amtibot_bart

This model is a fine-tuned version of facebook/bart-large-cnn on the None dataset. It achieves the following results on the evaluation set:

• Loss: 1.5905
• Rouge1: 0.4051
• Rouge2: 0.195
• Rougel: 0.3054
• Rougelsum: 0.3053
• Gen Len: 65.7532

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: 0.02
• train_batch_size: 4
• eval_batch_size: 4
• seed: 42
• gradient_accumulation_steps: 8
• total_train_batch_size: 32
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 4
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum	Gen Len
No log	0.9351	9	1.6594	0.4057	0.1833	0.3052	0.3048	67.9481
2.11	1.9740	19	1.6149	0.3938	0.192	0.3063	0.3058	64.8571
1.554	2.9091	28	1.5842	0.3956	0.1872	0.3039	0.3033	65.8182
1.3821	3.7403	36	1.5905	0.4051	0.195	0.3054	0.3053	65.7532

Framework versions

• PEFT 0.4.0
• Transformers 4.40.1
• Pytorch 2.1.2
• Datasets 2.18.0
• Tokenizers 0.19.1