--- license: apache-2.0 base_model: Helsinki-NLP/opus-mt-en-zh tags: - generated_from_trainer metrics: - bleu model-index: - name: opus-mt-cantonese-v1 results: [] --- # opus-mt-cantonese-v1 This model is a fine-tuned version of [Helsinki-NLP/opus-mt-en-zh](https://huggingface.co/Helsinki-NLP/opus-mt-en-zh) on an unknown dataset. It achieves the following results on the evaluation set: - Loss: 3.7189 - Bleu: 1.3095 - Gen Len: 12.8089 ## 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-06 - 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: 50 - mixed_precision_training: Native AMP ### Training results | Training Loss | Epoch | Step | Validation Loss | Bleu | Gen Len | |:-------------:|:-----:|:----:|:---------------:|:------:|:-------:| | No log | 1.0 | 62 | 4.0260 | 0.5212 | 12.1748 | | No log | 2.0 | 124 | 3.9917 | 0.5398 | 12.2033 | | No log | 3.0 | 186 | 3.9573 | 0.923 | 12.0894 | | No log | 4.0 | 248 | 3.9330 | 0.9257 | 12.252 | | No log | 5.0 | 310 | 3.9073 | 0.9197 | 12.2154 | | No log | 6.0 | 372 | 3.8840 | 0.9586 | 12.2561 | | No log | 7.0 | 434 | 3.8681 | 0.9702 | 12.3374 | | No log | 8.0 | 496 | 3.8540 | 0.9676 | 12.3415 | | 3.0832 | 9.0 | 558 | 3.8380 | 0.9564 | 12.4268 | | 3.0832 | 10.0 | 620 | 3.8276 | 0.963 | 12.5081 | | 3.0832 | 11.0 | 682 | 3.8159 | 0.9326 | 12.5528 | | 3.0832 | 12.0 | 744 | 3.8086 | 0.9326 | 12.5772 | | 3.0832 | 13.0 | 806 | 3.8007 | 0.9668 | 12.5813 | | 3.0832 | 14.0 | 868 | 3.7919 | 0.922 | 12.7073 | | 3.0832 | 15.0 | 930 | 3.7833 | 0.9319 | 12.626 | | 3.0832 | 16.0 | 992 | 3.7754 | 1.0907 | 12.7033 | | 2.6953 | 17.0 | 1054 | 3.7698 | 1.0914 | 12.7317 | | 2.6953 | 18.0 | 1116 | 3.7657 | 1.1198 | 12.7642 | | 2.6953 | 19.0 | 1178 | 3.7597 | 1.2304 | 12.6707 | | 2.6953 | 20.0 | 1240 | 3.7555 | 1.2345 | 12.7683 | | 2.6953 | 21.0 | 1302 | 3.7519 | 1.2465 | 12.7439 | | 2.6953 | 22.0 | 1364 | 3.7506 | 1.2322 | 12.7764 | | 2.6953 | 23.0 | 1426 | 3.7480 | 1.2558 | 12.7642 | | 2.6953 | 24.0 | 1488 | 3.7453 | 1.2465 | 12.7317 | | 2.4546 | 25.0 | 1550 | 3.7415 | 1.2614 | 12.7073 | | 2.4546 | 26.0 | 1612 | 3.7377 | 1.2339 | 12.7073 | | 2.4546 | 27.0 | 1674 | 3.7346 | 1.2664 | 12.7195 | | 2.4546 | 28.0 | 1736 | 3.7315 | 1.2664 | 12.7195 | | 2.4546 | 29.0 | 1798 | 3.7310 | 1.3041 | 12.7033 | | 2.4546 | 30.0 | 1860 | 3.7293 | 1.2715 | 12.687 | | 2.4546 | 31.0 | 1922 | 3.7266 | 1.2941 | 12.6748 | | 2.4546 | 32.0 | 1984 | 3.7266 | 1.2988 | 12.7398 | | 2.2894 | 33.0 | 2046 | 3.7260 | 1.3227 | 12.7439 | | 2.2894 | 34.0 | 2108 | 3.7243 | 1.3227 | 12.752 | | 2.2894 | 35.0 | 2170 | 3.7240 | 1.3227 | 12.752 | | 2.2894 | 36.0 | 2232 | 3.7230 | 1.3338 | 12.7276 | | 2.2894 | 37.0 | 2294 | 3.7242 | 1.3338 | 12.7724 | | 2.2894 | 38.0 | 2356 | 3.7224 | 1.3338 | 12.7764 | | 2.2894 | 39.0 | 2418 | 3.7210 | 1.3338 | 12.7642 | | 2.2894 | 40.0 | 2480 | 3.7214 | 1.351 | 12.7642 | | 2.1784 | 41.0 | 2542 | 3.7215 | 1.3283 | 12.7967 | | 2.1784 | 42.0 | 2604 | 3.7208 | 1.3173 | 12.7642 | | 2.1784 | 43.0 | 2666 | 3.7208 | 1.3519 | 12.7114 | | 2.1784 | 44.0 | 2728 | 3.7200 | 1.3519 | 12.7114 | | 2.1784 | 45.0 | 2790 | 3.7198 | 1.3173 | 12.7886 | | 2.1784 | 46.0 | 2852 | 3.7199 | 1.3519 | 12.752 | | 2.1784 | 47.0 | 2914 | 3.7192 | 1.3576 | 12.7642 | | 2.1784 | 48.0 | 2976 | 3.7191 | 1.3282 | 12.7967 | | 2.1365 | 49.0 | 3038 | 3.7189 | 1.3282 | 12.8089 | | 2.1365 | 50.0 | 3100 | 3.7189 | 1.3095 | 12.8089 | ### Framework versions - Transformers 4.39.2 - Pytorch 2.2.1+cu121 - Datasets 2.18.0 - Tokenizers 0.15.2