improvements

metadata.json

@@ -1,9 +1,9 @@
 {
     "Identifier": "eos9taz",
-    "Slug": "moler-generation",
+    "Slug": "moler-enamine-fragments",
     "Status": "Ready",
-    "Title": "Extending molecular scaffolds with structural motifs",
-    "Description": "MoLeR is a graph-based generative model that combines fragment-based and atom-by-atom generation of new molecules with scaffold-constrained optimization. It does not depend on generation history and therefore MoLeR is able to complete arbitrary scaffolds. The model has been trained on the GuacaMol dataset.",
+    "Title": "Extending molecular scaffolds with fragments",
+    "Description": "MoLeR is a graph-based generative model that combines fragment-based and atom-by-atom generation of new molecules with scaffold-constrained optimization. It does not depend on generation history and therefore MoLeR is able to complete arbitrary scaffolds. The model has been trained on the GuacaMol dataset. Here we sample a fragment library from Enamine.",
     "Mode": "Pretrained",
     "Task": [
         "Generative"
@@ -19,7 +19,7 @@
         "String"
     ],
     "Output Shape": "List",
-    "Interpretation": "100 new molecules are sampled for each input molecule. ",
+    "Interpretation": "1000 new molecules are sampled for each input molecule, preserving its scaffold. ",
     "Tag": [
         "Chemical graph model",
         "Compound generation"

model/checkpoints/README.md

@@ -1 +1,7 @@
-# Model pretrained parameters
+The fragments library has been downloaded from Enamine on the 1st of November, 2023.
+The following libraries were downloaded:
+- Enamine 3D-shape diverse
+- Enamine essential fragment library
+- Enamine high fidelity
+- Enamine natural product-like
+- Enamine single pharmacophore