m2m100_418M-finetuned-tq-to-ar-1

This model is a fine-tuned version of ad019el/m2m100_418M-finetuned-tq-to-ar on the None dataset. It achieves the following results on the evaluation set:

• Loss: 2.2002
• Bleu: 3.6349
• Gen Len: 35.5271

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: 8
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 15

Training results

Training Loss	Epoch	Step	Validation Loss	Bleu	Gen Len
2.7537	0.71	500	2.2710	4.2969	35.4312
2.6442	1.42	1000	2.2373	4.0784	35.1062
2.6329	2.13	1500	2.2257	3.8894	36.225
2.564	2.84	2000	2.2210	3.5513	36.076
2.5352	3.56	2500	2.2151	3.7339	35.0885
2.4991	4.27	3000	2.2078	3.4662	36.3333
2.4782	4.98	3500	2.2100	3.3332	36.4062
2.4363	5.69	4000	2.2085	3.3587	36.3135
2.4411	6.4	4500	2.2034	3.8744	34.5073
2.4002	7.11	5000	2.2036	3.6693	36.3448
2.3841	7.82	5500	2.2030	3.7486	35.076
2.3619	8.53	6000	2.1970	3.5687	35.8271
2.3627	9.25	6500	2.2016	3.5394	35.3583
2.3451	9.96	7000	2.1996	3.5863	34.9271
2.3323	10.67	7500	2.2002	3.6349	35.5271

Framework versions

• Transformers 4.32.0
• Pytorch 2.0.1+cu118
• Datasets 2.14.4
• Tokenizers 0.13.3