cs_m2m_2e-5_50_v0.2

This model is a fine-tuned version of facebook/m2m100_1.2B on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 2.3546
• Bleu: 46.5499
• Gen Len: 19.8571

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: 16
• eval_batch_size: 16
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 50

Training results

Training Loss	Epoch	Step	Validation Loss	Bleu	Gen Len
0.0	1.0	6	2.4739	49.3833	19.8571
0.0	2.0	12	2.4967	47.4622	20.7143
0.0087	3.0	18	2.6016	47.8384	20.9524
0.0	4.0	24	2.6004	49.8858	19.9048
0.0825	5.0	30	2.4731	50.7434	19.9524
0.0	6.0	36	2.4229	45.2602	20.7619
0.0002	7.0	42	2.4148	45.5274	20.5238
0.0001	8.0	48	2.3583	47.4096	19.9524
0.0	9.0	54	2.3559	49.1212	20.1905
0.0	10.0	60	2.3610	47.0296	20.0952
0.0001	11.0	66	2.3423	47.2022	19.8571
0.0002	12.0	72	2.2938	48.5473	20.0952
0.0	13.0	78	2.2591	49.6382	19.4762
0.0001	14.0	84	2.2492	49.5102	19.6667
0.0001	15.0	90	2.2740	49.1707	19.6667
0.0	16.0	96	2.2876	48.9631	19.3333
0.0023	17.0	102	2.2842	48.7639	19.6667
0.0001	18.0	108	2.2830	45.9993	19.5238
0.0	19.0	114	2.2872	49.1391	19.7619
0.0	20.0	120	2.2893	49.1623	19.8095
0.0	21.0	126	2.2948	48.5803	20.0
0.0	22.0	132	2.3048	48.9732	20.0476
0.0	23.0	138	2.3114	49.1156	19.9524
0.0	24.0	144	2.3169	49.1156	19.9524
0.0	25.0	150	2.3202	48.4435	20.0
0.0	26.0	156	2.3227	48.4435	20.0
0.0	27.0	162	2.3236	48.4435	20.0
0.0	28.0	168	2.3244	48.4435	20.0
0.0	29.0	174	2.3268	48.4435	20.0
0.0002	30.0	180	2.3296	45.9582	19.8571
0.0	31.0	186	2.3319	45.9582	19.8571
0.0	32.0	192	2.3338	45.9582	19.8571
0.0	33.0	198	2.3401	46.8428	19.8571
0.0	34.0	204	2.3473	46.586	19.8095
0.0001	35.0	210	2.3513	46.586	19.8095
0.0	36.0	216	2.3539	48.1767	20.0476
0.0	37.0	222	2.3554	48.1966	19.9048
0.0	38.0	228	2.3563	48.1966	19.9048
0.0	39.0	234	2.3563	48.1966	19.9048
0.0	40.0	240	2.3550	46.5682	19.8095
0.0001	41.0	246	2.3541	46.5499	19.9524
0.0	42.0	252	2.3534	46.5499	19.8571
0.0001	43.0	258	2.3533	46.5499	19.8571
0.0	44.0	264	2.3533	46.5499	19.8571
0.0	45.0	270	2.3537	46.5499	19.8571
0.0001	46.0	276	2.3540	46.5499	19.8571
0.0	47.0	282	2.3543	46.5499	19.8571
0.0	48.0	288	2.3544	46.5499	19.8571
0.0	49.0	294	2.3545	46.5499	19.8571
0.0	50.0	300	2.3546	46.5499	19.8571

Framework versions

• Transformers 4.38.2
• Pytorch 2.1.0+cu121
• Datasets 2.18.0
• Tokenizers 0.15.2