H2-keywordextractor-finetuned-scope-summarization

This model is a fine-tuned version of transformer3/H2-keywordextractor on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.2073
• Rouge1: 13.0222
• Rouge2: 10.4851
• Rougel: 13.0872
• Rougelsum: 13.1095

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: 5.6e-05
• train_batch_size: 15
• eval_batch_size: 15
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum
0.8852	1.0	23	0.3103	10.3278	6.2988	10.3528	10.3293
0.2901	2.0	46	0.2825	10.8308	7.5214	10.8428	10.8103
0.2625	3.0	69	0.2711	12.0182	8.6415	12.0115	12.0537
0.2453	4.0	92	0.2550	12.9535	9.6936	12.9952	13.0384
0.2353	5.0	115	0.2464	11.2808	7.8603	11.3196	11.281
0.2338	6.0	138	0.2389	12.6604	9.6355	12.6519	12.6377
0.2183	7.0	161	0.2307	13.2591	10.6628	13.2399	13.2554
0.2143	8.0	184	0.2252	13.537	11.1632	13.5668	13.5957
0.2055	9.0	207	0.2206	13.7032	11.6575	13.7226	13.774
0.2022	10.0	230	0.2158	13.7727	11.5365	13.7404	13.8018
0.1961	11.0	253	0.2166	13.4062	11.2919	13.4698	13.4854
0.2018	12.0	276	0.2116	13.8406	11.852	13.8309	13.8995
0.1946	13.0	299	0.2131	12.5757	9.5775	12.5738	12.6535
0.1943	14.0	322	0.2142	11.617	9.0291	11.5311	11.7201
0.2068	15.0	345	0.2080	12.9136	10.2865	12.9659	12.9787
0.2051	16.0	368	0.2041	13.6492	11.6388	13.6506	13.7041
0.1887	17.0	391	0.2119	11.4317	8.2482	11.386	11.4313
0.1886	18.0	414	0.2097	13.0287	10.6547	13.0829	13.118
0.1887	19.0	437	0.2079	13.0073	10.5381	13.0514	13.1089
0.186	20.0	460	0.2073	13.0222	10.4851	13.0872	13.1095

Framework versions

• Transformers 4.39.3
• Pytorch 2.2.2+cu121
• Datasets 2.18.0
• Tokenizers 0.15.2