Tounsify-v0.9-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: 1.2171
• Bleu: 47.287
• Gen Len: 9.1774

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	2.2319	11.5922	8.3226
No log	2.0	124	1.4979	22.8539	8.3871
No log	3.0	186	1.1749	31.2278	8.5323
No log	4.0	248	1.0500	39.4966	8.7097
No log	5.0	310	0.9562	42.3858	8.7742
No log	6.0	372	0.9306	43.1436	8.6935
No log	7.0	434	0.8928	42.3849	8.8387
No log	8.0	496	0.9243	42.8107	8.8548
0.9876	9.0	558	0.9293	44.3329	8.8548
0.9876	10.0	620	0.9398	42.859	8.871
0.9876	11.0	682	0.9637	44.6861	8.8548
0.9876	12.0	744	0.9514	45.1661	8.8387
0.9876	13.0	806	0.9780	45.5317	8.8226
0.9876	14.0	868	0.9832	48.3237	8.8548
0.9876	15.0	930	0.9618	49.9886	9.0484
0.9876	16.0	992	0.9980	47.1846	8.9516
0.0522	17.0	1054	0.9758	45.6558	8.9839
0.0522	18.0	1116	0.9907	45.325	9.0
0.0522	19.0	1178	1.0234	48.1955	8.9194
0.0522	20.0	1240	1.0339	47.0583	8.9839
0.0522	21.0	1302	1.0129	49.2604	8.8871
0.0522	22.0	1364	1.0407	49.847	8.8871
0.0522	23.0	1426	1.0656	48.4962	8.9839
0.0522	24.0	1488	1.0504	48.3458	8.9839
0.0153	25.0	1550	1.0556	49.455	9.0161
0.0153	26.0	1612	1.0522	48.9644	9.0323
0.0153	27.0	1674	1.0793	48.7056	8.9839
0.0153	28.0	1736	1.0859	48.8805	8.9839
0.0153	29.0	1798	1.1362	48.306	9.0806
0.0153	30.0	1860	1.0573	51.8905	9.2097
0.0153	31.0	1922	1.1220	48.3591	9.0806
0.0153	32.0	1984	1.0879	49.0288	9.129
0.0097	33.0	2046	1.1219	50.593	9.1129
0.0097	34.0	2108	1.1439	49.1391	9.0
0.0097	35.0	2170	1.1265	50.5195	9.0323
0.0097	36.0	2232	1.1031	50.2673	9.0806
0.0097	37.0	2294	1.1418	51.3256	8.9839
0.0097	38.0	2356	1.1419	50.8617	9.0968
0.0097	39.0	2418	1.1166	51.2853	9.1452
0.0097	40.0	2480	1.1309	50.6103	9.0806
0.0082	41.0	2542	1.1501	50.7017	9.0
0.0082	42.0	2604	1.1108	51.6167	9.0806
0.0082	43.0	2666	1.1176	51.1365	9.0968
0.0082	44.0	2728	1.1544	49.703	9.0645
0.0082	45.0	2790	1.1655	51.432	9.1935
0.0082	46.0	2852	1.1460	50.1011	9.1774
0.0082	47.0	2914	1.1377	50.0643	9.129
0.0082	48.0	2976	1.1406	50.1912	9.1129
0.0081	49.0	3038	1.1452	47.2465	9.1774
0.0081	50.0	3100	1.1532	49.9986	9.0806
0.0081	51.0	3162	1.1596	47.8461	9.0806
0.0081	52.0	3224	1.1643	48.3596	9.0968
0.0081	53.0	3286	1.1577	47.1237	9.0806
0.0081	54.0	3348	1.1599	48.6692	9.0968
0.0081	55.0	3410	1.1613	48.1806	9.0806
0.0081	56.0	3472	1.1668	47.5471	9.1613
0.0069	57.0	3534	1.1749	50.0805	9.0806
0.0069	58.0	3596	1.1784	49.3841	9.1774
0.0069	59.0	3658	1.1666	49.4183	9.0645
0.0069	60.0	3720	1.1768	47.8488	9.1774
0.0069	61.0	3782	1.1908	48.7428	9.0968
0.0069	62.0	3844	1.1882	49.2957	8.9677
0.0069	63.0	3906	1.1869	49.5255	9.0323
0.0069	64.0	3968	1.1866	48.8917	9.0161
0.0068	65.0	4030	1.1858	48.5308	9.0968
0.0068	66.0	4092	1.1951	49.2041	9.0806
0.0068	67.0	4154	1.1828	49.1255	9.0806
0.0068	68.0	4216	1.1923	48.0252	9.0484
0.0068	69.0	4278	1.1947	48.0764	9.1129
0.0068	70.0	4340	1.1927	48.2729	9.0484
0.0068	71.0	4402	1.1907	47.9908	9.129
0.0068	72.0	4464	1.1920	48.8939	9.0968
0.0062	73.0	4526	1.1939	49.0374	9.0968
0.0062	74.0	4588	1.1952	49.0374	9.0968
0.0062	75.0	4650	1.1954	49.2333	9.0323
0.0062	76.0	4712	1.1951	48.3221	9.1129
0.0062	77.0	4774	1.1971	48.3221	9.1129
0.0062	78.0	4836	1.1978	49.5615	9.1129
0.0062	79.0	4898	1.1994	48.947	9.0484
0.0062	80.0	4960	1.2009	48.0436	9.0806
0.0045	81.0	5022	1.2021	47.9908	9.129
0.0045	82.0	5084	1.2048	47.9908	9.129
0.0045	83.0	5146	1.2045	49.5615	9.0968
0.0045	84.0	5208	1.2065	49.4183	9.0968
0.0045	85.0	5270	1.2081	48.9864	9.0968
0.0045	86.0	5332	1.2131	46.327	9.0968
0.0045	87.0	5394	1.2144	47.2291	9.1452
0.0045	88.0	5456	1.2135	47.2291	9.1452
0.0047	89.0	5518	1.2163	46.8533	9.1452
0.0047	90.0	5580	1.2207	47.3713	9.1452
0.0047	91.0	5642	1.2188	47.3713	9.1452
0.0047	92.0	5704	1.2193	47.3713	9.1452
0.0047	93.0	5766	1.2188	48.9917	9.1452
0.0047	94.0	5828	1.2175	47.2291	9.1452
0.0047	95.0	5890	1.2177	48.9917	9.1452
0.0047	96.0	5952	1.2177	47.3713	9.1452
0.0043	97.0	6014	1.2165	47.3713	9.1452
0.0043	98.0	6076	1.2167	47.287	9.1774
0.0043	99.0	6138	1.2169	47.287	9.1774
0.0043	100.0	6200	1.2171	47.287	9.1774

Framework versions

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