smita1988/english-hindi-translation-model

hindi-colloquial-translator

This model is a fine-tuned version of unsloth/tinyllama-chat-bnb-4bit on the None dataset. It achieves the following results on the evaluation set:

• Loss: 13.9788

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
• train_batch_size: 4
• eval_batch_size: 4
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 8
• 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_steps: 10
• num_epochs: 3
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
14.9448	0.0333	10	12.0851
10.5162	0.0667	20	12.1086
6.5944	0.1	30	12.1206
4.5955	0.1333	40	12.1867
3.9384	0.1667	50	12.3252
3.7184	0.2	60	12.6046
3.596	0.2333	70	13.1815
3.5198	0.2667	80	13.3539
3.6022	0.3	90	12.9698
3.5499	0.3333	100	13.1654
3.5272	0.3667	110	13.3949
3.5894	0.4	120	13.4733
3.5825	0.4333	130	13.6201
3.5728	0.4667	140	13.6792
3.5061	0.5	150	13.6643
3.6101	0.5333	160	13.6148
3.557	0.5667	170	13.6482
3.5036	0.6	180	13.6264
3.4877	0.6333	190	13.6439
3.5853	0.6667	200	13.6861
3.5512	0.7	210	13.6245
3.5651	0.7333	220	13.7202
3.6101	0.7667	230	13.7473
3.5352	0.8	240	13.7057
3.5732	0.8333	250	13.7293
3.5483	0.8667	260	13.8731
3.5032	0.9	270	13.9614
3.5584	0.9333	280	14.0209
3.4704	0.9667	290	14.0475
3.5769	1.0	300	13.9849
3.5266	1.0333	310	13.9310
3.5559	1.0667	320	13.9112
3.522	1.1	330	13.9922
3.5856	1.1333	340	14.0535
3.5635	1.1667	350	14.0731
3.5347	1.2	360	14.0378
3.5998	1.2333	370	14.0262
3.5353	1.2667	380	13.9983
3.5461	1.3	390	13.9643
3.5471	1.3333	400	13.9684
3.5621	1.3667	410	13.9681
3.497	1.4	420	13.9489
3.4891	1.4333	430	13.9449
3.5001	1.4667	440	13.9471
3.4829	1.5	450	13.9471
3.5042	1.5333	460	13.9469
3.5616	1.5667	470	13.9399
3.5716	1.6	480	13.9463
3.5792	1.6333	490	13.9335
3.5115	1.6667	500	13.9425
3.432	1.7	510	13.9456
3.6118	1.7333	520	13.9393
3.53	1.7667	530	13.9433
3.5557	1.8	540	13.9360
3.5061	1.8333	550	13.9372
3.5141	1.8667	560	13.9420
3.5219	1.9	570	13.9484
3.5528	1.9333	580	13.9506
3.5767	1.9667	590	13.9557
3.576	2.0	600	13.9629
3.5206	2.0333	610	13.9666
3.5492	2.0667	620	13.9642
3.4966	2.1	630	13.9700
3.5562	2.1333	640	13.9725
3.4972	2.1667	650	13.9789
3.5234	2.2	660	13.9711
3.53	2.2333	670	13.9741
3.553	2.2667	680	13.9691
3.5519	2.3	690	13.9713
3.5854	2.3333	700	13.9782
3.4821	2.3667	710	13.9779
3.501	2.4	720	13.9834
3.5755	2.4333	730	13.9854
3.5215	2.4667	740	13.9791
3.5579	2.5	750	13.9823
3.5026	2.5333	760	13.9829
3.5423	2.5667	770	13.9793
3.5334	2.6	780	13.9822
3.5388	2.6333	790	13.9845
3.5442	2.6667	800	13.9853
3.5159	2.7	810	13.9844
3.4691	2.7333	820	13.9812
3.5149	2.7667	830	13.9821
3.5756	2.8	840	13.9809
3.552	2.8333	850	13.9823
3.5843	2.8667	860	13.9786
3.5026	2.9	870	13.9794
3.5703	2.9333	880	13.9802
3.5147	2.9667	890	13.9810
3.5719	3.0	900	13.9788

Framework versions

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