ailab-bio/PROTAC-Splitter-Trial-15-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: 0.0002
• Tanimoto Similarity: 0.0
• E3 Graph Edit Distance Norm: inf
• Reassembly: 0.9994
• Poi Tanimoto Similarity: 0.0
• E3 Tanimoto Similarity: 0.0
• E3 Valid: 1.0
• Linker Heavy Atoms Difference: -0.0152
• E3 Heavy Atoms Difference: 0.0073
• Linker Valid: 1.0
• Linker Equal: 0.9885
• Has All Attachment Points: 1.0
• Reassembly Nostereo: 0.9994
• Num Fragments: 3.0
• E3 Equal: 0.9950
• Linker Heavy Atoms Difference Norm: -0.0024
• Linker Graph Edit Distance Norm: inf
• Linker Has Attachment Point(s): 1.0
• Poi Graph Edit Distance: inf
• E3 Heavy Atoms Difference Norm: 0.0003
• Linker Graph Edit Distance: inf
• Poi Graph Edit Distance Norm: inf
• All Ligands Equal: 0.9881
• Poi Heavy Atoms Difference: 0.0141
• Heavy Atoms Difference Norm: 0.0001
• Linker Tanimoto Similarity: 0.0
• Has Three Substructures: 1.0
• Valid: 0.9999
• E3 Has Attachment Point(s): 1.0
• E3 Graph Edit Distance: inf
• Poi Valid: 0.9999
• Heavy Atoms Difference: 0.0101
• Poi Has Attachment Point(s): 0.9999
• Poi Heavy Atoms Difference Norm: 0.0001
• Poi Equal: 0.9929

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
0.0011	6.5543	70000	0.9852	0.9940	inf	inf	1.0	0.0040	0.0002	0.0	1.0	0.9999	1.0	0.0052	0.0001	0.9861	inf	inf	1.0	-0.0203	-0.0027	0.0	1.0	0.0003	3.0	0.9920	inf	inf	0.9999	0.0184	0.0005	0.0	0.9999	0.9978	0.9981	0.0	0.9999
0.0009	7.4906	80000	0.9868	0.9943	inf	inf	1.0	0.0044	0.0002	0.0	1.0	1.0	1.0	0.0057	0.0001	0.9875	0.1061	0.0021	1.0	-0.0155	-0.0023	0.0	1.0	0.0003	3.0	0.9927	inf	inf	1.0	0.0110	0.0002	0.0	1.0	0.9986	0.9986	0.0	1.0
0.0006	16.3858	350000	0.0002	0.0	inf	0.9994	0.0	0.0	1.0	-0.0213	0.0055	1.0	0.9880	1.0	0.9994	3.0	0.9949	-0.0027	inf	1.0	inf	0.0002	0.1093	inf	0.9877	0.0156	0.0001	0.0	1.0	1.0	1.0	inf	1.0	0.0055	1.0	0.0005	0.9926
0.0005	16.8539	360000	0.0002	0.0	inf	0.9989	0.0	0.0	1.0	-0.0115	0.0042	1.0	0.9880	1.0	0.9990	3.0	0.9948	-0.0024	inf	1.0	inf	0.0002	0.1076	inf	0.9875	0.0081	0.0000	0.0	1.0	1.0	1.0	inf	1.0	0.0008	1.0	0.0001	0.9928
0.0005	17.3221	370000	0.0003	0.0	inf	0.9986	0.0	0.0	1.0	-0.0272	0.0080	0.9999	0.9870	1.0	0.9989	3.0	0.9947	-0.0034	inf	0.9999	inf	0.0003	716948666475480386420205582353984848982468782380308030816256.0000	inf	0.9864	0.0214	0.0001	0.0	1.0	0.9999	1.0	inf	1.0	0.0054	1.0	0.0006	0.9923
0.0006	17.7903	380000	0.0002	0.0	inf	0.9990	0.0	0.0	1.0	-0.0037	0.0063	1.0	0.9876	1.0	0.9990	3.0	0.9951	-0.0022	inf	1.0	inf	0.0003	inf	inf	0.9873	-0.0010	0.0000	0.0	1.0	1.0	1.0	inf	1.0	0.0016	1.0	-0.0004	0.9925
0.0005	18.2584	390000	0.0003	0.0	inf	0.9991	0.0	0.0	1.0	-0.0174	0.0053	1.0	0.9876	1.0	0.9991	3.0	0.9947	-0.0023	inf	1.0	inf	0.0002	inf	inf	0.9873	0.0125	0.0000	0.0	1.0	1.0	1.0	inf	1.0	0.0004	1.0	0.0003	0.9928
0.0005	18.7266	400000	0.0002	0.0	inf	0.9989	0.0	0.0	1.0	-0.0166	0.0061	1.0	0.9874	1.0	0.9990	3.0	0.9946	-0.0030	inf	1.0	inf	0.0003	inf	inf	0.9869	0.0121	0.0000	0.0	1.0	1.0	1.0	inf	1.0	0.0016	1.0	0.0002	0.9922
0.0005	19.1948	410000	0.0002	0.0	inf	0.9991	0.0	0.0	1.0	-0.0104	0.0074	1.0	0.9872	1.0	0.9991	3.0	0.9949	-0.0024	inf	1.0	inf	0.0003	0.1232	inf	0.9868	0.0052	0.0002	0.0	1.0	0.9999	1.0	inf	0.9999	0.0121	0.9999	0.0001	0.9918
0.0005	19.6629	420000	0.0002	0.0	inf	0.9994	0.0	0.0	1.0	-0.0229	0.0060	1.0	0.9882	1.0	0.9994	3.0	0.9951	-0.0029	inf	1.0	inf	0.0003	inf	inf	0.9880	0.0179	0.0000	0.0	1.0	1.0	1.0	inf	1.0	0.0010	1.0	0.0005	0.9928
0.0005	20.1311	430000	0.0003	0.0	inf	0.9988	0.0	0.0	1.0	-0.0036	0.0048	1.0	0.9867	0.9999	0.9988	3.0	0.9943	-0.0026	inf	1.0	inf	0.0002	inf	inf	0.9862	0.0016	0.0001	0.0	1.0	0.9999	1.0	inf	0.9999	0.0047	0.9999	-0.0003	0.9921
0.0005	20.5993	440000	0.0002	0.0	inf	0.9991	0.0	0.0	1.0	-0.0156	0.0054	1.0	0.9880	1.0	0.9991	3.0	0.9949	-0.0028	inf	1.0	inf	0.0002	0.1159	inf	0.9875	0.0099	-0.0000	0.0	1.0	1.0	1.0	inf	1.0	-0.0003	1.0	0.0002	0.9925
0.0005	21.0674	450000	0.0003	0.0	inf	0.9994	0.0	0.0	1.0	-0.0239	0.0054	1.0	0.9880	1.0	0.9994	3.0	0.9947	-0.0031	inf	1.0	inf	0.0002	0.1095	inf	0.9877	0.0193	0.0000	0.0	1.0	1.0	1.0	inf	1.0	0.0009	1.0	0.0005	0.9928
0.0004	21.5356	460000	0.0002	0.0	inf	0.9990	0.0	0.0	1.0	-0.0338	0.0057	1.0	0.9883	1.0	0.9991	3.0	0.9947	-0.0039	inf	1.0	inf	0.0003	0.1085	inf	0.9877	0.0296	0.0001	0.0	1.0	0.9999	1.0	inf	0.9999	0.0054	0.9999	0.0010	0.9932
0.0004	22.0037	470000	0.0002	0.0	inf	0.9997	0.0	0.0	1.0	-0.0293	0.0052	1.0	0.9885	1.0	0.9997	3.0	0.9951	-0.0032	inf	1.0	inf	0.0003	inf	inf	0.9884	0.0239	-0.0000	0.0	1.0	1.0	1.0	inf	1.0	-0.0002	1.0	0.0008	0.9933
0.0005	22.4719	480000	0.0003	0.0	inf	0.9993	0.0	0.0	1.0	-0.0159	0.0067	1.0	0.9881	1.0	0.9993	3.0	0.9947	-0.0028	inf	1.0	inf	0.0003	inf	inf	0.9878	0.0080	-0.0000	0.0	1.0	1.0	1.0	inf	1.0	-0.0012	1.0	0.0001	0.9933
0.0004	22.9401	490000	0.0003	0.0	inf	0.9993	0.0	0.0	1.0	-0.0014	0.0061	1.0	0.9879	1.0	0.9994	3.0	0.9948	-0.0021	inf	1.0	inf	0.0003	0.1189	inf	0.9877	-0.0053	0.0001	0.0	1.0	1.0	1.0	inf	1.0	0.0052	1.0	-0.0005	0.9930
0.0005	23.4082	500000	0.0002	0.0	inf	0.9994	0.0	0.0	1.0	-0.0152	0.0073	1.0	0.9885	1.0	0.9994	3.0	0.9950	-0.0024	inf	1.0	inf	0.0003	inf	inf	0.9881	0.0141	0.0001	0.0	1.0	0.9999	1.0	inf	0.9999	0.0101	0.9999	0.0001	0.9929

Framework versions

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