bart-large-cnn-YT-transcript-sum

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

• Loss: 1.4849
• Rouge1: 48.0422
• Rouge2: 22.8938
• Rougel: 34.0775
• Rougelsum: 44.7056
• Gen Len: 108.8009

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: 2
• eval_batch_size: 2
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 15

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum	Gen Len
No log	1.0	432	1.5362	49.0506	22.9422	35.5667	45.7219	88.0602
1.5312	2.0	864	1.4849	48.0422	22.8938	34.0775	44.7056	108.8009
0.9026	3.0	1296	1.5761	50.0558	23.9657	36.247	46.4508	96.0231
0.5642	4.0	1728	1.8304	50.6862	24.4638	36.3568	47.2607	93.1667
0.3629	5.0	2160	1.9355	51.2362	25.1077	37.772	47.4362	88.9583
0.2335	6.0	2592	2.1215	49.5831	23.4294	35.9861	45.9306	94.2917
0.1603	7.0	3024	2.2890	49.8716	23.4756	36.2617	46.2866	88.7639
0.1603	8.0	3456	2.3604	49.5627	23.6399	35.9596	45.7914	88.8333
0.1049	9.0	3888	2.5252	50.358	24.1986	36.5297	46.5519	90.5463
0.0744	10.0	4320	2.6694	50.46	24.1493	37.0205	46.8988	91.0139
0.049	11.0	4752	2.7840	50.8805	24.5482	36.5901	46.9176	90.8380
0.0312	12.0	5184	2.8330	50.4793	24.6444	37.2087	46.7151	86.9444
0.0156	13.0	5616	2.9540	50.3911	24.4843	36.8037	46.8691	94.9352
0.0083	14.0	6048	3.0214	51.0557	25.127	37.1368	47.3072	92.5787
0.0083	15.0	6480	3.0340	51.3998	25.5847	37.5635	47.7132	90.5602

Framework versions

• Transformers 4.33.2
• Pytorch 2.0.1+cu117
• Datasets 2.14.5
• Tokenizers 0.13.3