Renamed output_indices to intermediate_feature_indices

src/transformers/models/seggpt/configuration_seggpt.py

@@ -73,7 +73,7 @@ class SegGptConfig(PretrainedConfig):
             Whether to use relative position encoding in the Attention
         merge_index (`int`, *optional*, defaults to 2):
             The index of the encoder layer to merge the embeddings.
-        output_indicies (`List[int]`, *optional*, defaults to `[5, 11, 17, 23]`):
+        intermediate_feature_indices (`List[int]`, *optional*, defaults to `[5, 11, 17, 23]`):
             The indices of the encoder layers which we store as features for the decoder.
         beta (`float`, *optional*, defaults to 0.01):
             Regularization factor for SegGptLoss (smooth-l1 loss).
@@ -114,15 +114,15 @@ def __init__(
         decoder_hidden_size=64,
         use_relative_position_embeddings=True,
         merge_index=2,
-        output_indicies=[5, 11, 17, 23],
+        intermediate_feature_indices=[5, 11, 17, 23],
         beta=0.01,
         **kwargs,
     ):
         super().__init__(**kwargs)
 
-        if merge_index > min(output_indicies):
+        if merge_index > min(intermediate_feature_indices):
             raise ValueError(
-                f"Merge index must be less than the minimum encoder output index, but got {merge_index=} and {output_indicies=}"
+                f"Merge index must be less than the minimum encoder output index, but got {merge_index=} and {intermediate_feature_indices=}"
             )
         self.hidden_size = hidden_size
         self.num_hidden_layers = num_hidden_layers
@@ -140,6 +140,6 @@ def __init__(
         self.decoder_hidden_size = decoder_hidden_size
         self.use_relative_position_embeddings = use_relative_position_embeddings
         self.merge_index = merge_index
-        self.output_indicies = output_indicies
+        self.intermediate_feature_indices = intermediate_feature_indices
         self.beta = beta
         self.mlp_dim = int(hidden_size * 4) if mlp_dim is None else mlp_dim

src/transformers/models/seggpt/modeling_seggpt.py

@@ -92,9 +92,9 @@ class SegGptEncoderOutput(BaseModelOutput):
         attentions (`Tuple[torch.FloatTensor]`, `optional`, returned when `config.output_attentions=True`):
             Tuple of *torch.FloatTensor* (one for each layer) of shape
             `(batch_size, num_heads, seq_len, seq_len)`.
-        intermediate_features (`Tuple[torch.FloatTensor]`, `optional`, returned when `config.output_indicies` is set):
+        intermediate_features (`Tuple[torch.FloatTensor]`, `optional`, returned when `config.intermediate_feature_indices` is set):
             Tuple of `torch.FloatTensor` of shape `(batch_size, patch_height, patch_width, hidden_size)`.
-            Each element in the Tuple corresponds to the output of the layer specified in `config.output_indicies`.
+            Each element in the Tuple corresponds to the output of the layer specified in `config.intermediate_feature_indices`.
             Additionaly, each feature passes through a LayerNorm.
     """
 
@@ -538,7 +538,7 @@ def forward(
                     hidden_states[: hidden_states.shape[0] // 2] + hidden_states[hidden_states.shape[0] // 2 :]
                 ) * 0.5
 
-            if i in self.config.output_indicies:
+            if i in self.config.intermediate_feature_indices:
                 intermediate_features.append(self.layernorm(hidden_states))
 
             if output_attentions:
@@ -620,7 +620,7 @@ class SegGptDecoder(nn.Module):
     def __init__(self, config):
         super().__init__()
         self.decoder_embed = nn.Linear(
-            config.hidden_size * len(config.output_indicies),
+            config.hidden_size * len(config.intermediate_feature_indices),
             config.patch_size**2 * config.decoder_hidden_size,
             bias=True,
         )

tests/models/seggpt/test_modeling_seggpt.py

@@ -65,7 +65,7 @@ def __init__(
         initializer_range=0.02,
         mlp_ratio=2.0,
         merge_index=0,
-        output_indicies=[1],
+        intermediate_feature_indices=[1],
         pretrain_image_size=10,
         decoder_hidden_size=10,
     ):
@@ -85,7 +85,7 @@ def __init__(
         self.initializer_range = initializer_range
         self.mlp_ratio = mlp_ratio
         self.merge_index = merge_index
-        self.output_indicies = output_indicies
+        self.intermediate_feature_indices = intermediate_feature_indices
         self.pretrain_image_size = pretrain_image_size
         self.decoder_hidden_size = decoder_hidden_size
 
@@ -121,7 +121,7 @@ def get_config(self):
             initializer_range=self.initializer_range,
             mlp_ratio=self.mlp_ratio,
             merge_index=self.merge_index,
-            output_indicies=self.output_indicies,
+            intermediate_feature_indices=self.intermediate_feature_indices,
             pretrain_image_size=self.pretrain_image_size,
             decoder_hidden_size=self.decoder_hidden_size,
         )