gendered_balanced

This model is a fine-tuned version of samzirbo/mT5.en-es.pretrained on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 1.1770
• Bleu: 43.8724
• Meteor: 0.6916
• Chrf++: 62.661

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: 0.0005
• train_batch_size: 64
• eval_batch_size: 64
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine
• lr_scheduler_warmup_steps: 1000
• training_steps: 50000

Training results

Training Loss	Epoch	Step	Validation Loss	Bleu	Meteor	Chrf++
4.3102	0.44	2500	2.0045	27.8405	0.5556	49.2062
2.408	0.88	5000	1.7104	33.3638	0.6039	54.2966
2.1298	1.31	7500	1.5654	36.5253	0.6312	56.719
1.9966	1.75	10000	1.4823	37.7206	0.6409	57.8933
1.8882	2.19	12500	1.4168	39.2333	0.6518	59.0052
1.8055	2.63	15000	1.3736	39.9916	0.6585	59.5057
1.756	3.06	17500	1.3454	40.4973	0.6622	59.9928
1.6829	3.5	20000	1.3109	41.6502	0.6732	60.8574
1.6575	3.94	22500	1.2729	41.7069	0.6729	60.9659
1.5944	4.38	25000	1.2549	42.2762	0.6783	61.5535
1.5758	4.81	27500	1.2349	42.6516	0.6803	61.6749
1.5407	5.25	30000	1.2231	42.6637	0.6817	61.7779
1.5146	5.69	32500	1.2082	42.8507	0.6856	61.9916
1.4953	6.13	35000	1.1988	43.387	0.6873	62.2332
1.4687	6.56	37500	1.1874	43.4174	0.6876	62.3151
1.4664	7.0	40000	1.1825	43.7501	0.6913	62.5472
1.4415	7.44	42500	1.1803	43.7744	0.6909	62.6208
1.4408	7.88	45000	1.1771	43.8914	0.692	62.6599
1.4346	8.31	47500	1.1772	43.9709	0.6919	62.7096
1.4311	8.75	50000	1.1770	43.8724	0.6916	62.661

Framework versions

• Transformers 4.38.0
• Pytorch 2.2.1+cu121
• Datasets 2.19.1
• Tokenizers 0.15.2