1650	1.7863
2.2443	7.31	1800	1.7835
2.2665	7.92	1950	1.7795

Framework versions

• Transformers 4.36.2
• Pytorch 2.1.2+cu121
• Datasets 2.15.0
• Tokenizers 0.15.0