fix modular

huggingface · Dec 29, 2024 · 350cafe · 350cafe
1 parent 6135be4
commit 350cafe
Show file tree

Hide file tree

Showing 11 changed files with 2,737 additions and 105 deletions.
diff --git a/docs/source/en/index.md b/docs/source/en/index.md
@@ -298,6 +298,7 @@ Flax), PyTorch, and/or TensorFlow.
 |                      [RoFormer](model_doc/roformer)                      |       ✅        |         ✅         |      ✅      |
 |                       [RT-DETR](model_doc/rt_detr)                       |       ✅        |         ❌         |      ❌      |
 |                [RT-DETR-ResNet](model_doc/rt_detr_resnet)                |       ✅        |         ❌         |      ❌      |
+|                     [RtDetrV2](model_doc/rt_detr_v2)                     |       ✅        |         ❌         |      ❌      |
 |                          [RWKV](model_doc/rwkv)                          |       ✅        |         ❌         |      ❌      |
 |                           [SAM](model_doc/sam)                           |       ✅        |         ✅         |      ❌      |
 |                  [SeamlessM4T](model_doc/seamless_m4t)                   |       ✅        |         ❌         |      ❌      |

diff --git a/src/transformers/__init__.py b/src/transformers/__init__.py
@@ -730,7 +730,7 @@
         "RoFormerTokenizer",
     ],
     "models.rt_detr": ["RTDetrConfig", "RTDetrResNetConfig"],
-    "models.rt_detr_v2": ["RtDetrV2Config"],
+    "models.rt_detr_v2": ["RtDetrV2Config", "RtDetrV2ResNetConfig"],
     "models.rwkv": ["RwkvConfig"],
     "models.sam": [
         "SamConfig",
@@ -3365,11 +3365,7 @@
         ]
     )
     _import_structure["models.rt_detr_v2"].extend(
-        [
-            "RtDetrV2ForObjectDetection",
-            "RtDetrV2Model",
-            "RtDetrV2PreTrainedModel"
-        ]
+        ["RtDetrV2ForObjectDetection", "RtDetrV2Model", "RtDetrV2PreTrainedModel"]
     )
     _import_structure["models.rwkv"].extend(
         [
@@ -5743,7 +5739,7 @@
         RTDetrConfig,
         RTDetrResNetConfig,
     )
-    from .models.rt_detr_v2 import RtDetrV2Config
+    from .models.rt_detr_v2 import RtDetrV2Config, RtDetrV2ResNetConfig
     from .models.rwkv import RwkvConfig
     from .models.sam import (
         SamConfig,

diff --git a/src/transformers/models/auto/configuration_auto.py b/src/transformers/models/auto/configuration_auto.py
@@ -252,6 +252,7 @@
         ("rt_detr", "RTDetrConfig"),
         ("rt_detr_resnet", "RTDetrResNetConfig"),
         ("rt_detr_v2", "RtDetrV2Config"),
+        ("rt_detr_v2_resnet", "RtDetrV2ResNetConfig"),
         ("rwkv", "RwkvConfig"),
         ("sam", "SamConfig"),
         ("seamless_m4t", "SeamlessM4TConfig"),
@@ -580,7 +581,8 @@
         ("roformer", "RoFormer"),
         ("rt_detr", "RT-DETR"),
         ("rt_detr_resnet", "RT-DETR-ResNet"),
-        ("rt_detr_v2", "RtDetrV2"),
+        ("rt_detr_v2", "RT-DETR-V2-ResNet"),
+        ("rt_detr_v2_resnet", "RtDetrV2ResNetConfig"),
         ("rwkv", "RWKV"),
         ("sam", "SAM"),
         ("seamless_m4t", "SeamlessM4T"),
@@ -713,6 +715,7 @@
         ("siglip_vision_model", "siglip"),
         ("chinese_clip_vision_model", "chinese_clip"),
         ("rt_detr_resnet", "rt_detr"),
+        ("rt_detr_v2_resnet", "rt_detr_v2"),
     ]
 )
 

diff --git a/src/transformers/models/rt_detr_v2/__init__.py b/src/transformers/models/rt_detr_v2/__init__.py
@@ -0,0 +1,21 @@
+# Copyright 2024 The HuggingFace Team. All rights reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+
+from typing import TYPE_CHECKING
+
+from ...utils import _LazyModule
+from ...utils.import_utils import define_import_structure
+from .configuration_rt_detr_v2 import *
+from .modeling_rt_detr_v2 import *
diff --git a/src/transformers/models/rt_detr_v2/configuration_rt_detr_v2.py b/src/transformers/models/rt_detr_v2/configuration_rt_detr_v2.py
diff --git a/src/transformers/models/rt_detr_v2/convert_rt_detr_v2_weights_to_hf.py b/src/transformers/models/rt_detr_v2/convert_rt_detr_v2_weights_to_hf.py
@@ -24,8 +24,7 @@
 from PIL import Image
 from torchvision import transforms
 
-from transformers import RTDetrImageProcessor
-from transformers.models.rt_detr_v2.modular_rt_detr_v2 import RtDetrV2Config, RtDetrV2ForObjectDetection
+from transformers import RTDetrImageProcessor, RtDetrV2Config, RtDetrV2ForObjectDetection
 from transformers.utils import logging
 
 
@@ -558,7 +557,6 @@ def convert_rt_detr_checkpoint(model_name, pytorch_dump_folder_path, push_to_hub
         "ema"
     ]["module"]
 
-
     # rename keys
     for src, dest in create_rename_keys(config):
         rename_key(state_dict, src, dest)
@@ -581,7 +579,6 @@ def convert_rt_detr_checkpoint(model_name, pytorch_dump_folder_path, push_to_hub
     model.load_state_dict(state_dict)
     model.eval()
 
-
     # load image processor
     image_processor = RTDetrImageProcessor()