Merge pull request #4977 from WenmuZhou/fix_vqa

fix win train bug

ppstructure/vqa/README.md

@@ -165,6 +165,7 @@ python3.7 train_ser.py \
     --learning_rate 5e-5 \
     --warmup_steps 50 \
     --evaluate_during_training \
+    --num_workers 8 \
     --seed 2048 \
     --resume
 ```
@@ -177,6 +178,7 @@ python3 eval_ser.py \
     --eval_data_dir "XFUND/zh_val/image" \
     --eval_label_path "XFUND/zh_val/xfun_normalize_val.json" \
     --per_gpu_eval_batch_size 8 \
+    --num_workers 8 \
     --output_dir "output/ser/"  \
     --seed 2048
 ```
@@ -234,6 +236,7 @@ python3 train_re.py \
     --warmup_steps 50 \
     --per_gpu_train_batch_size 8 \
     --per_gpu_eval_batch_size 8 \
+    --num_workers 8 \
     --evaluate_during_training \
     --seed 2048
 
@@ -257,6 +260,7 @@ python3 train_re.py \
     --warmup_steps 50 \
     --per_gpu_train_batch_size 8 \
     --per_gpu_eval_batch_size 8 \
+    --num_workers 8 \
     --evaluate_during_training \
     --seed 2048 \
     --resume
@@ -276,6 +280,7 @@ python3 eval_re.py \
     --label_map_path 'labels/labels_ser.txt' \
     --output_dir "output/re_test/"  \
     --per_gpu_eval_batch_size 8 \
+    --num_workers 8 \
     --seed 2048
 ```
 最终会打印出`precision`, `recall`, `f1`等指标

ppstructure/vqa/eval_re.py

@@ -112,7 +112,7 @@ def eval(args):
     eval_dataloader = paddle.io.DataLoader(
         eval_dataset,
         batch_size=args.per_gpu_eval_batch_size,
-        num_workers=8,
+        num_workers=args.num_workers,
         shuffle=False,
         collate_fn=DataCollator())
 

ppstructure/vqa/eval_ser.py

@@ -61,7 +61,7 @@ def eval(args):
     eval_dataloader = paddle.io.DataLoader(
         eval_dataset,
         batch_size=args.per_gpu_eval_batch_size,
-        num_workers=0,
+        num_workers=args.num_workers,
         use_shared_memory=True,
         collate_fn=None, )
 

ppstructure/vqa/train_re.py

@@ -97,14 +97,14 @@ def train(args):
     train_dataloader = paddle.io.DataLoader(
         train_dataset,
         batch_sampler=train_sampler,
-        num_workers=8,
+        num_workers=args.num_workers,
         use_shared_memory=True,
         collate_fn=DataCollator())
 
     eval_dataloader = paddle.io.DataLoader(
         eval_dataset,
         batch_size=args.per_gpu_eval_batch_size,
-        num_workers=8,
+        num_workers=args.num_workers,
         shuffle=False,
         collate_fn=DataCollator())
 

ppstructure/vqa/train_ser.py

@@ -94,14 +94,14 @@ def train(args):
     train_dataloader = paddle.io.DataLoader(
         train_dataset,
         batch_sampler=train_sampler,
-        num_workers=0,
+        num_workers=args.num_workers,
         use_shared_memory=True,
         collate_fn=None, )
 
     eval_dataloader = paddle.io.DataLoader(
         eval_dataset,
         batch_size=args.per_gpu_eval_batch_size,
-        num_workers=0,
+        num_workers=args.num_workers,
         use_shared_memory=True,
         collate_fn=None, )
 

ppstructure/vqa/utils.py

@@ -363,6 +363,7 @@ def parse_args():
     parser.add_argument("--output_dir", default=None, type=str, required=True,)
     parser.add_argument("--max_seq_length", default=512, type=int,)
     parser.add_argument("--evaluate_during_training", action="store_true",)
+    parser.add_argument("--num_workers", default=8, type=int,)
     parser.add_argument("--per_gpu_train_batch_size", default=8,
                         type=int, help="Batch size per GPU/CPU for training.",)
     parser.add_argument("--per_gpu_eval_batch_size", default=8,

ppstructure/vqa/xfun.py

@@ -79,14 +79,36 @@ class XFUNDataset(Dataset):
 
         self.entities_labels = {'HEADER': 0, 'QUESTION': 1, 'ANSWER': 2}
         self.return_keys = {
-            'bbox': 'np',
-            'input_ids': 'np',
-            'labels': 'np',
-            'attention_mask': 'np',
-            'image': 'np',
-            'token_type_ids': 'np',
-            'entities': 'dict',
-            'relations': 'dict',
+            'bbox': {
+                'type': 'np',
+                'dtype': 'int64'
+            },
+            'input_ids': {
+                'type': 'np',
+                'dtype': 'int64'
+            },
+            'labels': {
+                'type': 'np',
+                'dtype': 'int64'
+            },
+            'attention_mask': {
+                'type': 'np',
+                'dtype': 'int64'
+            },
+            'image': {
+                'type': 'np',
+                'dtype': 'float32'
+            },
+            'token_type_ids': {
+                'type': 'np',
+                'dtype': 'int64'
+            },
+            'entities': {
+                'type': 'dict'
+            },
+            'relations': {
+                'type': 'dict'
+            }
         }
 
         if load_mode == "all":
@@ -103,7 +125,7 @@ class XFUNDataset(Dataset):
                       return_special_tokens_mask=False):
         # Padding
         needs_to_be_padded = pad_to_max_seq_len and \
-                             max_seq_len and len(encoded_inputs["input_ids"]) < max_seq_len
+            max_seq_len and len(encoded_inputs["input_ids"]) < max_seq_len
 
         if needs_to_be_padded:
             difference = max_seq_len - len(encoded_inputs["input_ids"])
@@ -412,8 +434,8 @@ class XFUNDataset(Dataset):
         return_data = {}
         for k, v in data.items():
             if k in self.return_keys:
-                if self.return_keys[k] == 'np':
-                    v = np.array(v)
+                if self.return_keys[k]['type'] == 'np':
+                    v = np.array(v, dtype=self.return_keys[k]['dtype'])
                 return_data[k] = v
         return return_data