zhko_mbartLarge_50p_tokenize_run1

This model is a fine-tuned version of facebook/mbart-large-50-many-to-many-mmt on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 1.8482
• Bleu: 30.6847
• Gen Len: 14.6895

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: 5e-05
• train_batch_size: 4
• eval_batch_size: 4
• seed: 42
• distributed_type: multi-GPU
• num_devices: 4
• gradient_accumulation_steps: 2
• total_train_batch_size: 32
• total_eval_batch_size: 16
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 500
• num_epochs: 8

Training results

Training Loss	Epoch	Step	Validation Loss	Bleu	Gen Len
1.6487	1.0	2786	1.5519	26.5985	15.0925
1.1763	2.0	5572	1.4910	29.1024	14.8538
0.8697	3.0	8358	1.5510	29.5842	14.7611
0.6221	4.0	11144	1.6445	29.6959	14.7091
0.4444	5.0	13930	1.7176	29.6231	14.6204
0.3137	6.0	16716	1.7916	29.6666	14.524
0.2303	7.0	19502	1.8368	30.4697	14.5571
0.1888	8.0	22288	1.8482	30.6847	14.6895

Framework versions

• Transformers 4.34.1
• Pytorch 2.1.0+cu121
• Datasets 2.14.6
• Tokenizers 0.14.1