mt5-small-synthetic-data-plus-translated-bs32ep32

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: 0.9041
• Rouge1: 0.6137
• Rouge2: 0.4715
• Rougel: 0.5917
• Rougelsum: 0.5922

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: 32
• eval_batch_size: 32
• seed: 42
• optimizer: Use OptimizerNames.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: 32

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum
19.5396	1.0	38	11.5658	0.0077	0.0018	0.0067	0.0068
13.1079	2.0	76	7.8457	0.0077	0.0008	0.0075	0.0070
9.4043	3.0	114	3.9170	0.0234	0.0025	0.0217	0.0210
6.388	4.0	152	2.7482	0.1030	0.0252	0.0916	0.0910
4.5077	5.0	190	2.0482	0.1057	0.0414	0.0917	0.0914
3.3242	6.0	228	1.6075	0.1778	0.0897	0.1525	0.1532
2.7	7.0	266	1.3881	0.3601	0.2141	0.3479	0.3487
2.3089	8.0	304	1.2989	0.4295	0.2607	0.4095	0.4091
2.1141	9.0	342	1.2346	0.4337	0.2603	0.4147	0.4146
1.9442	10.0	380	1.1888	0.4926	0.3337	0.4642	0.4644
1.8082	11.0	418	1.1418	0.5101	0.3560	0.4920	0.4929
1.7142	12.0	456	1.1052	0.5341	0.3809	0.5154	0.5155
1.6345	13.0	494	1.0775	0.5605	0.4071	0.5394	0.5394
1.5983	14.0	532	1.0539	0.5790	0.4262	0.5585	0.5580
1.5376	15.0	570	1.0322	0.5713	0.4206	0.5531	0.5532
1.5059	16.0	608	1.0137	0.5807	0.4302	0.5605	0.5605
1.4434	17.0	646	0.9970	0.6069	0.4656	0.5874	0.5871
1.442	18.0	684	0.9826	0.6104	0.4671	0.5869	0.5874
1.4059	19.0	722	0.9688	0.6102	0.4666	0.5886	0.5883
1.3618	20.0	760	0.9636	0.6127	0.4683	0.5901	0.5906
1.3341	21.0	798	0.9517	0.6065	0.4632	0.5852	0.5860
1.3019	22.0	836	0.9397	0.6092	0.4669	0.5886	0.5882
1.3114	23.0	874	0.9343	0.6091	0.4663	0.5869	0.5871
1.2906	24.0	912	0.9272	0.6137	0.4702	0.5913	0.5912
1.255	25.0	950	0.9201	0.6153	0.4723	0.5934	0.5934
1.261	26.0	988	0.9186	0.6174	0.4749	0.5966	0.5965
1.2363	27.0	1026	0.9124	0.6155	0.4738	0.5948	0.5952
1.2993	28.0	1064	0.9078	0.6153	0.4738	0.5937	0.5939
1.2653	29.0	1102	0.9054	0.6126	0.4712	0.5919	0.5922
1.2287	30.0	1140	0.9046	0.6099	0.4676	0.5894	0.5901
1.2279	31.0	1178	0.9041	0.6137	0.4715	0.5917	0.5922
1.2348	32.0	1216	0.9041	0.6137	0.4715	0.5917	0.5922

Framework versions

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