bart-large-cnn-samsum-rescom-finetuned-resume-summarizer-10-epoch-tweak-lr-8-10-1

This model is a fine-tuned version of Ameer05/model-token-repo on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 1.4855
• Rouge1: 58.3832
• Rouge2: 49.9973
• Rougel: 55.3055
• Rougelsum: 57.7139

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: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 10
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum
No log	0.91	5	2.0183	52.3098	45.5304	49.2759	51.7456
No log	1.91	10	1.6564	61.815	53.9035	58.4243	60.784
No log	2.91	15	1.5330	61.3032	54.12	58.9152	60.7178
No log	3.91	20	1.4539	63.3012	56.2987	61.0907	62.5217
1.5646	4.91	25	1.4578	62.4815	55.1453	60.3921	61.6067
1.5646	5.91	30	1.4284	61.5347	54.1271	58.8474	60.5427
1.5646	6.91	35	1.4467	61.5081	53.8512	59.2782	60.6928
1.5646	7.91	40	1.4653	59.5349	51.8208	56.5996	58.8211
0.6692	8.91	45	1.4740	57.2917	49.5416	54.8409	56.6276
0.6692	9.91	50	1.4855	58.3832	49.9973	55.3055	57.7139

Framework versions

• Transformers 4.15.0
• Pytorch 1.9.1
• Datasets 1.18.4
• Tokenizers 0.10.3