am_hi_model

This model is a fine-tuned version of Piyush41/Sanskrit-Hindi on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 5.3728
• Sacrebleu: 0.1368
• Bleu: 0.0014
• Rouge1: 0.0
• Rouge2: 0.0
• Rougel: 0.0
• Ter: 97.8434
• Gen Len: 20.0

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: 16
• eval_batch_size: 16
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Sacrebleu	Bleu	Rouge1	Rouge2	Rougel	Ter	Gen Len
7.9097	1.0	509	6.8376	0.001	0.0	0.0	0.0	0.0	99.2079	20.0
6.6631	2.0	1018	6.2731	0.012	0.0001	0.0	0.0	0.0	98.8741	20.0
6.2666	3.0	1527	5.9668	0.0782	0.0008	0.0	0.0	0.0	98.3207	20.0
6.0185	4.0	2036	5.7664	0.0793	0.0008	0.0	0.0	0.0	98.3292	20.0
5.8656	5.0	2545	5.6281	0.112	0.0011	0.0	0.0	0.0	98.0673	20.0
5.7354	6.0	3054	5.5269	0.1582	0.0016	0.0	0.0	0.0	97.8476	20.0
5.6509	7.0	3563	5.4556	0.1561	0.0016	0.0	0.0	0.0	97.949	20.0
5.5912	8.0	4072	5.4093	0.1368	0.0014	0.0	0.0	0.0	97.9405	20.0
5.5534	9.0	4581	5.3812	0.1329	0.0013	0.0	0.0	0.0	97.856	20.0
5.5356	10.0	5090	5.3728	0.1368	0.0014	0.0	0.0	0.0	97.8434	20.0

Framework versions

• Transformers 4.34.0
• Pytorch 2.1.0+cu118
• Datasets 2.14.5
• Tokenizers 0.14.1