MIX3_ja-en_helsinki

This model is a fine-tuned version of Helsinki-NLP/opus-mt-ja-en on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 1.4832

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: 64
• eval_batch_size: 64
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 4
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
2.8699	0.01	5000	2.3465
2.6168	0.02	10000	2.2205
2.5083	0.03	15000	2.2382
2.4359	0.04	20000	2.1670
2.3821	0.06	25000	2.1122
2.3358	0.07	30000	2.0902
2.3045	0.08	35000	2.0461
2.2782	0.09	40000	2.0290
2.2481	0.1	45000	1.9910
2.2267	0.11	50000	2.0059
2.2056	0.12	55000	1.9858
2.1903	0.13	60000	1.9725
2.173	0.15	65000	1.9797
2.154	0.16	70000	1.9654
2.1429	0.17	75000	1.9567
2.1304	0.18	80000	1.9348
2.1232	0.19	85000	1.9361
2.116	0.2	90000	1.9277
2.1016	0.21	95000	1.9193
2.0984	0.22	100000	1.9064
2.0797	0.24	105000	1.9177
2.0767	0.25	110000	1.8975
2.0642	0.26	115000	1.8782
2.0595	0.27	120000	1.9012
2.0533	0.28	125000	1.8977
2.044	0.29	130000	1.8984
2.0374	0.3	135000	1.9221
2.0305	0.31	140000	1.9243
2.02	0.32	145000	1.8773
2.0195	0.34	150000	1.8676
2.0151	0.35	155000	1.8637
2.0065	0.36	160000	1.8556
2.0037	0.37	165000	1.8399
1.9963	0.38	170000	1.8452
1.9878	0.39	175000	1.8644
1.9871	0.4	180000	1.8576
1.9779	0.41	185000	1.8509
1.9721	0.43	190000	1.8405
1.9724	0.44	195000	1.8594
1.9685	0.45	200000	1.8540
1.9634	0.46	205000	1.8694
1.9583	0.47	210000	1.8591
1.9557	0.48	215000	1.8539
1.9494	0.49	220000	1.8673
1.9484	0.5	225000	1.8021
1.9395	0.52	230000	1.8309
1.9384	0.53	235000	1.7933
1.937	0.54	240000	1.8199
1.9315	0.55	245000	1.8065
1.9276	0.56	250000	1.7857
1.9248	0.57	255000	1.8207
1.9195	0.58	260000	1.7898
1.9187	0.59	265000	1.8097
1.9138	0.6	270000	1.7909
1.9094	0.62	275000	1.7995
1.9098	0.63	280000	1.8165
1.9038	0.64	285000	1.8132
1.9034	0.65	290000	1.7951
1.899	0.66	295000	1.7880
1.8965	0.67	300000	1.7953
1.8941	0.68	305000	1.7986
1.8919	0.69	310000	1.7964
1.8875	0.71	315000	1.8041
1.884	0.72	320000	1.7764
1.8798	0.73	325000	1.8019
1.8801	0.74	330000	1.7790
1.8809	0.75	335000	1.7849
1.8736	0.76	340000	1.7800
1.8727	0.77	345000	1.7900
1.8722	0.78	350000	1.7727
1.8699	0.8	355000	1.7597
1.8672	0.81	360000	1.7824
1.8638	0.82	365000	1.7674
1.8609	0.83	370000	1.7715
1.8584	0.84	375000	1.7694
1.8568	0.85	380000	1.7776
1.8523	0.86	385000	1.7697
1.8584	0.87	390000	1.7436
1.8474	0.88	395000	1.7644
1.8492	0.9	400000	1.7732
1.8465	0.91	405000	1.7611
1.846	0.92	410000	1.7717
1.8431	0.93	415000	1.7514
1.8402	0.94	420000	1.7353
1.8398	0.95	425000	1.7720
1.8314	0.96	430000	1.7728
1.8322	0.97	435000	1.7491
1.8284	0.99	440000	1.7561
1.8301	1.0	445000	1.7499
1.8182	1.01	450000	1.7514
1.8111	1.02	455000	1.7596
1.8116	1.03	460000	1.7455
1.8098	1.04	465000	1.7495
1.809	1.05	470000	1.7446
1.8088	1.06	475000	1.7290
1.8127	1.08	480000	1.7453
1.8051	1.09	485000	1.7495
1.8026	1.1	490000	1.7453
1.8028	1.11	495000	1.7615
1.8046	1.12	500000	1.7491
1.8052	1.13	505000	1.7280
1.7997	1.14	510000	1.7482
1.7976	1.15	515000	1.7368
1.7981	1.16	520000	1.7354
1.7949	1.18	525000	1.7076
1.7943	1.19	530000	1.7020
1.7911	1.2	535000	1.7121
1.7909	1.21	540000	1.7170
1.7926	1.22	545000	1.7310
1.7856	1.23	550000	1.7218
1.7875	1.24	555000	1.7362
1.7801	1.25	560000	1.7484
1.7854	1.27	565000	1.7466
1.7799	1.28	570000	1.7248
1.7823	1.29	575000	1.7355
1.7765	1.3	580000	1.7188
1.7779	1.31	585000	1.6993
1.7751	1.32	590000	1.7154
1.7762	1.33	595000	1.7348
1.7725	1.34	600000	1.7272
1.7701	1.36	605000	1.7157
1.7644	1.37	610000	1.7161
1.7707	1.38	615000	1.6961
1.764	1.39	620000	1.6930
1.7639	1.4	625000	1.6927
1.7654	1.41	630000	1.6989
1.7623	1.42	635000	1.6892
1.7598	1.43	640000	1.6911
1.7575	1.44	645000	1.7199
1.7574	1.46	650000	1.6992
1.7526	1.47	655000	1.6981
1.7556	1.48	660000	1.6860
1.7558	1.49	665000	1.7099
1.7539	1.5	670000	1.6950
1.7454	1.51	675000	1.6999
1.748	1.52	680000	1.6871
1.7476	1.53	685000	1.6884
1.7493	1.55	690000	1.6984
1.745	1.56	695000	1.6999
1.7397	1.57	700000	1.7036
1.7429	1.58	705000	1.7223
1.7367	1.59	710000	1.7111
1.7403	1.6	715000	1.6691
1.7361	1.61	720000	1.6693
1.737	1.62	725000	1.6884
1.7347	1.63	730000	1.6641
1.7323	1.65	735000	1.6628
1.7329	1.66	740000	1.6759
1.7292	1.67	745000	1.6654
1.7275	1.68	750000	1.6738
1.7266	1.69	755000	1.6792
1.7259	1.7	760000	1.6752
1.7231	1.71	765000	1.6641
1.7238	1.72	770000	1.6676
1.7223	1.74	775000	1.6563
1.722	1.75	780000	1.6541
1.7195	1.76	785000	1.6560
1.7171	1.77	790000	1.6786
1.7187	1.78	795000	1.6434
1.7186	1.79	800000	1.6538
1.7115	1.8	805000	1.6535
1.7119	1.81	810000	1.6738
1.7106	1.83	815000	1.6597
1.7088	1.84	820000	1.6486
1.7079	1.85	825000	1.6576
1.7062	1.86	830000	1.6676
1.7084	1.87	835000	1.6449
1.7059	1.88	840000	1.6515
1.7057	1.89	845000	1.6609
1.7021	1.9	850000	1.6482
1.7005	1.91	855000	1.6653
1.6988	1.93	860000	1.6801
1.6964	1.94	865000	1.6830
1.6954	1.95	870000	1.6589
1.693	1.96	875000	1.6553
1.689	1.97	880000	1.6554
1.69	1.98	885000	1.6424
1.6893	1.99	890000	1.6628
1.6772	2.0	895000	1.6709
1.6703	2.02	900000	1.6627
1.6726	2.03	905000	1.6612
1.669	2.04	910000	1.6595
1.6696	2.05	915000	1.6427
1.6672	2.06	920000	1.6497
1.669	2.07	925000	1.6288
1.6675	2.08	930000	1.6443
1.6685	2.09	935000	1.6316
1.6671	2.11	940000	1.6451
1.6673	2.12	945000	1.6313
1.6649	2.13	950000	1.6363
1.6655	2.14	955000	1.6440
1.6637	2.15	960000	1.6238
1.6632	2.16	965000	1.6226
1.6599	2.17	970000	1.6171
1.6602	2.18	975000	1.6466
1.658	2.19	980000	1.6341
1.6571	2.21	985000	1.6500
1.6572	2.22	990000	1.6225
1.6572	2.23	995000	1.6296
1.6552	2.24	1000000	1.6437
1.6548	2.25	1005000	1.6162
1.6552	2.26	1010000	1.6223
1.6544	2.27	1015000	1.6355
1.6464	2.28	1020000	1.6250
1.652	2.3	1025000	1.6217
1.6481	2.31	1030000	1.6079
1.6466	2.32	1035000	1.6110
1.6462	2.33	1040000	1.6210
1.6448	2.34	1045000	1.5993
1.6461	2.35	1050000	1.6096
1.6396	2.36	1055000	1.6137
1.644	2.37	1060000	1.6189
1.6396	2.39	1065000	1.6211
1.639	2.4	1070000	1.6149
1.6358	2.41	1075000	1.6144
1.6356	2.42	1080000	1.6018
1.6364	2.43	1085000	1.5999
1.6352	2.44	1090000	1.6095
1.634	2.45	1095000	1.6114
1.6279	2.46	1100000	1.6156
1.6272	2.47	1105000	1.6124
1.6319	2.49	1110000	1.6046
1.6276	2.5	1115000	1.6152
1.6285	2.51	1120000	1.6129
1.6242	2.52	1125000	1.5984
1.6261	2.53	1130000	1.6116
1.623	2.54	1135000	1.6061
1.6203	2.55	1140000	1.6182
1.62	2.56	1145000	1.5887
1.6177	2.58	1150000	1.5731
1.6172	2.59	1155000	1.5990
1.6179	2.6	1160000	1.5965
1.6206	2.61	1165000	1.6000
1.6156	2.62	1170000	1.5873
1.6124	2.63	1175000	1.5899
1.613	2.64	1180000	1.5910
1.6134	2.65	1185000	1.6017
1.609	2.67	1190000	1.5822
1.6084	2.68	1195000	1.5906
1.6101	2.69	1200000	1.6218
1.6077	2.7	1205000	1.6149
1.6057	2.71	1210000	1.5994
1.6018	2.72	1215000	1.5839
1.6049	2.73	1220000	1.5864
1.6012	2.74	1225000	1.5994
1.6013	2.75	1230000	1.5821
1.5957	2.77	1235000	1.5964
1.5971	2.78	1240000	1.5897
1.5967	2.79	1245000	1.5774
1.5927	2.8	1250000	1.5861
1.5954	2.81	1255000	1.5789
1.5937	2.82	1260000	1.5739
1.5895	2.83	1265000	1.5701
1.5912	2.84	1270000	1.5622
1.5922	2.86	1275000	1.5730
1.5883	2.87	1280000	1.5775
1.5864	2.88	1285000	1.5726
1.5837	2.89	1290000	1.5679
1.5824	2.9	1295000	1.5683
1.5817	2.91	1300000	1.5508
1.5778	2.92	1305000	1.5620
1.5822	2.93	1310000	1.5556
1.5783	2.95	1315000	1.5693
1.5751	2.96	1320000	1.5781
1.5716	2.97	1325000	1.5655
1.5765	2.98	1330000	1.5528
1.5728	2.99	1335000	1.5748
1.5672	3.0	1340000	1.5597
1.5467	3.01	1345000	1.5461
1.547	3.02	1350000	1.5516
1.5462	3.03	1355000	1.5519
1.5464	3.05	1360000	1.5593
1.5457	3.06	1365000	1.5576
1.5441	3.07	1370000	1.5653
1.544	3.08	1375000	1.5662
1.5467	3.09	1380000	1.5611
1.5439	3.1	1385000	1.5635
1.5449	3.11	1390000	1.5467
1.5417	3.12	1395000	1.5495
1.5428	3.14	1400000	1.5552
1.5432	3.15	1405000	1.5347
1.5401	3.16	1410000	1.5394
1.5391	3.17	1415000	1.5497
1.539	3.18	1420000	1.5431
1.5368	3.19	1425000	1.5479
1.5365	3.2	1430000	1.5513
1.5327	3.21	1435000	1.5467
1.5337	3.23	1440000	1.5477
1.5317	3.24	1445000	1.5398
1.5315	3.25	1450000	1.5481
1.532	3.26	1455000	1.5385
1.5312	3.27	1460000	1.5520
1.5328	3.28	1465000	1.5423
1.5288	3.29	1470000	1.5489
1.5271	3.3	1475000	1.5395
1.5273	3.31	1480000	1.5335
1.5235	3.33	1485000	1.5381
1.5224	3.34	1490000	1.5289
1.5206	3.35	1495000	1.5331
1.5189	3.36	1500000	1.5343
1.5152	3.37	1505000	1.5246
1.5225	3.38	1510000	1.5280
1.5168	3.39	1515000	1.5315
1.5161	3.4	1520000	1.5284
1.5111	3.42	1525000	1.5278
1.5154	3.43	1530000	1.5148
1.515	3.44	1535000	1.5286
1.5117	3.45	1540000	1.5291
1.5099	3.46	1545000	1.5320
1.5097	3.47	1550000	1.5323
1.5075	3.48	1555000	1.5157
1.5059	3.49	1560000	1.5214
1.5011	3.51	1565000	1.5199
1.5074	3.52	1570000	1.5114
1.5033	3.53	1575000	1.5145
1.5009	3.54	1580000	1.5184
1.4994	3.55	1585000	1.5125
1.5041	3.56	1590000	1.5048
1.5002	3.57	1595000	1.5156
1.4967	3.58	1600000	1.5176
1.4923	3.59	1605000	1.5128
1.495	3.61	1610000	1.5188
1.4929	3.62	1615000	1.5149
1.4921	3.63	1620000	1.5097
1.4916	3.64	1625000	1.5161
1.4852	3.65	1630000	1.5134
1.4881	3.66	1635000	1.5101
1.4873	3.67	1640000	1.5027
1.4911	3.68	1645000	1.4968
1.488	3.7	1650000	1.4962
1.4842	3.71	1655000	1.5030
1.4829	3.72	1660000	1.5041
1.4816	3.73	1665000	1.5076
1.479	3.74	1670000	1.5029
1.4768	3.75	1675000	1.5053
1.4769	3.76	1680000	1.5026
1.4781	3.77	1685000	1.5016
1.4781	3.79	1690000	1.5034
1.4777	3.8	1695000	1.4976
1.4736	3.81	1700000	1.5002
1.4715	3.82	1705000	1.4995
1.4716	3.83	1710000	1.4996
1.4648	3.84	1715000	1.4952
1.4711	3.85	1720000	1.4934
1.4682	3.86	1725000	1.4965
1.4659	3.87	1730000	1.4932
1.4689	3.89	1735000	1.4920
1.4656	3.9	1740000	1.4910
1.4666	3.91	1745000	1.4893
1.4611	3.92	1750000	1.4888
1.4623	3.93	1755000	1.4898
1.4637	3.94	1760000	1.4909
1.4585	3.95	1765000	1.4858
1.4586	3.96	1770000	1.4847
1.4579	3.98	1775000	1.4841
1.458	3.99	1780000	1.4840
1.4572	4.0	1785000	1.4832

Framework versions

• Transformers 4.19.2
• Pytorch 1.11.0+cu113
• Datasets 2.2.2
• Tokenizers 0.12.1