nllb-200-distilled-600M-finetuned_ramayana_sns_lexr

This model is a fine-tuned version of facebook/nllb-200-distilled-600M on the None dataset. It achieves the following results on the evaluation set:

• Loss: 3.2899
• Rouge1: 18.7327
• Rouge2: 2.1067
• Rougel: 14.4307
• Rougelsum: 16.856

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: 5.6e-06
• train_batch_size: 1
• eval_batch_size: 1
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum
4.5091	1.0	427	4.0603	16.2779	1.6057	13.7589	14.9914
4.1329	2.0	854	3.8728	16.1259	1.3608	13.0307	14.5339
3.9846	3.0	1281	3.7534	17.0824	1.7524	13.8469	15.5679
3.8863	4.0	1708	3.6693	16.9334	1.7307	13.8506	15.5077
3.8144	5.0	2135	3.5996	18.0132	1.9515	14.3725	16.3556
3.7542	6.0	2562	3.5460	17.4338	1.8704	14.1021	15.9508
3.7068	7.0	2989	3.5052	17.5684	2.0413	14.0332	15.9409
3.665	8.0	3416	3.4695	18.2728	2.0033	14.2529	16.6322
3.6288	9.0	3843	3.4347	18.0755	2.2161	14.0724	16.4611
3.5969	10.0	4270	3.4081	18.3883	2.1437	14.3847	16.7879
3.5718	11.0	4697	3.3840	18.9654	2.2593	14.6253	17.3421
3.5464	12.0	5124	3.3619	18.9897	2.352	14.679	17.4166
3.5302	13.0	5551	3.3465	18.9671	2.246	14.4441	17.16
3.5102	14.0	5978	3.3309	18.5565	2.1854	14.1515	16.6572
3.4992	15.0	6405	3.3171	19.0665	2.1941	14.7519	17.2236
3.4851	16.0	6832	3.3075	18.5714	2.1059	14.3258	16.7627
3.4765	17.0	7259	3.2998	18.5252	2.083	14.1246	16.724
3.464	18.0	7686	3.2944	18.9694	2.1503	14.6684	17.0653
3.4647	19.0	8113	3.2911	18.6916	2.1447	14.4372	16.8423
3.4585	20.0	8540	3.2899	18.7327	2.1067	14.4307	16.856

Framework versions

• Transformers 4.37.2
• Pytorch 2.1.2+cu121
• Datasets 2.14.4
• Tokenizers 0.15.2