m2m100_lr_2e5_gradd_accum_1

This model is a fine-tuned version of facebook/m2m100_418M on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 3.2511
• Bleu: 10.7253
• Gen Len: 45.9543
• Meteor: 0.3079
• Chrf: 33.934

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: 12
• eval_batch_size: 12
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 32.0

Training results

Training Loss	Epoch	Step	Validation Loss	Bleu	Gen Len	Meteor	Chrf
3.6632	3.97	2406	2.8977	7.6826	49.2935	0.2567	28.2679
2.1329	7.94	4812	2.7600	10.0915	47.2241	0.2999	32.4294
1.552	11.91	7218	2.8218	10.3506	45.6178	0.3041	33.2923
1.1577	15.88	9624	2.9258	10.3313	46.668	0.3058	33.4639
0.8781	19.85	12030	3.0435	10.5266	46.2384	0.3063	33.6649
0.6935	23.82	14436	3.1381	10.4391	46.0441	0.3062	33.7686
0.5683	27.79	16842	3.2178	10.6801	45.8612	0.309	33.8163
0.5037	31.76	19248	3.2511	10.7253	45.9543	0.3079	33.934

Framework versions

• Transformers 4.30.2
• Pytorch 1.11.0+cu113
• Datasets 2.10.0
• Tokenizers 0.12.1