ailab-bio/PROTAC-Splitter-EncoderDecoder-lr_reduce-rand-smiles

This model is a fine-tuned version of seyonec/ChemBERTa-zinc-base-v1 on the ailab-bio/PROTAC-Splitter-Dataset dataset. It achieves the following results on the evaluation set:

• Loss: 0.3692
• All Ligands Equal: 0.6035
• E3 Valid: 0.9927
• Linker Has Attachment Point(s): 0.9984
• E3 Graph Edit Distance: inf
• Poi Valid: 0.9629
• Reassembly: 0.6101
• E3 Has Attachment Point(s): 0.9927
• E3 Tanimoto Similarity: 0.0
• Poi Heavy Atoms Difference: 1.0907
• Linker Graph Edit Distance Norm: inf
• Heavy Atoms Difference: 4.2504
• E3 Heavy Atoms Difference Norm: 0.0208
• Has All Attachment Points: 0.9934
• Linker Heavy Atoms Difference: 0.1354
• Poi Equal: 0.7987
• Poi Graph Edit Distance: inf
• Tanimoto Similarity: 0.0
• Num Fragments: 2.9997
• Poi Heavy Atoms Difference Norm: 0.0332
• Poi Graph Edit Distance Norm: inf
• Valid: 0.9550
• Reassembly Nostereo: 0.6440
• Linker Graph Edit Distance: 15934844192634561100698130285457299401122608004817217891860480.0000
• E3 Equal: 0.8303
• Linker Heavy Atoms Difference Norm: -0.0025
• Poi Tanimoto Similarity: 0.0
• Linker Tanimoto Similarity: 0.0
• E3 Graph Edit Distance Norm: inf
• Linker Valid: 0.9984
• E3 Heavy Atoms Difference: 0.6159
• Linker Equal: 0.8640
• Poi Has Attachment Point(s): 0.9629
• Heavy Atoms Difference Norm: 0.0571
• Has Three Substructures: 0.9997

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: 5e-05
• train_batch_size: 128
• eval_batch_size: 64
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: reduce_lr_on_plateau
• lr_scheduler_warmup_steps: 400
• training_steps: 100000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	All Ligands Equal	E3 Equal	E3 Graph Edit Distance	E3 Graph Edit Distance Norm	E3 Has Attachment Point(s)	E3 Heavy Atoms Difference	E3 Heavy Atoms Difference Norm	E3 Tanimoto Similarity	E3 Valid	Has All Attachment Points	Has Three Substructures	Heavy Atoms Difference	Heavy Atoms Difference Norm	Linker Equal	Linker Graph Edit Distance	Linker Graph Edit Distance Norm	Linker Has Attachment Point(s)	Linker Heavy Atoms Difference	Linker Heavy Atoms Difference Norm	Linker Tanimoto Similarity	Linker Valid	Validation Loss	Num Fragments	Poi Equal	Poi Graph Edit Distance	Poi Graph Edit Distance Norm	Poi Has Attachment Point(s)	Poi Heavy Atoms Difference	Poi Heavy Atoms Difference Norm	Poi Tanimoto Similarity	Poi Valid	Reassembly	Reassembly Nostereo	Tanimoto Similarity	Valid
0.0107	0.4932	5000	0.5011	0.7894	inf	inf	0.9859	0.4474	0.0105	0.0	0.9859	0.9880	0.9993	4.6779	0.0617	0.7242	inf	inf	0.9978	0.5834	0.0173	0.0	0.9978	0.3008	3.0002	0.7466	inf	inf	0.9624	1.1262	0.0317	0.0	0.9624	0.5097	0.5343	0.0	0.9481
0.0058	0.7398	7500	0.5325	0.8015	inf	inf	0.9835	0.6172	0.0172	0.0	0.9835	0.9898	0.9992	4.7889	0.0626	0.7643	59313031161473085686992099539504630098717768674496038276431872.0000	inf	0.9941	0.2995	0.0039	0.0	0.9941	0.3187	2.9999	0.7586	inf	inf	0.9616	0.9886	0.0264	0.0	0.9616	0.5402	0.5643	0.0	0.9417
0.0049	0.9864	10000	0.5455	0.8076	inf	inf	0.9866	0.4690	0.0123	0.0	0.9866	0.9866	0.9990	5.7446	0.0744	0.7856	inf	inf	0.9952	0.2002	-0.0013	0.0	0.9952	0.3251	2.9997	0.7632	inf	inf	0.9458	1.7548	0.0510	0.0	0.9458	0.5548	0.5789	0.0	0.9308
0.0007	6.9047	70000	0.3657	0.5933	0.9918	0.9974	inf	0.9509	0.6012	0.9918	0.0	1.4603	inf	5.0341	0.0084	0.9904	0.1724	0.7852	inf	0.0	2.9999	0.0467	inf	0.9408	0.6375	25672804532577903995569209904347871257364201785538851047997440.0000	0.8272	0.0013	0.0	0.0	inf	0.9974	0.4047	0.8478	0.9509	0.0678	0.9999
0.0006	7.1513	72500	0.3683	0.5906	0.9896	0.9978	inf	0.9499	0.5974	0.9896	0.0	1.4326	inf	5.1042	0.0098	0.9856	0.1323	0.7870	inf	0.0	3.0001	0.0473	inf	0.9390	0.6340	inf	0.8245	-0.0034	0.0	0.0	inf	0.9978	0.4132	0.8443	0.9499	0.0684	0.9999
0.0006	7.3979	75000	0.3680	0.5893	0.9945	0.9958	inf	0.9505	0.5966	0.9945	0.0	1.3942	inf	4.8410	0.0119	0.9904	0.2242	0.7831	inf	0.0	2.9993	0.0443	inf	0.9429	0.6321	inf	0.8291	0.0022	0.0	0.0	inf	0.9958	0.3916	0.8384	0.9505	0.0649	0.9993
0.0006	7.6445	77500	0.3600	0.5914	0.9844	0.9988	inf	0.9499	0.5994	0.9844	0.0	1.5098	0.0298	5.7886	0.0231	0.9919	0.2759	0.7857	inf	0.0	3.0	0.0486	inf	0.9340	0.6316	12393767705382436411654101333133455089762028448191169471447040.0000	0.8292	0.0055	0.0	0.0	inf	0.9988	0.6907	0.8437	0.9499	0.0776	1.0
0.0006	7.8911	80000	0.3442	0.5889	0.9830	0.9964	inf	0.9527	0.5960	0.9830	0.0	1.4641	0.0292	5.3543	0.0158	0.9889	0.2057	0.7866	472733711048158617442424011016035594400917981819674736470261760.0000	0.0	2.9997	0.0450	inf	0.9352	0.6297	36296033994334280917196682173239166308017879354407269074731008.0000	0.8214	0.0011	0.0	0.0	inf	0.9964	0.5424	0.8451	0.9527	0.0723	0.9996
0.0006	8.1377	82500	0.3475	0.5930	0.9927	0.9975	inf	0.9552	0.6004	0.9927	0.0	1.4236	inf	4.7973	0.0125	0.9905	0.2185	0.7876	inf	0.0	3.0	0.0429	inf	0.9459	0.6358	inf	0.8276	0.0033	0.0	0.0	inf	0.9975	0.4410	0.8479	0.9552	0.0636	1.0
0.0006	8.3843	85000	0.3691	0.5923	0.9934	0.9977	inf	0.9467	0.5998	0.9934	0.0	1.4565	0.0290	5.2051	0.0106	0.9875	0.2453	0.7859	inf	0.0	3.0001	0.0449	inf	0.9383	0.6351	23016997167138810478786188190104988023843767118069314732687360.0000	0.8279	0.0043	0.0	0.0	inf	0.9977	0.4311	0.8380	0.9467	0.0692	0.9999
0.001	8.6309	87500	0.3526	0.5934	0.9881	0.9972	inf	0.9449	0.6006	0.9881	0.0	1.7053	inf	5.8499	0.0174	0.9930	0.2047	0.7833	inf	0.0	2.9997	0.0520	inf	0.9324	0.6352	28328611898016991803361460794751230257740758655027841832321024.0000	0.8293	-0.0009	0.0	0.0	inf	0.9972	0.6076	0.8429	0.9449	0.0772	0.9997
0.0011	8.8775	90000	0.3572	0.5955	0.9827	0.9969	inf	0.9503	0.6022	0.9827	0.0	1.5165	inf	5.9425	0.0283	0.9924	0.2921	0.7941	inf	0.0	2.9996	0.0475	inf	0.9310	0.6323	30984419263456093883630638744753399840977010019468196444110848.0000	0.8275	0.0069	0.0	0.0	inf	0.9969	0.7812	0.8484	0.9503	0.0805	0.9996
0.0003	9.1241	92500	0.3578	0.6011	0.9813	0.9979	inf	0.9595	0.6080	0.9813	0.0	1.2296	inf	5.2612	0.0331	0.9928	0.1763	0.7982	inf	0.0	2.9998	0.0392	inf	0.9396	0.6372	inf	0.8302	0.0002	0.0	0.0	inf	0.9979	0.8631	0.8592	0.9595	0.0723	0.9996
0.0003	9.3707	95000	0.3646	0.6000	0.9880	0.9979	inf	0.9614	0.6070	0.9880	0.0	1.1567	0.0261	4.6484	0.0210	0.9926	0.1959	0.7989	inf	0.0	2.9997	0.0352	inf	0.9479	0.6393	21246458923512748134264173713943065868163477339756290522480640.0000	0.8310	0.0018	0.0	0.0	inf	0.9979	0.6215	0.8579	0.9614	0.0627	0.9997
0.0003	9.6173	97500	0.3708	0.6024	0.9886	0.9982	inf	0.9598	0.6090	0.9886	0.0	1.1906	inf	4.7226	0.0237	0.9923	0.1565	0.7978	inf	0.0	2.9996	0.0370	inf	0.9478	0.6402	17705382436260623445220144761619221556802897783130242102067200.0000	0.8306	-0.0017	0.0	0.0	inf	0.9982	0.6769	0.8600	0.9598	0.0640	0.9996
0.0003	9.8639	100000	0.3692	0.6035	0.9927	0.9984	inf	0.9629	0.6101	0.9927	0.0	1.0907	inf	4.2504	0.0208	0.9934	0.1354	0.7987	inf	0.0	2.9997	0.0332	inf	0.9550	0.6440	15934844192634561100698130285457299401122608004817217891860480.0000	0.8303	-0.0025	0.0	0.0	inf	0.9984	0.6159	0.8640	0.9629	0.0571	0.9997

Framework versions

• Transformers 4.44.2
• Pytorch 2.4.1+cu121
• Datasets 3.0.0
• Tokenizers 0.19.1