gendered

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: 1.1747
• Bleu: 43.5153
• Meteor: 0.6866
• Chrf++: 62.4637

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++
4.5293	0.26	2500	2.0219	28.0137	0.5556	49.0872
2.4187	0.53	5000	1.7260	33.4178	0.6039	54.238
2.1694	0.79	7500	1.5868	35.7611	0.6256	56.142
2.0239	1.05	10000	1.4968	37.4453	0.6394	57.6176
1.9076	1.32	12500	1.4333	38.2643	0.6473	58.2946
1.8514	1.58	15000	1.3902	39.5846	0.6554	59.3566
1.7986	1.84	17500	1.3454	39.9905	0.6612	59.701
1.7351	2.11	20000	1.3096	40.8823	0.6664	60.3625
1.6834	2.37	22500	1.2872	41.4499	0.6711	60.837
1.6589	2.64	25000	1.2591	42.0603	0.6743	61.2625
1.6353	2.9	27500	1.2353	42.3878	0.6798	61.5963
1.5929	3.16	30000	1.2242	42.7944	0.6794	61.8097
1.5635	3.43	32500	1.2094	43.0109	0.6822	61.9923
1.5537	3.69	35000	1.2011	43.1533	0.6839	62.1374
1.5449	3.95	37500	1.1873	43.2691	0.6844	62.1844
1.5131	4.22	40000	1.1830	43.3538	0.6868	62.3435
1.5067	4.48	42500	1.1790	43.4627	0.6861	62.4529
1.502	4.74	45000	1.1761	43.4613	0.686	62.3993
1.4976	5.01	47500	1.1749	43.6051	0.6866	62.4784
1.4889	5.27	50000	1.1747	43.5153	0.6866	62.4637

Framework versions

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