ailab-bio/PROTAC-Splitter-Trial-15

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: 0.0002
• Valid: 1.0
• Linker Equal: 0.9882
• Linker Graph Edit Distance: 0.1158
• Linker Tanimoto Similarity: 0.0
• E3 Graph Edit Distance Norm: inf
• Poi Graph Edit Distance: inf
• E3 Has Attachment Point(s): 1.0
• Has Three Substructures: 1.0
• Reassembly: 0.9997
• E3 Heavy Atoms Difference Norm: 0.0002
• Linker Heavy Atoms Difference Norm: -0.0027
• Linker Graph Edit Distance Norm: inf
• E3 Tanimoto Similarity: 0.0
• Poi Heavy Atoms Difference: 0.0124
• Linker Valid: 1.0
• Heavy Atoms Difference: -0.0001
• Num Fragments: 3.0
• All Ligands Equal: 0.9880
• Has All Attachment Points: 1.0
• E3 Graph Edit Distance: inf
• Linker Has Attachment Point(s): 1.0
• Poi Heavy Atoms Difference Norm: 0.0003
• Heavy Atoms Difference Norm: -0.0000
• Linker Heavy Atoms Difference: -0.0166
• Tanimoto Similarity: 0.0
• E3 Equal: 0.9951
• Poi Valid: 1.0
• E3 Valid: 1.0
• Poi Tanimoto Similarity: 0.0
• Poi Graph Edit Distance Norm: inf
• Poi Equal: 0.9930
• Poi Has Attachment Point(s): 1.0
• E3 Heavy Atoms Difference: 0.0041
• Reassembly Nostereo: 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: cosine_with_restarts
• lr_scheduler_warmup_steps: 699
• training_steps: 500000
• 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
0.0008	6.5543	70000	0.9846	0.9935	inf	inf	1.0	0.0063	0.0003	0.0	1.0	1.0	1.0	-0.0009	-0.0000	0.9859	0.1331	inf	1.0	-0.0189	-0.0030	0.0	1.0	0.0003	3.0	0.9915	inf	inf	1.0	0.0118	0.0002	0.0	1.0	0.9976	0.9978	0.0	1.0
0.0011	7.4906	80000	0.9849	0.9937	inf	inf	1.0	0.0048	0.0002	0.0	1.0	1.0	1.0	0.0326	0.0003	0.9869	inf	inf	1.0	-0.0146	-0.0026	0.0	1.0	0.0004	3.0	0.9916	inf	inf	0.9998	0.0318	0.0005	0.0	0.9998	0.9968	0.9968	0.0	0.9998
0.0006	8.4270	90000	0.9864	0.9938	inf	inf	1.0	0.0047	0.0002	0.0	1.0	1.0	1.0	0.0009	0.0000	0.9867	inf	inf	1.0	-0.0003	-0.0018	0.0	1.0	0.0003	3.0	0.9925	inf	inf	1.0	-0.0036	-0.0005	0.0	1.0	0.9990	0.9990	0.0	1.0
0.0005	9.3633	100000	0.9874	0.9944	inf	inf	1.0	0.0068	0.0003	0.0	1.0	1.0	1.0	0.0	-0.0000	0.9878	0.1165	inf	1.0	-0.0186	-0.0023	0.0	1.0	0.0003	3.0	0.9928	inf	inf	1.0	0.0118	0.0003	0.0	1.0	0.9991	0.9991	0.0	1.0
0.0004	32.7715	350000	0.0002	1.0	0.9888	0.1068	0.0	inf	inf	1.0	1.0	0.9996	0.0003	-0.0037	inf	0.0	0.0260	1.0	0.0004	3.0	0.9886	1.0	inf	1.0	0.0009	0.0000	-0.0323	0.0	0.9949	1.0	1.0	0.0	inf	0.9937	1.0	0.0066	0.9996
0.0003	33.7079	360000	0.0002	1.0	0.9880	0.1166	0.0	inf	inf	1.0	1.0	0.9994	0.0002	-0.0030	inf	0.0	0.0190	1.0	0.0021	3.0	0.9877	1.0	inf	1.0	0.0005	0.0000	-0.0203	0.0	0.9948	1.0	1.0	0.0	inf	0.9930	1.0	0.0034	0.9994
0.0003	34.6442	370000	0.0002	1.0	0.9880	inf	0.0	inf	inf	1.0	1.0	0.9995	0.0002	-0.0026	inf	0.0	0.0116	1.0	-0.0012	3.0	0.9879	1.0	inf	1.0	0.0003	-0.0000	-0.0175	0.0	0.9953	1.0	1.0	0.0	inf	0.9927	1.0	0.0047	0.9995
0.0003	35.5805	380000	0.0002	1.0	0.9884	inf	0.0	inf	inf	1.0	1.0	0.9997	0.0002	-0.0032	inf	0.0	0.0189	1.0	-0.0005	3.0	0.9883	1.0	inf	1.0	0.0005	-0.0000	-0.0239	0.0	0.9951	1.0	1.0	0.0	inf	0.9933	1.0	0.0045	0.9997
0.0003	36.5169	390000	0.0002	1.0	0.9890	0.1027	0.0	inf	inf	1.0	1.0	0.9996	0.0002	-0.0030	inf	0.0	0.0237	1.0	-0.0002	3.0	0.9887	1.0	inf	1.0	0.0008	-0.0000	-0.0285	0.0	0.9950	1.0	1.0	0.0	inf	0.9936	1.0	0.0046	0.9996
0.0003	37.4532	400000	0.0002	1.0	0.9890	0.1059	0.0	inf	inf	1.0	1.0	0.9995	0.0002	-0.0025	inf	0.0	0.0115	1.0	-0.0002	3.0	0.9887	1.0	inf	1.0	0.0002	-0.0000	-0.0148	0.0	0.9949	1.0	1.0	0.0	inf	0.9938	1.0	0.0031	0.9995
0.0003	38.3895	410000	0.0002	1.0	0.9885	0.1101	0.0	inf	inf	1.0	1.0	0.9995	0.0002	-0.0033	inf	0.0	0.0214	1.0	-0.0002	3.0	0.9882	1.0	inf	1.0	0.0006	-0.0000	-0.0250	0.0	0.9950	1.0	1.0	0.0	inf	0.9930	1.0	0.0034	0.9995
0.0003	39.3258	420000	0.0002	0.9999	0.9887	inf	0.0	inf	inf	1.0	1.0	0.9993	0.0002	-0.0026	inf	0.0	0.0184	1.0	0.0064	3.0	0.9881	1.0	inf	1.0	0.0005	0.0001	-0.0207	0.0	0.9950	0.9999	1.0	0.0	inf	0.9930	0.9999	0.0047	0.9993
0.0003	40.2622	430000	0.0002	1.0	0.9884	0.1125	0.0	inf	inf	1.0	1.0	0.9999	0.0002	-0.0034	inf	0.0	0.0236	1.0	0.0	3.0	0.9883	1.0	0.0333	1.0	0.0007	0.0	-0.0272	0.0	0.9949	1.0	1.0	0.0	inf	0.9933	1.0	0.0036	0.9999
0.0003	41.1985	440000	0.0002	1.0	0.9882	0.1105	0.0	inf	inf	1.0	1.0	0.9993	0.0002	-0.0032	0.0022	0.0	0.0229	1.0	-0.0008	3.0	0.9877	1.0	inf	1.0	0.0007	-0.0000	-0.0275	0.0	0.9943	1.0	1.0	0.0	inf	0.9930	1.0	0.0038	0.9994
0.0003	42.1348	450000	0.0002	1.0	0.9880	inf	0.0	inf	inf	1.0	1.0	0.9994	0.0002	-0.0027	inf	0.0	0.0171	1.0	0.0009	3.0	0.9877	1.0	inf	1.0	0.0005	0.0000	-0.0201	0.0	0.9949	1.0	1.0	0.0	inf	0.9929	1.0	0.0039	0.9995
0.0003	43.0712	460000	0.0002	1.0	0.9883	0.1074	0.0	inf	inf	1.0	1.0	0.9997	0.0002	-0.0033	inf	0.0	0.0244	1.0	0.0009	3.0	0.9882	1.0	inf	1.0	0.0008	0.0000	-0.0282	0.0	0.9950	1.0	1.0	0.0	inf	0.9930	1.0	0.0047	0.9997
0.0003	44.0075	470000	0.0002	1.0	0.9887	inf	0.0	inf	inf	1.0	1.0	0.9994	0.0002	-0.0032	inf	0.0	0.0232	1.0	0.0009	3.0	0.9885	1.0	inf	1.0	0.0007	0.0000	-0.0272	0.0	0.9950	1.0	1.0	0.0	inf	0.9936	1.0	0.0049	0.9994
0.0003	44.9438	480000	0.0002	1.0	0.9892	inf	0.0	inf	inf	1.0	1.0	0.9997	0.0002	-0.0037	inf	0.0	0.0295	1.0	-0.0001	3.0	0.9890	1.0	inf	1.0	0.0010	-0.0000	-0.0341	0.0	0.9951	1.0	1.0	0.0	inf	0.9938	1.0	0.0044	0.9997
0.0003	45.8801	490000	0.0002	1.0	0.9884	inf	0.0	inf	inf	1.0	1.0	0.9997	0.0002	-0.0027	inf	0.0	0.0108	1.0	-0.0001	3.0	0.9882	1.0	inf	1.0	0.0002	-0.0000	-0.0148	0.0	0.9951	1.0	1.0	0.0	inf	0.9930	1.0	0.0040	0.9997
0.0003	46.8165	500000	0.0002	1.0	0.9882	0.1158	0.0	inf	inf	1.0	1.0	0.9997	0.0002	-0.0027	inf	0.0	0.0124	1.0	-0.0001	3.0	0.9880	1.0	inf	1.0	0.0003	-0.0000	-0.0166	0.0	0.9951	1.0	1.0	0.0	inf	0.9930	1.0	0.0041	0.9997

Framework versions

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