marianmt-santali-ipa-to-bangla_normalSplit

thunderboltc/marianmt-santali-ipa-to-bangla_normalSplit

marianmt-santali-ipa-to-bangla_normalSplit is a machine learning model from thunderboltc. Use it for the machine learning task on the model card, and read the license before you ship it in a product. It is set up for transformers. The card lists the license as apache-2.0.

should probably proofread and complete it, then remove this comment. --

Downloads · 30 days

5% of all-time downloads

All-time downloads

367

Public

Parameters

77M

27.7 GB on disk

Likes

Public

Hugging Face

Repo makeup

Click a slice to open those files.

.safetensors616 MB · 50%

At a glance

Library: transformers
License: apache-2.0
Model type: marian
Access: Public
Created: Aug 23, 2026
Updated: Aug 23, 2026
SHA: 8b46a02f

Base models

Helsinki-NLP/opus-mt-en-mul

Model card

From the Hugging Face model README

marianmt-santali-ipa-to-bangla_normalSplit

This model is a fine-tuned version of Helsinki-NLP/opus-mt-en-mul on the None dataset. It achieves the following results on the evaluation set:

Loss: 1.7438
Bleu: 10.8230
Chrf: 35.9474
Meteor: 0.3051
Bertscore: 0.8376

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: 8
eval_batch_size: 8
seed: 42
optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
lr_scheduler_type: linear
lr_scheduler_warmup_ratio: 0.1
num_epochs: 30
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Bleu	Chrf	Meteor	Bertscore
2.8745	1.0	194	2.5988	0.3358	10.8306	0.0336	0.7126
2.5044	2.0	388	2.2998	0.5050	12.3618	0.0469	0.7310
2.2231	3.0	582	2.1207	1.4073	15.1310	0.0825	0.7558
2.024	4.0	776	1.9867	2.1319	18.0802	0.1291	0.7761
1.7752	5.0	970	1.8718	2.5581	22.2991	0.1773	0.7907
1.4984	6.0	1164	1.7765	3.4326	25.0497	0.2008	0.7978
1.3704	7.0	1358	1.7767	3.3985	25.0114	0.2034	0.7984
1.2993	8.0	1552	1.7245	5.3114	25.9564	0.2172	0.8083
1.1183	9.0	1746	1.7270	4.9278	27.5015	0.2295	0.8131
1.116	10.0	1940	1.6764	7.0247	29.4231	0.2548	0.8206
0.9505	11.0	2134	1.6977	5.7992	30.0031	0.2531	0.8200
0.8805	12.0	2328	1.6682	7.0811	31.1998	0.2605	0.8226
0.7953	13.0	2522	1.6678	7.1778	32.4209	0.2786	0.8301
0.7558	14.0	2716	1.6792	6.8708	32.6178	0.2809	0.8298
0.7147	15.0	2910	1.6694	8.7661	33.8030	0.2923	0.8311
0.6098	16.0	3104	1.6749	8.2144	34.1549	0.2960	0.8363
0.6088	17.0	3298	1.6897	8.8893	33.5770	0.2940	0.8359
0.5913	18.0	3492	1.6957	8.9283	33.5152	0.2961	0.8318
0.5362	19.0	3686	1.6878	9.1546	33.9310	0.2973	0.8342
0.4918	20.0	3880	1.7068	9.4396	35.3776	0.3056	0.8357
0.4642	21.0	4074	1.6999	10.2171	35.5317	0.3062	0.8399
0.4272	22.0	4268	1.7180	9.9976	35.1329	0.3048	0.8360
0.4119	23.0	4462	1.7229	10.3259	35.4247	0.3050	0.8373
0.3877	24.0	4656	1.7212	10.5739	35.4685	0.3052	0.8356
0.3964	25.0	4850	1.7379	10.3918	36.3235	0.3105	0.8397
0.3571	26.0	5044	1.7341	10.9371	36.3835	0.3137	0.8413
0.3462	27.0	5238	1.7381	10.5170	35.7879	0.3105	0.8378
0.3184	28.0	5432	1.7400	10.5324	35.6082	0.3129	0.8391
0.354	29.0	5626	1.7422	11.3840	36.3976	0.3079	0.8379
0.3315	30.0	5820	1.7438	10.8230	35.9474	0.3051	0.8376

Framework versions

Transformers 4.46.3
Pytorch 2.11.0+cu128
Datasets 4.0.0
Tokenizers 0.20.3

Library: transformers
Type: marian
License: apache-2.0
Created: Aug 23, 2026
Updated: Aug 23, 2026