b2b_cnn_retrain

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

• Loss: 8.0538
• Rouge2 Precision: 0.0033
• Rouge2 Recall: 0.0089
• Rouge2 Fmeasure: 0.0048

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: 16
• eval_batch_size: 16
• seed: 42
• gradient_accumulation_steps: 32
• total_train_batch_size: 512
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 5
• training_steps: 50
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge2 Precision	Rouge2 Recall	Rouge2 Fmeasure
0.587	5.0	5	8.3529	0.0	0.0	0.0
0.4646	10.0	10	8.1390	0.0033	0.003	0.0031
0.4335	15.0	15	8.1031	0.0	0.0	0.0
0.3966	20.0	20	8.1701	0.0	0.0	0.0
0.3476	25.0	25	8.2264	0.0	0.0	0.0
0.2928	30.0	30	8.0323	0.0029	0.017	0.0049
0.244	35.0	35	7.9815	0.0024	0.0057	0.0034
0.2059	40.0	40	7.9555	0.0035	0.0114	0.0053
0.1791	45.0	45	8.0112	0.0046	0.0114	0.0066
0.1637	50.0	50	8.0538	0.0033	0.0089	0.0048

Framework versions

• Transformers 4.41.1
• Pytorch 2.3.0+cu121
• Datasets 2.19.2
• Tokenizers 0.19.1