You can deploy this model and start earning money today!

cshim-cmu/
k60

Name: k60
Availability: InStock

pytorch

marian

generated_from_trainer

apache-2.0

Views

No views yet

k60

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: 0.4673
Bleu: 1.931
Chrf: 34.6649
Gen Len: 33.6117

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
0.6855	0.09	1000	0.6880	0.6513	16.7059	49.668
0.5739	0.18	2000	0.6414	0.8782	19.8503	42.2465
0.525	0.27	3000	0.6110	0.9947	21.594	43.7022
0.4913	0.36	4000	0.5891	0.7998	23.3189	41.4658
0.4658	0.45	5000	0.5716	0.7458	23.9484	43.7586
0.4345	0.54	6000	0.5540	1.1043	25.7669	42.3431
0.4206	0.63	7000	0.5451	1.2621	26.2873	41.4437
0.3957	0.72	8000	0.5336	1.5355	27.5189	39.4135
0.3887	0.81	9000	0.5254	1.67	28.0731	38.2716
0.373	0.9	10000	0.5188	1.9531	29.1203	35.2183
0.3608	0.99	11000	0.5114	1.8166	29.5736	33.7897
0.3446	1.08	12000	0.5084	2.0298	30.218	34.2686
0.3431	1.18	13000	0.5034	2.1373	30.8485	35.0805
0.336	1.27	14000	0.4986	1.8225	31.4563	34.3471
0.3254	1.36	15000	0.4940	2.0065	31.361	34.8229
0.3224	1.45	16000	0.4930	2.0902	31.809	35.6328
0.3188	1.54	17000	0.4904	1.938	32.2144	35.33
0.313	1.63	18000	0.4865	2.0539	32.6906	34.1328
0.3101	1.72	19000	0.4817	2.175	32.8017	34.3893
0.3018	1.81	20000	0.4819	2.1489	32.9338	35.7173
0.3009	1.9	21000	0.4812	2.1229	33.0706	32.8722
0.2904	1.99	22000	0.4796	2.0758	33.2377	33.7072
0.2885	2.08	23000	0.4782	2.0332	33.5191	34.4004
0.2825	2.17	24000	0.4779	2.0486	33.6665	34.5322
0.2786	2.26	25000	0.4737	2.1297	33.725	34.834
0.2779	2.35	26000	0.4735	2.5071	34.0979	34.8612
0.2764	2.44	27000	0.4727	1.9358	33.9464	34.161
0.2712	2.53	28000	0.4734	1.7888	34.0369	35.5855
0.2735	2.62	29000	0.4739	1.9447	33.9657	34.5302
0.268	2.71	30000	0.4710	2.3687	34.0985	34.2334
0.2696	2.8	31000	0.4736	1.9553	34.2237	34.0131
0.2673	2.89	32000	0.4689	1.9401	34.2569	33.5976
0.2643	2.98	33000	0.4670	2.3772	34.7587	33.8551
0.2587	3.07	34000	0.4683	2.2288	34.7731	33.8602
0.2556	3.16	35000	0.4691	2.0693	34.4555	34.2817
0.2509	3.25	36000	0.4702	2.198	34.5518	33.4266
0.248	3.34	37000	0.4680	2.1579	34.7251	34.1881
0.2458	3.43	38000	0.4700	2.4393	34.7171	33.9437
0.2465	3.53	39000	0.4673	1.931	34.6649	33.6117

Framework versions

Transformers 4.28.1
Pytorch 2.0.0+cu117
Datasets 2.11.0
Tokenizers 0.13.3

cshim-cmu/k60

k60

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

cshim-cmu/
k60