ailab-bio/PROTAC-Splitter-Trial-18-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
• Linker Heavy Atoms Difference: -0.0081
• Has Three Substructures: 1.0
• Linker Equal: 0.9877
• Poi Graph Edit Distance Norm: inf
• Linker Heavy Atoms Difference Norm: -0.0023
• E3 Valid: 1.0
• Poi Tanimoto Similarity: 0.0
• Tanimoto Similarity: 0.0
• Poi Heavy Atoms Difference: 0.0036
• E3 Heavy Atoms Difference: 0.0054
• E3 Tanimoto Similarity: 0.0
• Reassembly Nostereo: 0.9994
• E3 Has Attachment Point(s): 1.0
• E3 Graph Edit Distance: inf
• Poi Has Attachment Point(s): 1.0
• Linker Valid: 1.0
• E3 Heavy Atoms Difference Norm: 0.0002
• All Ligands Equal: 0.9875
• Poi Valid: 1.0
• Reassembly: 0.9994
• Poi Graph Edit Distance: inf
• Heavy Atoms Difference: 0.0009
• E3 Equal: 0.9948
• Linker Has Attachment Point(s): 1.0
• Linker Tanimoto Similarity: 0.0
• Poi Equal: 0.9925
• Poi Heavy Atoms Difference Norm: -0.0001
• Num Fragments: 3.0
• Heavy Atoms Difference Norm: 0.0000
• Linker Graph Edit Distance: inf
• E3 Graph Edit Distance Norm: inf
• Valid: 1.0
• Linker Graph Edit Distance Norm: inf
• Has All Attachment Points: 1.0

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: 100
• 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
0.0005	6.5543	70000	0.9874	0.9943	inf	inf	1.0	0.0052	0.0002	0.0	1.0	1.0	1.0	0.0033	0.0000	0.9881	716948666475480386420205582353984848982468782380308030816256.0000	inf	0.9999	-0.0227	-0.0031	0.0	0.9999	0.0003	3.0	0.9933	inf	inf	1.0	0.0182	0.0006	0.0	1.0	0.9984	0.9986	0.0	0.9999
0.0004	16.3858	350000	0.0002	-0.0228	1.0	0.9884	inf	-0.0029	1.0	0.0	0.0	0.0184	0.0040	0.0	0.9994	1.0	inf	1.0	1.0	0.0002	0.9880	1.0	0.9994	inf	-0.0004	0.9948	1.0	0.0	0.9933	0.0006	3.0	-0.0000	inf	inf	1.0	inf	1.0
0.0004	16.8539	360000	0.0002	-0.0243	1.0	0.9878	inf	-0.0029	1.0	0.0	0.0	0.0197	0.0050	0.0	0.9994	1.0	inf	1.0	1.0	0.0002	0.9875	1.0	0.9994	inf	0.0004	0.9948	1.0	0.0	0.9929	0.0006	3.0	0.0000	0.1189	inf	1.0	inf	1.0
0.0004	17.3221	370000	0.0002	-0.0290	1.0	0.9884	inf	-0.0033	1.0	0.0	0.0	0.0249	0.0044	0.0	0.9994	1.0	inf	1.0	1.0	0.0002	0.9880	1.0	0.9993	inf	0.0003	0.9948	1.0	0.0	0.9933	0.0008	3.0	0.0000	inf	inf	1.0	inf	1.0
0.0004	17.7903	380000	0.0002	-0.0218	1.0	0.9877	inf	-0.0029	1.0	0.0	0.0	0.0168	0.0048	0.0	0.9996	1.0	0.0399	1.0	1.0	0.0002	0.9876	1.0	0.9996	inf	-0.0001	0.9946	1.0	0.0	0.9931	0.0004	3.0	-0.0000	inf	inf	1.0	inf	1.0
0.0004	18.2584	390000	0.0002	-0.0115	1.0	0.9885	inf	-0.0022	1.0	0.0	0.0	0.0069	0.0045	0.0	0.9996	1.0	inf	1.0	1.0	0.0002	0.9884	1.0	0.9996	inf	-0.0001	0.9951	1.0	0.0	0.9933	0.0001	3.0	-0.0000	0.1098	inf	1.0	inf	1.0
0.0004	18.7266	400000	0.0002	-0.0264	1.0	0.9882	inf	-0.0030	1.0	0.0	0.0	0.0211	0.0046	0.0	0.9994	1.0	inf	1.0	1.0	0.0002	0.9880	1.0	0.9994	inf	-0.0007	0.9950	1.0	0.0	0.9931	0.0006	3.0	-0.0000	0.1136	inf	1.0	inf	1.0
0.0003	19.1948	410000	0.0002	-0.0184	1.0	0.9885	inf	-0.0028	1.0	0.0	0.0	0.0134	0.0056	0.0	0.9995	1.0	inf	1.0	0.9999	0.0003	0.9883	1.0	0.9995	inf	0.0042	0.9951	0.9999	0.0	0.9933	0.0003	3.0	0.0001	716948666475480386420205582353984848982468782380308030816256.0000	inf	0.9999	inf	1.0
0.0003	19.6629	420000	0.0002	-0.0088	1.0	0.9878	inf	-0.0024	1.0	0.0	0.0	0.0042	0.0047	0.0	0.9995	1.0	inf	1.0	1.0	0.0002	0.9876	1.0	0.9995	inf	0.0001	0.9948	1.0	0.0	0.9928	-0.0001	3.0	0.0000	inf	inf	1.0	inf	1.0
0.0003	20.1311	430000	0.0002	-0.0211	1.0	0.9884	inf	-0.0029	1.0	0.0	0.0	0.0145	0.0044	0.0	0.9994	1.0	inf	1.0	1.0	0.0002	0.9881	1.0	0.9994	inf	-0.0023	0.9946	1.0	0.0	0.9935	0.0004	3.0	-0.0000	inf	inf	1.0	inf	1.0
0.0003	20.5993	440000	0.0002	-0.0241	1.0	0.9884	inf	-0.0031	1.0	0.0	0.0	0.0196	0.0052	0.0	0.9999	1.0	inf	1.0	1.0	0.0002	0.9882	1.0	0.9999	inf	0.0006	0.9953	1.0	0.0	0.9927	0.0005	3.0	0.0000	inf	inf	1.0	inf	1.0
0.0003	21.0674	450000	0.0002	-0.0252	1.0	0.9879	inf	-0.0032	1.0	0.0	0.0	0.0194	0.0058	0.0	0.9995	1.0	inf	1.0	1.0	0.0003	0.9877	1.0	0.9995	inf	0.0	0.9950	1.0	0.0	0.9928	0.0005	3.0	-0.0000	0.1149	inf	1.0	inf	1.0
0.0003	21.5356	460000	0.0002	-0.0274	1.0	0.9893	inf	-0.0029	1.0	0.0	0.0	0.0236	0.0049	0.0	0.9997	1.0	inf	1.0	1.0	0.0002	0.9892	1.0	0.9997	inf	0.0011	0.9952	1.0	0.0	0.9940	0.0008	3.0	0.0000	inf	inf	1.0	inf	1.0
0.0003	22.0037	470000	0.0002	-0.0216	1.0	0.9887	inf	-0.0031	1.0	0.0	0.0	0.0156	0.0058	0.0	0.9995	1.0	inf	1.0	1.0	0.0003	0.9885	1.0	0.9995	inf	-0.0001	0.9951	1.0	0.0	0.9933	0.0003	3.0	-0.0000	inf	inf	1.0	inf	1.0
0.0003	22.4719	480000	0.0002	-0.0238	1.0	0.9875	inf	-0.0029	1.0	0.0	0.0	0.0168	0.0070	0.0	0.9996	1.0	inf	1.0	1.0	0.0003	0.9872	1.0	0.9996	inf	0.0	0.9949	1.0	0.0	0.9920	0.0005	3.0	-0.0000	0.1162	inf	1.0	inf	1.0
0.0003	22.9401	490000	0.0002	-0.0244	1.0	0.9884	inf	-0.0031	1.0	0.0	0.0	0.0199	0.0051	0.0	0.9995	1.0	inf	1.0	1.0	0.0002	0.9881	1.0	0.9994	inf	0.0068	0.9948	1.0	0.0	0.9933	0.0005	3.0	0.0001	inf	inf	1.0	inf	1.0
0.0003	23.4082	500000	0.0002	-0.0081	1.0	0.9877	inf	-0.0023	1.0	0.0	0.0	0.0036	0.0054	0.0	0.9994	1.0	inf	1.0	1.0	0.0002	0.9875	1.0	0.9994	inf	0.0009	0.9948	1.0	0.0	0.9925	-0.0001	3.0	0.0000	inf	inf	1.0	inf	1.0

Framework versions

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