esm2_t33_650M_UR50D-finetuned-localization

This model is a fine-tuned version of facebook/esm2_t33_650M_UR50D on the None dataset. It achieves the following results on the evaluation set:

• Loss: 1.0689
• Rmse: 1.0339

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: 2
• eval_batch_size: 2
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 25

Training results

Training Loss	Epoch	Step	Validation Loss	Rmse
No log	1.0	226	1.2216	1.1052
No log	2.0	452	1.7920	1.3387
1.7878	3.0	678	1.0784	1.0385
1.7878	4.0	904	1.4254	1.1939
1.2236	5.0	1130	1.5014	1.2253
1.2236	6.0	1356	1.3869	1.1777
0.6751	7.0	1582	0.9855	0.9927
0.6751	8.0	1808	1.1011	1.0493
0.2989	9.0	2034	1.3254	1.1512
0.2989	10.0	2260	1.1216	1.0590
0.2989	11.0	2486	1.1718	1.0825
0.1584	12.0	2712	1.0833	1.0408
0.1584	13.0	2938	1.0868	1.0425
0.0783	14.0	3164	1.0736	1.0362
0.0783	15.0	3390	1.0607	1.0299
0.0467	16.0	3616	1.0792	1.0388
0.0467	17.0	3842	1.0528	1.0261
0.0199	18.0	4068	1.0405	1.0201
0.0199	19.0	4294	1.0931	1.0455
0.0129	20.0	4520	1.0766	1.0376
0.0129	21.0	4746	1.0486	1.0240
0.0129	22.0	4972	1.0801	1.0393
0.0086	23.0	5198	1.0636	1.0313
0.0086	24.0	5424	1.0675	1.0332
0.0032	25.0	5650	1.0689	1.0339

Framework versions

• Transformers 4.42.4
• Pytorch 2.3.1
• Datasets 2.20.0
• Tokenizers 0.19.1