ff

model_training_nnn_tpu/dataset_tf.py

@@ -852,20 +852,22 @@ def analyze_dataset_shapes(dataset_tf: BrainToTextDatasetTF, sample_size: int =
 def create_input_fn(dataset_tf: BrainToTextDatasetTF,
                    transform_args: Dict[str, Any],
                    training: bool = True,
-                   cache_path: Optional[str] = None,
-                   auto_analyze_shapes: bool = True) -> tf.data.Dataset:
+                   cache_path: Optional[str] = None) -> tf.data.Dataset:
     """
-    Create input function for TPU training with fixed-shape batching and data augmentation
+    Create input function for TPU training with DYNAMIC padding and data augmentation
+
+    This function uses dynamic shapes to avoid the "pad to a smaller size" error.
+    All variable-length dimensions use tf.TensorShape([None, ...]) to allow
+    TensorFlow to automatically determine the appropriate padding size for each batch.
 
     Args:
         dataset_tf: BrainToTextDatasetTF instance
         transform_args: Data transformation configuration
         training: Whether this is for training (applies augmentations)
         cache_path: Optional path for disk caching to improve I/O performance
-        auto_analyze_shapes: Whether to automatically analyze dataset for optimal shapes
 
     Returns:
-        tf.data.Dataset ready for TPU training with fixed shapes
+        tf.data.Dataset ready for TPU training with dynamic shapes
     """
 
     # Create individual example dataset with file-grouping I/O optimization
@@ -914,58 +916,27 @@ def create_input_fn(dataset_tf: BrainToTextDatasetTF,
         num_parallel_calls=tf.data.AUTOTUNE
     )
 
-    # Determine shapes for TPU compatibility
-    if auto_analyze_shapes:
-        # Dynamically analyze dataset to determine optimal shapes
-        shape_info = analyze_dataset_shapes(dataset_tf, sample_size=100)
-        max_time_steps = shape_info['max_time_steps']
-        max_phone_seq_len = shape_info['max_phone_seq_len']
-        max_transcription_len = shape_info['max_transcription_len']
-        n_features = shape_info['n_features']
-        print(f"🔧 Using auto-analyzed shapes: time_steps={max_time_steps}, phone_seq={max_phone_seq_len}, transcription={max_transcription_len}")
-    else:
-        # Use dynamic shapes for maximum compatibility - let TensorFlow handle padding automatically
-        # This avoids the "pad to a smaller size" error by allowing dynamic sizing
-        print(f"🔧 Using dynamic shapes for maximum compatibility")
-        # Calculate number of features based on subset
-        n_features = len(dataset_tf.feature_subset) if dataset_tf.feature_subset else 512
+    # ========================= DYNAMIC SHAPES SOLUTION =========================
+    # 使用动态形状避免 "pad to a smaller size" 错误
+    # 这是最简单、最健壮的解决方案
+    print("🔧 Using DYNAMIC shapes for maximum compatibility and robustness.")
 
-        padded_shapes = {
-            'input_features': tf.TensorShape([None, n_features]),
-            'seq_class_ids': tf.TensorShape([None]),
-            'n_time_steps': tf.TensorShape([]),  # Scalar
-            'phone_seq_lens': tf.TensorShape([]),  # Scalar
-            'day_indices': tf.TensorShape([]),  # Scalar
-            'transcriptions': tf.TensorShape([None]),
-            'block_nums': tf.TensorShape([]),  # Scalar
-            'trial_nums': tf.TensorShape([])  # Scalar
-        }
+    # Calculate number of features based on subset
+    n_features = len(dataset_tf.feature_subset) if dataset_tf.feature_subset else 512
 
-        # Create fixed-shape batches with dynamic padding
-        dataset = dataset.padded_batch(
-            batch_size=dataset_tf.batch_size,
-            padded_shapes=padded_shapes,
-            padding_values=padding_values,
-            drop_remainder=True  # Critical for TPU: ensures all batches have same size
-        )
-
-        # Prefetch for optimal performance
-        dataset = dataset.prefetch(tf.data.AUTOTUNE)
-
-        return dataset
-
-    # If using auto-analyzed shapes, create fixed-size padded shapes
+    # Define dynamic padded shapes - all variable dimensions use None
     padded_shapes = {
-        'input_features': [max_time_steps, n_features],
-        'seq_class_ids': [max_phone_seq_len],
-        'n_time_steps': [],  # Scalar
-        'phone_seq_lens': [],  # Scalar
-        'day_indices': [],  # Scalar
-        'transcriptions': [max_transcription_len],
-        'block_nums': [],  # Scalar
-        'trial_nums': []  # Scalar
+        'input_features': tf.TensorShape([None, n_features]),  # 时间维度动态
+        'seq_class_ids': tf.TensorShape([None]),              # 序列长度动态
+        'n_time_steps': tf.TensorShape([]),                   # 标量
+        'phone_seq_lens': tf.TensorShape([]),                 # 标量
+        'day_indices': tf.TensorShape([]),                    # 标量
+        'transcriptions': tf.TensorShape([None]),             # 转录长度动态
+        'block_nums': tf.TensorShape([]),                     # 标量
+        'trial_nums': tf.TensorShape([])                      # 标量
     }
 
+    # Define padding values for each field
     padding_values = {
         'input_features': 0.0,
         'seq_class_ids': 0,
@@ -977,7 +948,8 @@ def create_input_fn(dataset_tf: BrainToTextDatasetTF,
         'trial_nums': 0
     }
 
-    # Create fixed-shape batches with padding
+    # Create batches with dynamic padding - TensorFlow will automatically
+    # determine the appropriate padding size for each batch
     dataset = dataset.padded_batch(
         batch_size=dataset_tf.batch_size,
         padded_shapes=padded_shapes,