mt5-small-finetuned-DEPlain

This model is a fine-tuned version of google/mt5-small on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 1.6826
• Rouge1: 55.2811
• Rouge2: 33.4022
• Rougel: 49.0555
• Rougelsum: 49.8535
• Gen Len: 16.3063

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: 8
• eval_batch_size: 8
• 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	Gen Len
6.1176	1.0	1333	2.2908	44.5349	24.6346	39.4206	40.1782	14.1202
2.9848	2.0	2666	2.0502	50.238	29.4215	44.3707	45.2709	15.4249
2.7551	3.0	3999	1.9383	52.4619	30.9025	46.0413	47.0121	15.814
2.5148	4.0	5332	1.8737	53.5574	31.7729	47.2411	48.1681	15.9764
2.3973	5.0	6665	1.8404	54.3931	32.4048	47.8747	48.7838	16.1422
2.3364	6.0	7998	1.7939	54.1979	32.253	47.8117	48.6785	16.0699
2.2807	7.0	9331	1.7804	54.3636	32.486	48.0067	48.8565	16.1324
2.2326	8.0	10664	1.7678	54.7743	32.7932	48.3657	49.1933	16.1608
2.1839	9.0	11997	1.7436	54.8052	32.8554	48.4859	49.3042	16.2047
2.1508	10.0	13330	1.7309	54.8064	32.8648	48.4888	49.3149	16.1779
2.1245	11.0	14663	1.7251	55.0598	33.1609	48.7331	49.6079	16.2705
2.1003	12.0	15996	1.7104	54.9449	33.1058	48.7477	49.5681	16.2185
2.0486	13.0	17329	1.6998	55.2225	33.3383	48.9821	49.8075	16.2494
2.0494	14.0	18662	1.6966	55.1758	33.3602	48.9134	49.7473	16.2648
2.0307	15.0	19995	1.6912	55.2276	33.3542	49.0322	49.8302	16.2721
2.0296	16.0	21328	1.6845	55.153	33.289	48.8609	49.7004	16.2754
2.01	17.0	22661	1.6842	55.3664	33.4755	49.1395	49.9518	16.3168
1.989	18.0	23994	1.6836	55.2333	33.3763	49.0259	49.8265	16.2794
2.0067	19.0	25327	1.6829	55.3122	33.4023	49.1034	49.8989	16.3022
2.0174	20.0	26660	1.6826	55.2811	33.4022	49.0555	49.8535	16.3063

Framework versions

• Transformers 4.36.2
• Pytorch 2.1.2+cu121
• Datasets 2.16.1
• Tokenizers 0.15.0