添加：EDG++选择蒸馏

.gitignore

@@ -5,6 +5,16 @@ __pycache__/
 *$py.class
 *.pth
 
+# Data files
+*.pt
+*.npz
+*.csv
+*.sdf
+*.gz
+*.swp
+downstream/datasets/*/processed/
+downstream/datasets/*/raw/
+
 # Project file
 logs/
 application/

EDG/data_utils.py

@@ -1,50 +1,50 @@
-import collections.abc as container_abcs
-
-import numpy as np
-import torch
-from torch.utils.data.dataloader import default_collate
-from torch_geometric.data import Data, Batch
-
-string_classes, int_classes = str, int
-
-
-class DualCollater(object):
-    def __init__(self, follow_batch, multigpu=False):
-        self.follow_batch = follow_batch
-        self.multigpu = multigpu
-
-    def collate(self, batch):
-
-        elem = batch[0]
-        if isinstance(elem, Data):
-            if self.multigpu:
-                return batch
-            else:
-                batch = Batch.from_data_list(batch, self.follow_batch)
-                if "img_feat" in batch:
-                    # assert len(batch.img_feat.shape) == 1  # img_feat 被拉平了
-                    # n = batch.num_graphs
-                    batch.img_feat = torch.from_numpy(np.stack(batch.img_feat))
-                    # batch.img_feat = batch.img_feat.resize(n, 512)
-                return batch
-        elif isinstance(elem, torch.Tensor):
-            return default_collate(batch)
-        elif isinstance(elem, np.ndarray):
-            return default_collate(batch)
-        elif isinstance(elem, float):
-            return torch.tensor(batch, dtype=torch.float)
-        elif isinstance(elem, int_classes):
-            return torch.tensor(batch)
-        elif isinstance(elem, string_classes):
-            return batch
-        elif isinstance(elem, container_abcs.Mapping):
-            return {key: self.collate([d[key] for d in batch]) for key in elem}
-        elif isinstance(elem, tuple) and hasattr(elem, '_fields'):
-            return type(elem)(*(self.collate(s) for s in zip(*batch)))
-        elif isinstance(elem, container_abcs.Sequence):
-            return [self.collate(s) for s in zip(*batch)]
-
-        raise TypeError('DataLoader found invalid type: {}'.format(type(elem)))
-
-    def __call__(self, batch):
-        return self.collate(batch)
+import collections.abc as container_abcs
+
+import numpy as np
+import torch
+from torch.utils.data.dataloader import default_collate
+from torch_geometric.data import Data, Batch
+
+string_classes, int_classes = str, int
+
+
+class DualCollater(object):
+    def __init__(self, follow_batch, multigpu=False):
+        self.follow_batch = follow_batch
+        self.multigpu = multigpu
+
+    def collate(self, batch):
+
+        elem = batch[0]
+        if isinstance(elem, Data):
+            if self.multigpu:
+                return batch
+            else:
+                batch = Batch.from_data_list(batch, self.follow_batch)
+                if "img_feat" in batch:
+                    # assert len(batch.img_feat.shape) == 1  # img_feat 被拉平了
+                    # n = batch.num_graphs
+                    batch.img_feat = torch.from_numpy(np.stack(batch.img_feat))
+                    # batch.img_feat = batch.img_feat.resize(n, 512)
+                return batch
+        elif isinstance(elem, torch.Tensor):
+            return default_collate(batch)
+        elif isinstance(elem, np.ndarray):
+            return default_collate(batch)
+        elif isinstance(elem, float):
+            return torch.tensor(batch, dtype=torch.float)
+        elif isinstance(elem, int_classes):
+            return torch.tensor(batch)
+        elif isinstance(elem, string_classes):
+            return batch
+        elif isinstance(elem, container_abcs.Mapping):
+            return {key: self.collate([d[key] for d in batch]) for key in elem}
+        elif isinstance(elem, tuple) and hasattr(elem, '_fields'):
+            return type(elem)(*(self.collate(s) for s in zip(*batch)))
+        elif isinstance(elem, container_abcs.Sequence):
+            return [self.collate(s) for s in zip(*batch)]
+
+        raise TypeError('DataLoader found invalid type: {}'.format(type(elem)))
+
+    def __call__(self, batch):
+        return self.collate(batch)

EDG/dataset/QM9/raw/QM9_README

@@ -0,0 +1,30 @@
+QM9 is a comprehensive dataset that provides geometric, energetic, electronic and thermodynamic properties for a subset of GDB-17 database, comprising 134 thousand stable organic molecules with up to 9 heavy atoms. All moleucles are modeled using density functional theory (B3LYP/6-31G(2df,p) based DFT).
+
+The data files:
+qm9.sdf: molecular structures
+qm9.sdf.csv: tables for molecular properties
+	"mol_id" - Molecule ID (gdb9 index) mapping to the .sdf file
+	"A" - Rotational constant (unit: GHz)
+	"B" - Rotational constant (unit: GHz)
+	"C" - Rotational constant (unit: GHz)
+	"mu" - Dipole moment (unit: D)
+	"alpha" - Isotropic polarizability (unit: Bohr^3)
+	"homo" - Highest occupied molecular orbital energy (unit: Hartree)
+	"lumo" - Lowest unoccupied molecular orbital energy (unit: Hartree)
+	"gap" - Gap between HOMO and LUMO (unit: Hartree)
+	"r2" - Electronic spatial extent (unit: Bohr^2)
+	"zpve" - Zero point vibrational energy (unit: Hartree)
+	"u0" - Internal energy at 0K (unit: Hartree)
+	"u298" - Internal energy at 298.15K (unit: Hartree)
+	"h298" - Enthalpy at 298.15K (unit: Hartree)
+	"g298" - Free energy at 298.15K (unit: Hartree)
+	"cv" - Heat capavity at 298.15K (unit: cal/(mol*K))
+	"u0_atom" - Atomization energy at 0K (unit: kcal/mol)
+	"u298_atom" - Atomization energy at 298.15K (unit: kcal/mol)
+	"h298_atom" - Atomization enthalpy at 298.15K (unit: kcal/mol)
+	"g298_atom" - Atomization free energy at 298.15K (unit: kcal/mol)
+*"u0_atom" ~ "g298_atom" are calculated from the differences between "u0" ~ "g298" and sum of reference energies of all atoms in the molecules, as given in https://figshare.com/articles/Atomref%3A_Reference_thermochemical_energies_of_H%2C_C%2C_N%2C_O%2C_F_atoms./1057643
+
+Reference:
+Blum, Lorenz C., and Jean-Louis Reymond. "970 million druglike small molecules for virtual screening in the chemical universe database GDB-13." Journal of the American Chemical Society 131.25 (2009): 8732-8733.
+Ramakrishnan, Raghunathan, et al. "Quantum chemistry structures and properties of 134 kilo molecules." Scientific data 1 (2014): 140022.

EDG/dataset/QM9/raw/atomref.txt

@@ -0,0 +1,11 @@
+
+=========================================================================================================
+  Ele-    ZPVE         U (0 K)      U (298.15 K)    H (298.15 K)    G (298.15 K)     CV
+  ment   Hartree       Hartree        Hartree         Hartree         Hartree        Cal/(Mol Kelvin)
+=========================================================================================================
+   H     0.000000     -0.500273      -0.498857       -0.497912       -0.510927       2.981
+   C     0.000000    -37.846772     -37.845355      -37.844411      -37.861317       2.981
+   N     0.000000    -54.583861     -54.582445      -54.581501      -54.598897       2.981
+   O     0.000000    -75.064579     -75.063163      -75.062219      -75.079532       2.981
+   F     0.000000    -99.718730     -99.717314      -99.716370      -99.733544       2.981
+=========================================================================================================