Tounsify-v0.10-shuffle

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

• Loss: 1.5449
• Bleu: 35.7955
• Gen Len: 9.4516

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: 4
• eval_batch_size: 4
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 100
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Bleu	Gen Len
No log	1.0	62	2.2061	11.8171	9.5161
No log	2.0	124	1.5962	21.0419	9.3065
No log	3.0	186	1.2980	24.4645	9.1935
No log	4.0	248	1.2160	28.5544	9.2742
No log	5.0	310	1.1631	31.2379	9.1452
No log	6.0	372	1.1556	32.39	9.2097
No log	7.0	434	1.1650	32.8728	9.2419
No log	8.0	496	1.1666	28.6033	9.0645
0.9718	9.0	558	1.1976	25.7388	10.0484
0.9718	10.0	620	1.1648	32.5972	9.2258
0.9718	11.0	682	1.1939	31.9682	9.0806
0.9718	12.0	744	1.2021	33.2574	9.371
0.9718	13.0	806	1.2006	32.1413	9.2258
0.9718	14.0	868	1.2208	33.0105	9.3065
0.9718	15.0	930	1.2888	31.9994	9.4194
0.9718	16.0	992	1.2568	33.6234	9.3226
0.0484	17.0	1054	1.2758	32.9602	9.3226
0.0484	18.0	1116	1.2841	33.2857	9.2903
0.0484	19.0	1178	1.2968	32.4006	9.2419
0.0484	20.0	1240	1.3066	32.7878	9.1935
0.0484	21.0	1302	1.3192	32.1068	9.3871
0.0484	22.0	1364	1.3158	32.7501	9.3226
0.0484	23.0	1426	1.3553	33.1188	9.3065
0.0484	24.0	1488	1.3182	33.9851	9.6613
0.0137	25.0	1550	1.3493	32.7566	9.3226
0.0137	26.0	1612	1.3419	33.8387	9.5
0.0137	27.0	1674	1.3501	32.2899	9.371
0.0137	28.0	1736	1.3520	32.1795	9.3226
0.0137	29.0	1798	1.3676	33.7723	9.5645
0.0137	30.0	1860	1.3832	32.8767	9.3548
0.0137	31.0	1922	1.3814	33.3269	9.5
0.0137	32.0	1984	1.3833	32.8231	9.371
0.0108	33.0	2046	1.3828	32.1068	9.4194
0.0108	34.0	2108	1.3976	33.9396	9.4677
0.0108	35.0	2170	1.4015	32.1225	9.1613
0.0108	36.0	2232	1.4058	32.7627	9.371
0.0108	37.0	2294	1.4213	32.1195	9.2581
0.0108	38.0	2356	1.4300	33.0876	9.4355
0.0108	39.0	2418	1.4263	32.7883	9.3387
0.0108	40.0	2480	1.4390	31.3041	9.3387
0.0107	41.0	2542	1.4405	33.2307	9.3871
0.0107	42.0	2604	1.4421	32.5338	9.3871
0.0107	43.0	2666	1.4617	31.5815	9.3548
0.0107	44.0	2728	1.4517	32.2336	9.3226
0.0107	45.0	2790	1.4708	32.5791	9.4194
0.0107	46.0	2852	1.4665	33.5456	9.4516
0.0107	47.0	2914	1.4574	32.4045	9.3871
0.0107	48.0	2976	1.4585	34.6859	9.4677
0.0099	49.0	3038	1.4733	34.7	9.4355
0.0099	50.0	3100	1.4713	34.7405	9.4032
0.0099	51.0	3162	1.4740	34.6316	9.4355
0.0099	52.0	3224	1.4867	35.6172	9.4516
0.0099	53.0	3286	1.4845	34.6718	9.5
0.0099	54.0	3348	1.4891	35.3407	9.4194
0.0099	55.0	3410	1.4868	35.243	9.3871
0.0099	56.0	3472	1.4695	35.326	9.4839
0.0069	57.0	3534	1.4851	35.3597	9.4355
0.0069	58.0	3596	1.4960	34.1363	9.4032
0.0069	59.0	3658	1.4808	34.8965	9.4677
0.0069	60.0	3720	1.4891	34.7792	9.4839
0.0069	61.0	3782	1.4882	34.0964	9.4839
0.0069	62.0	3844	1.4952	34.0563	9.5
0.0069	63.0	3906	1.4995	35.5974	9.4839
0.0069	64.0	3968	1.5149	35.873	9.4194
0.0087	65.0	4030	1.5180	35.3407	9.4516
0.0087	66.0	4092	1.5169	35.9762	9.4677
0.0087	67.0	4154	1.5101	35.8579	9.4677
0.0087	68.0	4216	1.5098	35.2643	9.4677
0.0087	69.0	4278	1.5119	35.2643	9.4677
0.0087	70.0	4340	1.5109	35.0762	9.4194
0.0087	71.0	4402	1.5118	35.3965	9.4839
0.0087	72.0	4464	1.5099	35.3965	9.4839
0.0056	73.0	4526	1.5249	35.2791	9.5
0.0056	74.0	4588	1.5197	35.2791	9.4677
0.0056	75.0	4650	1.5288	35.2021	9.4516
0.0056	76.0	4712	1.5323	35.2021	9.4516
0.0056	77.0	4774	1.5264	35.2021	9.4516
0.0056	78.0	4836	1.5266	35.2021	9.4516
0.0056	79.0	4898	1.5285	35.2021	9.4516
0.0056	80.0	4960	1.5326	35.7955	9.4516
0.0058	81.0	5022	1.5339	35.7955	9.4516
0.0058	82.0	5084	1.5435	35.2021	9.4516
0.0058	83.0	5146	1.5421	35.2021	9.4516
0.0058	84.0	5208	1.5441	35.7955	9.4516
0.0058	85.0	5270	1.5484	35.7955	9.4516
0.0058	86.0	5332	1.5527	35.7955	9.4516
0.0058	87.0	5394	1.5497	35.7955	9.4516
0.0058	88.0	5456	1.5504	35.7955	9.4516
0.0055	89.0	5518	1.5485	35.7955	9.4516
0.0055	90.0	5580	1.5484	35.7955	9.4516
0.0055	91.0	5642	1.5496	35.7955	9.4516
0.0055	92.0	5704	1.5475	35.739	9.4516
0.0055	93.0	5766	1.5438	35.739	9.4516
0.0055	94.0	5828	1.5464	35.739	9.4516
0.0055	95.0	5890	1.5461	35.739	9.4516
0.0055	96.0	5952	1.5467	35.7955	9.4516
0.0045	97.0	6014	1.5452	35.7955	9.4516
0.0045	98.0	6076	1.5449	35.7955	9.4516
0.0045	99.0	6138	1.5449	35.7955	9.4516
0.0045	100.0	6200	1.5449	35.7955	9.4516

Framework versions

• Transformers 4.41.2
• Pytorch 2.3.0+cu121
• Datasets 2.20.0
• Tokenizers 0.19.1