video_transcript_summary

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: 5.5472
• Rouge1: 0.3909
• Rouge2: 0.1328
• Rougel: 0.2802
• Rougelsum: 0.2801
• Gen Len: 69.8235

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: 2e-05
• train_batch_size: 16
• eval_batch_size: 16
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 100
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum	Gen Len
No log	1.0	5	2.6381	0.3163	0.0838	0.2135	0.2138	72.1176
No log	2.0	10	2.4739	0.345	0.0969	0.2372	0.2364	69.7647
No log	3.0	15	2.6011	0.3719	0.116	0.2619	0.2605	69.7059
No log	4.0	20	2.7476	0.3769	0.1156	0.2588	0.2572	72.9412
No log	5.0	25	2.9283	0.3812	0.1206	0.2645	0.2625	74.4706
No log	6.0	30	3.3089	0.3722	0.1194	0.2779	0.2769	70.1765
No log	7.0	35	3.6585	0.3802	0.1177	0.2718	0.2719	73.2941
No log	8.0	40	3.8924	0.3764	0.1056	0.2661	0.2651	70.8235
No log	9.0	45	4.3638	0.3677	0.1144	0.2653	0.2634	74.0
No log	10.0	50	4.5590	0.354	0.1069	0.2491	0.2477	71.0
No log	11.0	55	4.5746	0.4103	0.1359	0.2935	0.293	72.5882
No log	12.0	60	4.9055	0.3869	0.118	0.2705	0.2695	66.1176
No log	13.0	65	5.0987	0.3947	0.1292	0.2713	0.2716	74.7647
No log	14.0	70	5.3199	0.3814	0.1135	0.2839	0.2816	72.2353
No log	15.0	75	5.4462	0.4092	0.1265	0.2896	0.29	68.0
No log	16.0	80	5.4262	0.412	0.1282	0.293	0.2932	72.0
No log	17.0	85	5.4873	0.3776	0.1051	0.2637	0.2636	68.8824
No log	18.0	90	5.6043	0.3932	0.1106	0.2734	0.2727	67.9412
No log	19.0	95	5.7495	0.3694	0.1075	0.2699	0.2692	69.7647
No log	20.0	100	5.6518	0.3849	0.1109	0.2714	0.2709	65.5882
No log	21.0	105	5.4524	0.3762	0.1101	0.2598	0.2596	64.5294
No log	22.0	110	5.5472	0.3909	0.1328	0.2802	0.2801	69.8235

Framework versions

• Transformers 4.41.2
• Pytorch 2.3.1+cu118
• Datasets 2.20.0
• Tokenizers 0.19.1