es_fi_all_quy

This model is a fine-tuned version of nouman-10/es_fi_all_quy on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.4858
• Bleu: 1.9345
• Chrf: 34.1553
• Gen Len: 38.1579

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: 16
• eval_batch_size: 16
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 20
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Bleu	Chrf	Gen Len
0.2239	0.09	1000	0.4842	1.613	34.0838	37.4266
0.2256	0.17	2000	0.4808	1.5738	32.9285	43.005
0.2178	0.26	3000	0.4820	1.499	32.8545	41.5795
0.2175	0.34	4000	0.4828	1.6725	33.3302	39.0342
0.223	0.43	5000	0.4864	2.343	34.8172	35.6197
0.2286	0.51	6000	0.4743	1.7218	33.6246	42.0755
0.2281	0.6	7000	0.4794	1.7699	34.7209	38.7716
0.2243	0.68	8000	0.4802	1.7605	34.16	37.6258
0.236	0.77	9000	0.4772	1.5908	33.9113	39.0543
0.2309	0.85	10000	0.4756	2.0459	34.2757	38.2404
0.2303	0.94	11000	0.4762	1.8315	34.3413	37.7465
0.2218	1.02	12000	0.4860	1.8359	33.5176	38.5392
0.217	1.11	13000	0.4811	1.9919	34.0441	37.3702
0.2158	1.19	14000	0.4845	1.6391	34.5521	37.0493
0.2211	1.28	15000	0.4858	1.9345	34.1553	38.1579

Framework versions

• Transformers 4.28.1
• Pytorch 2.0.0+cu117
• Datasets 2.11.0
• Tokenizers 0.13.3