You can deploy this model and start earning money today!

cshim-cmu/
d100-100-100-100-100-100

Name: d100-100-100-100-100-100
Availability: InStock

pytorch

marian

generated_from_trainer

apache-2.0

Views

No views yet

d100-100-100-100-100-100

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

Loss: 0.5078
Bleu: 2.0194
Chrf: 30.7967
Gen Len: 36.667

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: 20
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Bleu	Chrf	Gen Len
0.7103	0.09	1000	0.7040	0.7798	18.301	43.3581
0.6121	0.18	2000	0.6621	0.7864	20.1835	37.3672
0.5687	0.27	3000	0.6406	0.8489	21.6548	39.5191
0.545	0.36	4000	0.6233	1.0347	22.1915	38.5362
0.5275	0.45	5000	0.6106	0.9366	23.3081	39.9658
0.5007	0.54	6000	0.6003	1.0374	23.9695	38.8813
0.4925	0.63	7000	0.5943	1.0743	23.8719	44.2757
0.4709	0.72	8000	0.5825	0.9944	24.3937	49.6207
0.4671	0.81	9000	0.5790	0.9953	25.2133	41.8058
0.4535	0.9	10000	0.5742	1.0193	25.6907	34.1318
0.4416	0.99	11000	0.5669	1.2067	26.367	37.0292
0.4286	1.08	12000	0.5631	1.4208	27.3466	38.6781
0.4278	1.18	13000	0.5595	0.918	27.3633	37.5714
0.4216	1.27	14000	0.5562	1.3929	27.1099	35.5171
0.4105	1.36	15000	0.5532	1.1944	26.8407	42.3561
0.4084	1.45	16000	0.5509	1.2987	27.9877	38.5181
0.4054	1.54	17000	0.5465	1.198	28.6213	36.7565
0.3995	1.63	18000	0.5423	1.4734	28.2443	38.996
0.399	1.72	19000	0.5409	1.2935	28.3064	37.8581
0.3882	1.81	20000	0.5398	1.5491	28.4761	39.1751
0.3891	1.9	21000	0.5353	1.4344	28.8084	35.2978
0.3768	1.99	22000	0.5359	1.5729	28.425	40.1791
0.3772	2.08	23000	0.5306	1.5295	29.1184	38.5785
0.3711	2.17	24000	0.5308	1.1517	28.3518	38.3229
0.366	2.26	25000	0.5314	1.4113	29.1601	36.6952
0.3664	2.35	26000	0.5273	1.3947	28.8587	40.7958
0.3647	2.44	27000	0.5275	1.8229	29.4971	37.994
0.3579	2.53	28000	0.5235	1.4989	29.3623	39.2998
0.3619	2.62	29000	0.5278	1.6982	30.1368	34.84
0.3565	2.71	30000	0.5228	1.7473	30.2026	38.7012
0.357	2.8	31000	0.5243	1.7737	30.1197	36.2706
0.355	2.89	32000	0.5188	1.7271	30.4385	37.0262
0.3518	2.98	33000	0.5176	1.8531	30.3444	36.3038
0.3467	3.07	34000	0.5192	2.0895	30.7617	35.0734
0.3434	3.16	35000	0.5172	1.5883	29.301	39.2404
0.3373	3.25	36000	0.5164	1.8406	30.5954	34.8803
0.3339	3.34	37000	0.5151	1.6371	30.4521	38.3984
0.3314	3.43	38000	0.5158	1.8338	30.6019	35.835
0.3326	3.53	39000	0.5145	1.7891	31.1398	38.6358
0.3344	3.62	40000	0.5147	1.7449	30.3597	35.9286
0.336	3.71	41000	0.5100	1.7145	30.9239	35.9004
0.3327	3.8	42000	0.5102	2.0134	30.4394	36.0976
0.3289	3.89	43000	0.5087	2.0366	31.7571	37.9668
0.3317	3.98	44000	0.5075	1.4085	30.521	39.5352
0.327	4.07	45000	0.5109	1.8187	31.2169	38.9085
0.3229	4.16	46000	0.5107	1.9995	30.7497	36.1861
0.3243	4.25	47000	0.5078	2.0834	31.5256	35.6157
0.3207	4.34	48000	0.5078	2.0194	30.7967	36.667

Framework versions

Transformers 4.28.1
Pytorch 2.0.0+cu117
Datasets 2.11.0
Tokenizers 0.13.3

cshim-cmu/d100-100-100-100-100-100

d100-100-100-100-100-100

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

cshim-cmu/
d100-100-100-100-100-100