Layoutlm onnx support (Issue huggingface#13300) (huggingface#13562)

* Add support for exporting PyTorch LayoutLM to ONNX

* Added tests for converting LayoutLM to ONNX

* Add support for exporting PyTorch LayoutLM to ONNX

* Added tests for converting LayoutLM to ONNX

* cleanup

* Removed regression/ folder

* Add support for exporting PyTorch LayoutLM to ONNX

* Added tests for converting LayoutLM to ONNX

* cleanup

* Fixed import error

* Remove unnecessary import statements

* Changed max_2d_positions from class variable to instance variable of the config class

* Add support for exporting PyTorch LayoutLM to ONNX

* Added tests for converting LayoutLM to ONNX

* cleanup

* Add support for exporting PyTorch LayoutLM to ONNX

* cleanup

* Fixed import error

* Changed max_2d_positions from class variable to instance variable of the config class

* Use super class generate_dummy_inputs method

Co-authored-by: Michael Benayoun <mickbenayoun@gmail.com>

* Add support for Masked LM, sequence classification and token classification

Co-authored-by: Michael Benayoun <mickbenayoun@gmail.com>

* Removed uncessary import and method

* Fixed code styling

* Raise error if PyTorch is not installed

* Remove unnecessary import statement

Co-authored-by: Michael Benayoun <mickbenayoun@gmail.com>

src/transformers/models/layoutlm/configuration_layoutlm.py

@@ -71,6 +71,8 @@ class LayoutLMConfig(BertConfig):
             The standard deviation of the truncated_normal_initializer for initializing all weight matrices.
         layer_norm_eps (:obj:`float`, `optional`, defaults to 1e-12):
             The epsilon used by the layer normalization layers.
+        gradient_checkpointing (:obj:`bool`, `optional`, defaults to :obj:`False`):
+            If True, use gradient checkpointing to save memory at the expense of slower backward pass.
         max_2d_position_embeddings (:obj:`int`, `optional`, defaults to 1024):
             The maximum value that the 2D position embedding might ever used. Typically set this to something large
             just in case (e.g., 1024).
@@ -106,6 +108,7 @@ def __init__(
         initializer_range=0.02,
         layer_norm_eps=1e-12,
         pad_token_id=0,
+        gradient_checkpointing=False,
         max_2d_position_embeddings=1024,
         **kwargs
     ):
@@ -123,6 +126,7 @@ def __init__(
             initializer_range=initializer_range,
             layer_norm_eps=layer_norm_eps,
             pad_token_id=pad_token_id,
+            gradient_checkpointing=gradient_checkpointing,
             **kwargs,
         )
         self.max_2d_position_embeddings = max_2d_position_embeddings
@@ -183,6 +187,11 @@ def generate_dummy_inputs(
             raise ValueError("Cannot generate dummy inputs without PyTorch installed.")
         import torch
 
-        batch_size, seq_length = input_dict["input_ids"].shape
-        input_dict["bbox"] = torch.tensor([*[box] * seq_length]).tile(batch_size, 1, 1)
+        input_dict["bbox"] = torch.tensor(
+            [
+                [0] * 4,
+                *[box] * seq_length,
+                [self.max_2d_positions] * 4,
+            ]
+        ).tile(batch_size, 1, 1)
         return input_dict