esm2_t12_35M_UR50D-finetuned-rep7868aav2-v1

This model is a fine-tuned version of facebook/esm2_t12_35M_UR50D on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.0505
• Spearmanr: 0.7430

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: 32
• 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	Spearmanr
0.1227	1.0	590	0.1132	0.3426
0.117	2.0	1180	0.1105	0.3945
0.1177	3.0	1770	0.1150	0.2857
0.1143	4.0	2360	0.1107	0.3518
0.1176	5.0	2950	0.1133	0.3774
0.117	6.0	3540	0.1115	0.3230
0.1178	7.0	4130	0.1107	0.4432
0.1146	8.0	4720	0.1105	0.4608
0.1182	9.0	5310	0.1105	0.4004
0.114	10.0	5900	0.1122	0.5124
0.115	11.0	6490	0.1108	0.5551
0.1152	12.0	7080	0.1091	0.5391
0.1109	13.0	7670	0.0705	0.6234
0.0888	14.0	8260	0.0633	0.6810
0.0832	15.0	8850	0.0584	0.6952
0.0705	16.0	9440	0.0585	0.7102
0.0635	17.0	10030	0.0581	0.7167
0.0606	18.0	10620	0.0525	0.7326
0.0562	19.0	11210	0.0555	0.7403
0.055	20.0	11800	0.0505	0.7430

Framework versions

• Transformers 4.35.2
• Pytorch 2.1.0+cu121
• Datasets 2.17.1
• Tokenizers 0.15.2