LLM_Teached_Pegasus_100k_FS

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

• Loss: 1.4433
• Rouge1: 0.4961
• Rouge2: 0.2476
• Rougel: 0.4155
• Rougelsum: 0.4154
• Gen Len: 25.8629
• Precision: 0.9136
• Recall: 0.914
• F1: 0.9137

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
• gradient_accumulation_steps: 6
• total_train_batch_size: 96
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 20
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	F1	Gen Len	Validation Loss	Precision	Recall	Rouge1	Rouge2	Rougel	Rougelsum
1.781	2.0	1388	0.9088	26.8891	1.5797	0.908	0.91	0.4708	0.2219	0.3892	0.389
1.6618	3.0	2083	0.91	26.7282	1.5411	0.9094	0.9111	0.4776	0.2303	0.3977	0.3973
1.626	4.0	2776	0.911	26.7596	1.5171	0.9102	0.9121	0.4834	0.2345	0.402	0.402
1.5918	5.0	3471	0.9112	26.6476	1.5001	0.9106	0.9122	0.4853	0.2365	0.4045	0.4045
1.5586	6.0	4164	0.9116	26.7778	1.4880	0.9108	0.9127	0.4875	0.2373	0.4063	0.4063
1.5375	7.0	4858	0.912	26.3991	1.4768	0.9116	0.9128	0.4898	0.24	0.4083	0.4083
1.5146	8.0	5553	0.9126	26.156	1.4686	0.9123	0.9133	0.4907	0.241	0.4088	0.4089
1.5006	9.0	6247	0.9127	26.2629	1.4636	0.9122	0.9135	0.4914	0.2419	0.4097	0.4099
1.49	10.0	6942	0.9127	26.0273	1.4580	0.9125	0.9133	0.4911	0.2429	0.4109	0.411
1.4749	11.0	7636	0.9131	26.2304	1.4546	0.9127	0.9138	0.4932	0.244	0.4121	0.4123
1.4661	12.0	8331	0.9132	25.8778	1.4514	0.9133	0.9136	0.4937	0.2448	0.4126	0.4127
1.4575	13.0	9025	0.9133	26.1151	1.4499	0.913	0.914	0.4947	0.2453	0.4139	0.414
1.4511	14.0	9720	0.9133	26.0287	1.4478	0.9131	0.9138	0.4939	0.2451	0.4133	0.4134
1.4519	15.0	10414	0.9133	25.9078	1.4471	0.9132	0.9137	0.4938	0.2451	0.4134	0.4134
1.4439	16.0	11104	1.4474	0.4942	0.2456	0.4133	0.4134	26.0345	0.9131	0.9139	0.9133
1.4441	17.0	11799	1.4447	0.4945	0.2457	0.4139	0.414	25.9391	0.9133	0.9138	0.9134
1.444	18.0	12493	1.4446	0.4957	0.2473	0.415	0.4151	26.0107	0.9133	0.9141	0.9135
1.4375	19.0	13188	1.4433	0.4961	0.2473	0.4153	0.4153	25.8869	0.9136	0.914	0.9136
1.4361	20.0	13880	1.4433	0.4961	0.2476	0.4155	0.4154	25.8629	0.9136	0.914	0.9137

Framework versions

• Transformers 4.36.0
• Pytorch 2.0.1+cu117
• Datasets 2.14.5
• Tokenizers 0.15.0