esm2_t6_8M_UR50D_ft_peft-IA3_Aerin_Yang_et_al_2023_prepared-dataset

This model is a fine-tuned version of facebook/esm2_t6_8M_UR50D on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 70.9898
• Rmse: 21.7326
• Mae: 17.9645
• Spearmanr Corr: 0.8508
• Spearmanr Corr P Value: 0.0000
• Pearsonr Corr: 0.9197
• Pearsonr Corr P Value: 0.0000
• Spearmanr Corr Of Deltas: 0.8774
• Spearmanr Corr Of Deltas P Value: 0.0
• Pearsonr Corr Of Deltas: 0.9193
• Pearsonr Corr Of Deltas P Value: 0.0
• Ranking F1 Score: 0.7690
• Ranking Mcc: 0.6009
• Rmse Enriched: 7.7540
• Mae Enriched: 6.8827
• Spearmanr Corr Enriched: 0.5429
• Spearmanr Corr Enriched P Value: 0.0000
• Pearsonr Corr Enriched: 0.0133
• Pearsonr Corr Enriched P Value: 0.8018
• Spearmanr Corr Of Deltas Enriched: 0.4375
• Spearmanr Corr Of Deltas Enriched P Value: 0.0
• Pearsonr Corr Of Deltas Enriched: 0.0155
• Pearsonr Corr Of Deltas Enriched P Value: 0.0001
• Ranking F1 Score Enriched: 0.6552
• Ranking Mcc Enriched: 0.3704
• Classification Thresh: 0.2
• Mcc: 0.8953
• F1 Score: 0.9508
• Acc: 0.9478
• Auc: 0.9772
• Precision: 0.9471
• Recall: 0.9482

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: 0.00033
• train_batch_size: 32
• eval_batch_size: 32
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 64
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 200
• num_epochs: 100
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Rmse	Mae	Spearmanr Corr	Spearmanr Corr P Value	Pearsonr Corr	Pearsonr Corr P Value	Spearmanr Corr Of Deltas	Spearmanr Corr Of Deltas P Value	Pearsonr Corr Of Deltas	Pearsonr Corr Of Deltas P Value	Ranking F1 Score	Ranking Mcc	Rmse Enriched	Mae Enriched	Spearmanr Corr Enriched	Spearmanr Corr Enriched P Value	Pearsonr Corr Enriched	Pearsonr Corr Enriched P Value	Spearmanr Corr Of Deltas Enriched	Spearmanr Corr Of Deltas Enriched P Value	Pearsonr Corr Of Deltas Enriched	Pearsonr Corr Of Deltas Enriched P Value	Ranking F1 Score Enriched	Ranking Mcc Enriched	Classification Thresh	Mcc	F1 Score	Acc	Auc	Precision	Recall
460.7964	1.0	42	434.4817	18.6715	14.1108	0.7207	0.0000	0.7197	0.0000	0.7023	0.0	0.7183	0.0	0.7023	0.4735	2.6853	2.6581	0.2823	0.0000	-0.0060	0.9101	0.2369	0.0	-0.0062	0.1180	0.5573	0.1874	0.2	0.6429	0.8116	0.8149	0.9130	0.8225	0.8204
401.283	2.0	84	356.9070	19.0315	14.1069	0.7920	0.0000	0.8101	0.0000	0.8001	0.0	0.8092	0.0	0.7410	0.5476	1.5232	0.9534	0.4040	0.0000	0.0115	0.8284	0.3485	0.0	0.0136	0.0006	0.6080	0.2772	0.5	0.7784	0.8997	0.8896	0.9490	0.8916	0.8868
296.2103	3.0	126	266.6812	19.6865	15.2607	0.7767	0.0000	0.8307	0.0000	0.7993	0.0	0.8300	0.0	0.7236	0.5142	2.7693	2.4028	0.2798	0.0000	0.0089	0.8660	0.2542	0.0	0.0111	0.0048	0.5624	0.1953	0.5	0.7872	0.9034	0.8940	0.9518	0.8956	0.8917
276.8175	4.0	168	264.8241	21.4862	17.0255	0.6996	0.0000	0.7229	0.0000	0.7083	0.0	0.7216	0.0	0.6880	0.4460	4.7746	4.6865	0.0538	0.3096	-0.0040	0.9404	0.0691	0.0000	-0.0033	0.4020	0.4815	0.0470	0.5	0.6588	0.8512	0.8179	0.9274	0.8546	0.8060
186.8575	5.0	210	132.7174	21.4535	18.0706	0.8269	0.0000	0.8739	0.0000	0.8452	0.0	0.8734	0.0	0.7529	0.5703	7.6455	7.4059	0.4745	0.0000	0.0035	0.9477	0.3586	0.0	0.0054	0.1720	0.6306	0.3234	0.2	0.8445	0.9268	0.9224	0.9664	0.9216	0.9228
143.6457	6.0	252	129.9129	22.7771	19.2314	0.8202	0.0000	0.8590	0.0000	0.8268	0.0	0.8584	0.0	0.7534	0.5712	9.4125	9.2343	0.4869	0.0000	0.0092	0.8619	0.3867	0.0	0.0097	0.0143	0.6324	0.3272	0.4	0.8111	0.9138	0.9060	0.9608	0.9070	0.9041
121.6148	7.0	294	98.5414	23.5646	20.1629	0.8314	0.0000	0.8928	0.0000	0.8527	0.0	0.8924	0.0	0.7544	0.5733	11.1044	10.8920	0.4789	0.0000	0.0106	0.8414	0.3919	0.0	0.0098	0.0133	0.6331	0.3247	0.2	0.8619	0.9361	0.9313	0.9693	0.9312	0.9308
130.6186	8.0	336	83.4085	22.8860	19.5197	0.8440	0.0000	0.9093	0.0000	0.8670	0.0	0.9089	0.0	0.7628	0.5891	10.2468	9.8949	0.5147	0.0000	0.0144	0.7863	0.3918	0.0	0.0172	0.0000	0.6469	0.3523	0.4	0.8803	0.9438	0.9403	0.9747	0.9396	0.9406
111.8285	9.0	378	126.4507	22.5877	18.9855	0.8336	0.0000	0.8619	0.0000	0.8273	0.0	0.8613	0.0	0.7563	0.5769	8.6582	8.4736	0.4969	0.0000	0.0143	0.7865	0.4101	0.0	0.0162	0.0000	0.6403	0.3378	0.5	0.8169	0.9176	0.9075	0.9691	0.9137	0.9033
113.5487	10.0	420	96.3994	21.4343	18.1736	0.8491	0.0000	0.8923	0.0000	0.8688	0.0	0.8919	0.0	0.7643	0.5922	8.7046	8.0078	0.5176	0.0000	0.0177	0.7377	0.3835	0.0	0.0144	0.0003	0.6488	0.3507	0.4	0.8381	0.9179	0.9164	0.9787	0.9182	0.9199
123.4506	11.0	462	87.6575	23.5949	20.1224	0.8338	0.0000	0.9054	0.0000	0.8528	0.0	0.9051	0.0	0.7559	0.5760	10.7717	10.5746	0.4567	0.0000	0.0130	0.8061	0.3812	0.0	0.0147	0.0002	0.6241	0.3075	0.2	0.8800	0.9449	0.9403	0.9745	0.9409	0.9391
96.0486	12.0	504	128.6021	20.4876	17.3638	0.8378	0.0000	0.8470	0.0000	0.8524	0.0	0.8463	0.0	0.7622	0.5881	8.8083	7.5890	0.4958	0.0000	0.0487	0.3579	0.3968	0.0	0.0462	0.0000	0.6410	0.3376	0.01	0.7802	0.8740	0.8791	0.9739	0.8937	0.8865
109.1046	13.0	546	84.4819	23.2523	19.7547	0.8349	0.0000	0.9081	0.0000	0.8602	0.0	0.9078	0.0	0.7564	0.5769	10.6298	10.2565	0.4339	0.0000	0.0109	0.8364	0.3583	0.0	0.0140	0.0004	0.6131	0.2904	0.2	0.8892	0.9481	0.9448	0.9776	0.9442	0.9450
94.8759	14.0	588	78.4801	22.0612	18.5745	0.8409	0.0000	0.9109	0.0000	0.8682	0.0	0.9106	0.0	0.7568	0.5778	9.1375	8.3497	0.4721	0.0000	0.0105	0.8433	0.3769	0.0	0.0114	0.0040	0.6252	0.3134	0.2	0.8881	0.9456	0.9433	0.9774	0.9429	0.9451
84.9398	15.0	630	81.5671	21.6981	18.0842	0.8384	0.0000	0.9089	0.0000	0.8614	0.0	0.9086	0.0	0.7581	0.5802	7.6759	7.0524	0.4686	0.0000	0.0030	0.9549	0.3807	0.0	0.0042	0.2859	0.6245	0.3111	0.4	0.8799	0.9444	0.9403	0.9765	0.9402	0.9398
89.9718	16.0	672	96.1698	21.6908	17.9853	0.8413	0.0000	0.8914	0.0000	0.8532	0.0	0.8909	0.0	0.7591	0.5823	7.3056	6.7279	0.4899	0.0000	0.0175	0.7407	0.4178	0.0	0.0199	0.0000	0.6384	0.3335	0.4	0.8623	0.9372	0.9313	0.9750	0.9328	0.9295
93.6391	17.0	714	75.8101	22.3371	18.6930	0.8431	0.0000	0.9153	0.0000	0.8726	0.0	0.9149	0.0	0.7627	0.5889	9.3630	8.5478	0.4792	0.0000	0.0124	0.8146	0.3869	0.0	0.0133	0.0007	0.6338	0.3261	0.1	0.8881	0.9456	0.9433	0.9773	0.9429	0.9451
74.3313	18.0	756	70.9898	21.7326	17.9645	0.8508	0.0000	0.9197	0.0000	0.8774	0.0	0.9193	0.0	0.7690	0.6009	7.7540	6.8827	0.5429	0.0000	0.0133	0.8018	0.4375	0.0	0.0155	0.0001	0.6552	0.3704	0.2	0.8953	0.9508	0.9478	0.9772	0.9471	0.9482

Framework versions

• Transformers 4.45.1
• Pytorch 2.0.0+cu117
• Datasets 2.21.0
• Tokenizers 0.20.3