Added a yaml construct method for Trainer (#144)

* added a yaml construct method for Trainer

* used safe_load

* reformatted the yaml structure to use the same _construct method

* added test for YAML config loader

* added pyyaml dep

azure/install_dependencies.sh

@@ -2,5 +2,5 @@
 
 python -m pip install --upgrade pip
 pip install torch_nightly -f https://download.pytorch.org/whl/nightly/cpu/torch_nightly.html
-pip install tqdm ase tensorboardX
+pip install tqdm ase tensorboardX pyyaml
 pip install pytorch-ignite --no-deps

azure/runnable_submodules.yml

@@ -18,5 +18,8 @@ steps:
 - script: 'python -m torchani.neurochem.trainer --tqdm tests/test_data/inputtrain.ipt  dataset/ani_gdb_s01.h5 dataset/ani_gdb_s01.h5'
   displayName: NeuroChem Trainer
 
+- script: 'python -m torchani.neurochem.trainer --tqdm tests/test_data/inputtrain.yaml  dataset/ani_gdb_s01.h5 dataset/ani_gdb_s01.h5'
+  displayName: NeuroChem Trainer YAML config
+
 - script: 'python -m torchani.data.cache_aev tmp dataset/ani_gdb_s01.h5 256'
   displayName: Cache AEV

tests/test_data/inputtrain.yaml

@@ -0,0 +1,106 @@
+# InputFile for Force Prediction Network
+sflparamsfile: 'rHCNO-5.2R_16-3.5A_a4-8.params'
+ntwkStoreDir: 'networks/'
+atomEnergyFile: 'sae_linfit.dat'
+nmax: 0 # Maximum number of iterations (0 = inf)
+tolr: 1 # Tolerance - early stopping
+emult: 0.1 # Multiplier by eta after tol switch
+eta: 0.001 # Eta -- Learning rate
+tcrit: 1.0E-5 # Eta termination criterion
+tmax: 0 # Maximum time (0 = inf)
+tbtchsz: 2560
+vbtchsz: 2560
+gpuid: 0
+ntwshr: 0
+nkde: 2
+energy: 1
+force: 0
+fmult: 0.0
+pbc: 0
+cmult: 0.001
+runtype: 'ANNP_CREATE_HDNN_AND_TRAIN' # Create and train a HDN network
+
+network_setup:
+  inputsize: 384
+  atom_net:
+    H:
+      - nodes: 160
+        activation: 9
+        type: 0
+        # l2norm: 1
+        # l2valu: 0.0001
+      - nodes: 128
+        activation: 9
+        type: 0
+        # l2norm: 1
+        # l2valu: 0.00001
+      - nodes: 96
+        activation: 9
+        type: 0
+        # l2norm: 1
+        # l2valu: 0.000001
+      - nodes: 1
+        activation: 6
+        type: 0
+    C:
+      - nodes: 144
+        activation: 9
+        type: 0
+        # l2norm: 1
+        # l2valu: 0.0001
+      - nodes: 112
+        activation: 9
+        type: 0
+        # l2norm: 1
+        # l2valu: 0.00001
+      - nodes: 96
+        activation: 9
+        type: 0
+        # l2norm: 1
+        # l2valu: 0.000001
+      - nodes: 1
+        activation: 6
+        type: 0
+    N:
+      - nodes: 128
+        activation: 9
+        type: 0
+        # l2norm: 1
+        # l2valu: 0.0001
+      - nodes: 112
+        activation: 9
+        type: 0
+        # l2norm: 1
+        # l2valu: 0.00001
+      - nodes: 96
+        activation: 9
+        type: 0
+        # l2norm: 1
+        # l2valu: 0.000001
+      - nodes: 1
+        activation: 6
+        type: 0
+    O:
+      - nodes: 128
+        activation: 9
+        type: 0
+        # l2norm: 1
+        # l2valu: 0.0001
+      - nodes: 112
+        activation: 9
+        type: 0
+        # l2norm: 1
+        # l2valu: 0.00001
+      - nodes: 96
+        activation: 9
+        type: 0
+        # l2norm: 1
+        # l2valu: 0.000001
+      - nodes: 1
+        activation: 6
+        type: 0
+
+adptlrn: 'OFF' # Adaptive learning (OFF,RMSPROP)
+decrate: 0.9 # Decay rate of RMSPROP
+moment: 'ADAM' #  Turn on momentum or nesterov momentum (OFF,CNSTTEMP,TMANNEAL,REGULAR,NESTEROV)
+mu: 0.99 # Mu factor for momentum

torchani/neurochem/__init__.py

@@ -337,8 +337,12 @@ def __init__(self, filename, device=torch.device('cuda'), tqdm=False,
             self.training_eval_every = 20
         else:
             self.tensorboard = None
+
         with open(filename, 'r') as f:
-            network_setup, params = self._parse(f.read())
+            if filename.endswith('.yaml') or filename.endswith('.yml'):
+                network_setup, params = self._parse_yaml(f)
+            else:
+                network_setup, params = self._parse(f.read())
             self._construct(network_setup, params)
 
     def _parse(self, txt):
@@ -438,6 +442,14 @@ def input_size(self, v):
 
         return TreeExec().transform(tree)
 
+    def _parse_yaml(self, f):
+        import yaml
+        params = yaml.safe_load(f)
+        network_setup = params['network_setup']
+        del params['network_setup']
+        network_setup = (network_setup['inputsize'], network_setup['atom_net'])
+        return network_setup, params
+
     def _construct(self, network_setup, params):
         dir_ = os.path.dirname(os.path.abspath(self.filename))