mt5-small-finetuned

This model is a fine-tuned version of google/mt5-small on the None dataset. It achieves the following results on the evaluation set:

• Loss: 2.0192
• Rouge1: 0.3780
• Rouge2: 0.1970
• Rougel: 0.3508
• Rougelsum: 0.3527

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: 12
• eval_batch_size: 12
• seed: 42
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• num_epochs: 60

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum
21.4055	1.0	12	13.7151	0.0189	0.0084	0.0147	0.0189
17.1792	2.0	24	11.5227	0.0189	0.0084	0.0147	0.0189
15.0485	3.0	36	9.5193	0.0	0.0	0.0	0.0
13.0405	4.0	48	6.8529	0.0102	0.0	0.0102	0.0102
11.7418	5.0	60	5.8151	0.0331	0.0084	0.0303	0.0335
9.659	6.0	72	5.6024	0.0344	0.0084	0.0344	0.0357
8.6025	7.0	84	4.7311	0.0273	0.0036	0.0269	0.0277
7.5747	8.0	96	3.8319	0.0510	0.0031	0.0483	0.0456
6.916	9.0	108	3.5873	0.0578	0.0	0.0540	0.0520
6.3394	10.0	120	3.4854	0.0788	0.0076	0.0792	0.0794
5.5822	11.0	132	3.2956	0.0752	0.0158	0.0694	0.0697
5.0731	12.0	144	3.0977	0.0524	0.0115	0.0470	0.0475
4.7234	13.0	156	2.9120	0.0331	0.0105	0.0279	0.0285
4.3512	14.0	168	2.7709	0.0527	0.0304	0.0375	0.0377
4.136	15.0	180	2.6770	0.0616	0.0331	0.0494	0.0495
3.8591	16.0	192	2.5894	0.1028	0.0473	0.0817	0.0826
3.6558	17.0	204	2.5183	0.1814	0.0828	0.1541	0.1532
3.4821	18.0	216	2.4590	0.1940	0.0838	0.1618	0.1621
3.3248	19.0	228	2.3901	0.2062	0.0856	0.1667	0.1676
3.194	20.0	240	2.3352	0.1971	0.0918	0.1672	0.1684
3.0883	21.0	252	2.2934	0.1971	0.0918	0.1672	0.1684
2.9907	22.0	264	2.2471	0.2039	0.0943	0.1660	0.1675
2.9249	23.0	276	2.2038	0.1904	0.0843	0.1515	0.1537
2.8418	24.0	288	2.1643	0.1995	0.0939	0.1686	0.1705
2.7667	25.0	300	2.1296	0.2233	0.1002	0.1882	0.1890
2.7157	26.0	312	2.1176	0.3513	0.1825	0.3422	0.3432
2.7058	27.0	324	2.0969	0.3525	0.1803	0.3444	0.3457
2.5703	28.0	336	2.0761	0.3507	0.1847	0.3395	0.3409
2.4907	29.0	348	2.0688	0.3379	0.1741	0.3281	0.3290
2.3974	30.0	360	2.0706	0.3520	0.1872	0.3391	0.3402
2.4584	31.0	372	2.0635	0.3465	0.1840	0.3332	0.3344
2.3775	32.0	384	2.0560	0.3525	0.1826	0.3390	0.3411
2.4014	33.0	396	2.0544	0.3585	0.1860	0.3456	0.3469
2.3388	34.0	408	2.0583	0.3607	0.1865	0.3483	0.3496
2.3288	35.0	420	2.0487	0.3551	0.1835	0.3368	0.3379
2.3233	36.0	432	2.0394	0.3569	0.1803	0.3313	0.3326
2.2882	37.0	444	2.0361	0.3585	0.1867	0.3422	0.3446
2.2109	38.0	456	2.0324	0.3565	0.1858	0.3413	0.3429
2.212	39.0	468	2.0327	0.3585	0.1867	0.3422	0.3446
2.2059	40.0	480	2.0310	0.3612	0.1849	0.3421	0.3436
2.1866	41.0	492	2.0352	0.3612	0.1849	0.3421	0.3436
2.2122	42.0	504	2.0369	0.3612	0.1849	0.3421	0.3436
2.1305	43.0	516	2.0351	0.3604	0.1863	0.3419	0.3443
2.1174	44.0	528	2.0358	0.3578	0.1864	0.3397	0.3413
2.0972	45.0	540	2.0356	0.3602	0.1881	0.3390	0.3405
2.1051	46.0	552	2.0325	0.3606	0.1861	0.3359	0.3376
2.0632	47.0	564	2.0329	0.3606	0.1861	0.3359	0.3376
2.0601	48.0	576	2.0301	0.3621	0.1857	0.3346	0.3364
2.0487	49.0	588	2.0301	0.3621	0.1857	0.3346	0.3364
2.0538	50.0	600	2.0314	0.3617	0.1876	0.3380	0.3399
2.071	51.0	612	2.0308	0.3608	0.1871	0.3368	0.3385
2.0415	52.0	624	2.0283	0.3777	0.1993	0.3546	0.3559
2.007	53.0	636	2.0259	0.3777	0.1993	0.3546	0.3559
2.0238	54.0	648	2.0232	0.3777	0.1993	0.3546	0.3559
2.074	55.0	660	2.0207	0.3780	0.1970	0.3508	0.3527
2.0497	56.0	672	2.0202	0.3780	0.1970	0.3508	0.3527
2.0075	57.0	684	2.0200	0.3780	0.1970	0.3508	0.3527
2.0837	58.0	696	2.0193	0.3780	0.1970	0.3508	0.3527
2.0277	59.0	708	2.0194	0.3780	0.1970	0.3508	0.3527
2.0912	60.0	720	2.0192	0.3780	0.1970	0.3508	0.3527

Framework versions

• Transformers 4.47.1
• Pytorch 2.5.1+cu121
• Datasets 3.2.0
• Tokenizers 0.21.0