refactor: refactor codes for better GPU usage

Novartis · May 28, 2024 · 535326f · 535326f
1 parent 7bb5d9a
commit 535326f
Showing 1 changed file with 14 additions and 5 deletions.
diff --git a/scar/main/_scar.py b/scar/main/_scar.py
@@ -494,6 +494,11 @@ def train(
 
                 vae_nets.train()
                 for x_batch, ambient_freq, batch_id_onehot in training_generator:
+                    # Move data to device
+                    x_batch = x_batch.to(self.device)
+                    ambient_freq = ambient_freq.to(self.device)
+                    batch_id_onehot = batch_id_onehot.to(self.device)
+
                     optim.zero_grad()
                     dec_nr, dec_prob, means, var, dec_dp = vae_nets(x_batch, batch_id_onehot)
                     recon_loss_minibatch, kld_loss_minibatch, loss_minibatch = loss_fn(
@@ -603,6 +608,11 @@ def inference(
         )
 
         for x_batch_tot, ambient_freq_tot, x_batch_id_onehot_tot in generator_full_data:
+            # Move data to device
+            x_batch_tot = x_batch_tot.to(self.device)
+            x_batch_id_onehot_tot = x_batch_id_onehot_tot.to(self.device)
+            ambient_freq_tot = ambient_freq_tot.to(self.device)
+
             minibatch_size = x_batch_tot.shape[
                 0
             ]  # if not the last batch, equals to batch size
@@ -706,13 +716,12 @@ def assignment(self, cutoff=3, moi=None):
 class UMIDataset(torch.utils.data.Dataset):
     """Characterizes dataset for PyTorch"""
 
-    def __init__(self, raw_count, ambient_profile, batch_id, device, list_ids=None):
+    def __init__(self, raw_count, ambient_profile, batch_id, list_ids=None):
         """Initialization"""
-        self.device = device
-        self.raw_count = torch.from_numpy(raw_count).int().to(self.device)
-        self.ambient_profile = torch.from_numpy(ambient_profile).float().to(device)
+        self.raw_count = torch.from_numpy(raw_count).int()
+        self.ambient_profile = torch.from_numpy(ambient_profile).float()
         self.batch_id = torch.from_numpy(batch_id).to(torch.int64)
-        self.batch_onehot = torch.from_numpy(np.eye(len(np.unique(batch_id)))).to(torch.int64).to(self.device)
+        self.batch_onehot = torch.from_numpy(np.eye(len(np.unique(batch_id)))).to(torch.int64)
 
         if list_ids:
             self.list_ids = list_ids