You can deploy this model and start earning money today!

thunderboltc/
mbart50-sanlish-to-bangla_1934_lr2e5_bleufixed

Name: mbart50-sanlish-to-bangla_1934_lr2e5_bleufixed
Availability: InStock

Views

No views yet

mbart50-sanlish-to-bangla_1934_lr2e5_bleufixed

This model is a fine-tuned version of facebook/mbart-large-50-many-to-many-mmt on the None dataset. It achieves the following results on the evaluation set:

More information needed

More information needed

More information needed

The following hyperparameters were used during training:

learning_rate: 2e-05
train_batch_size: 8
eval_batch_size: 8
seed: 42
gradient_accumulation_steps: 2
total_train_batch_size: 16
optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
lr_scheduler_type: linear
num_epochs: 25
mixed_precision_training: Native AMP

Training Loss	Epoch	Step	Validation Loss	Bleu	Chrf
8.3315	1.0	98	3.5064	0.7883	12.1081
5.5967	2.0	196	3.0986	2.5229	17.2748
4.1783	3.0	294	2.9789	2.3490	19.8955
3.0388	4.0	392	2.9608	4.4033	23.7514
2.1599	5.0	490	3.0397	3.6133	25.3107
1.5631	6.0	588	3.0399	5.9597	27.0649
1.1581	7.0	686	3.1427	5.0788	27.3468
0.8721	8.0	784	3.1384	7.2174	30.3129
0.6972	9.0	882	3.1405	7.5372	30.1748
0.5659	10.0	980	3.2135	7.3507	29.4237
0.4848	11.0	1078	3.2691	7.6517	29.9946
0.4137	12.0	1176	3.2783	8.0592	28.8309
0.3703	13.0	1274	3.2653	8.6232	31.8898
0.3202	14.0	1372	3.2907	8.7342	31.8681
0.2878	15.0	1470	3.3074	8.2271	31.2518
0.2554	16.0	1568	3.2872	8.2875	31.6010
0.2243	17.0	1666	3.3393	9.0506	32.5280
0.1985	18.0	1764	3.3172	10.3651	32.2873
0.1854	19.0	1862	3.3454	9.5312	32.4380
0.1674	20.0	1960	3.3242	9.2245	32.4313
0.1566	21.0	2058	3.3230	9.1774	33.7570
0.1497	22.0	2156	3.3441	9.8885	33.2189
0.1409	23.0	2254	3.3407	8.4301	33.0052