Modelpublic

CreAtlons/english-tamil-colloquial-translator

sourceHugging Faceapache-2.0updated 2y agoView on Hugging Face

0likes11downloads

Model Card

english-tamil-colloquial-translator

This model is a fine-tuned version of unsloth/tinyllama-chat-bnb-4bit on an unknown dataset. It achieves the following results on the evaluation set:

—Loss: 6.4734

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: 0.0003
—trainbatchsize: 14
—evalbatchsize: 14
—seed: 42
—optimizer: Use OptimizerNames.ADAMWTORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizerargs=No additional optimizer arguments
—lrschedulertype: linear
—lrschedulerwarmup_steps: 10
—num_epochs: 1
—mixedprecisiontraining: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
5.0805	0.0219	20	6.3932
2.7551	0.0438	40	5.7325
2.1716	0.0657	60	6.5299
2.2547	0.0876	80	6.7338
2.1684	0.1095	100	6.4174
2.4315	0.1314	120	6.0452
2.2081	0.1533	140	6.3193
1.951	0.1752	160	6.4108
1.9445	0.1972	180	6.2947
1.9979	0.2191	200	6.2312
2.1081	0.2410	220	6.4618
1.7883	0.2629	240	6.5584
1.922	0.2848	260	6.4296
2.2168	0.3067	280	6.4355
1.8854	0.3286	300	6.4379
2.0228	0.3505	320	6.3622
1.9864	0.3724	340	6.3381
2.0443	0.3943	360	6.4798
1.9478	0.4162	380	6.5400
1.773	0.4381	400	6.6497
2.0066	0.4600	420	6.6103
1.8865	0.4819	440	6.5956
2.1984	0.5038	460	6.5025
1.9825	0.5257	480	6.5883
1.975	0.5476	500	6.4721
2.0906	0.5696	520	6.6262
1.9389	0.5915	540	6.4519
1.734	0.6134	560	6.3723
2.2446	0.6353	580	6.3517
2.102	0.6572	600	6.4213
1.9183	0.6791	620	6.3832
1.7529	0.7010	640	6.4442
2.1462	0.7229	660	6.4660
1.6759	0.7448	680	6.5409
1.6992	0.7667	700	6.5130
1.8339	0.7886	720	6.5549
1.9626	0.8105	740	6.5354
1.7808	0.8324	760	6.4610
1.9102	0.8543	780	6.4915
1.8514	0.8762	800	6.5019
1.7297	0.8981	820	6.5373
1.7459	0.9200	840	6.5211
2.0906	0.9419	860	6.5073
2.1094	0.9639	880	6.4955
1.9157	0.9858	900	6.4734

Framework versions

—PEFT 0.14.0
—Transformers 4.48.3
—Pytorch 2.6.0+cu124
—Datasets 3.3.2
—Tokenizers 0.21.0