Sync

_ShanghaiA-train.sh

@@ -6,6 +6,8 @@
 #SBATCH --mem=24000
 #SBATCH --time=3-00:00:00
 
+set -e
+
 export CUDA_HOME=/opt/cuda-9.0.176.1/
 export CUDNN_HOME=/opt/cuDNN-7.0/
 export STUDENT_ID=$(whoami)
@@ -21,8 +23,11 @@ export TMPDIR=/disk/scratch/${STUDENT_ID}/
 export TMP=/disk/scratch/${STUDENT_ID}/
 
 source /home/${STUDENT_ID}/miniconda3/bin/activate mlp-cuda
+
 python train.py \
-    --model='stn' \
+    --debug True \
+    --model 'stn' \
     --save_path ./save_file/ShanghaiA \
     --batch_size 4 \
-    --gpus 0,1,2,3,4,5
+    --gpus 0,1,2,3,4,5 \
+    --print_freq 100

model/stn.py

@@ -24,7 +24,6 @@ class STNet(nn.Module):
             _dummy_size_ = input_size
 
         # shape checking
-        print("STN: dummy_size {}".format(_dummy_size_))
         _dummy_x_ = torch.zeros(_dummy_size_)
 
         # (3.1) Spatial transformer localization-network

train.py

@@ -89,7 +89,7 @@ def build_test_loader(data_keys, args):
     test_loader = DataLoader(
         dataset=test_dataset,
         sampler=test_dist_sampler,
-        batch_size=1
+        batch_size=4
     )
     return test_loader
 
@@ -299,13 +299,17 @@ def valid_one_epoch(test_loader, model, device, args):
 
         with torch.no_grad():
             out, gt_count = model(img, kpoint)
+            if args.debug:
+                print("out: {} | gt_count: {}".format(
+                    out.shape, gt_count.shape
+                ))
             count = torch.sum(out).item()
-
             gt_count = torch.sum(gt_count).item()
+
         mae += abs(gt_count - count)
         mse += abs(gt_count - count) ** 2
 
-        if i % 15 == 0:
+        # if i % 15 == 0:
         print("[valid_one_epoch] {} Gt {:.2f} Pred {}".format(
             fname[0], gt_count, count
         ))