[Model] Add XLMRoBERTaModel in paddlenlp (#9720)

* add xlm_roberta in paddlenlp

Author: jie-z-0607

paddlenlp/transformers/__init__.py

@@ -209,6 +209,9 @@
 from .xlm.modeling import *
 from .xlm.tokenizer import *
 from .xlm.configuration import *
+from .xlm_roberta.modeling import *
+from .xlm_roberta.tokenizer import *
+from .xlm_roberta.configuration import *
 from .gau_alpha.modeling import *
 from .gau_alpha.tokenizer import *
 from .gau_alpha.configuration import *

paddlenlp/transformers/auto/configuration.py

@@ -113,6 +113,7 @@
         ("unimo", "UNIMOConfig"),
         ("visualglm", "VisualGLMConfig"),
         ("xlm", "XLMConfig"),
+        ("xlm-roberta", "XLMRobertaConfig"),
         ("xlnet", "XLNetConfig"),
         ("yuan", "YuanConfig"),
     ]
@@ -202,6 +203,7 @@
         ("unimo", "UNIMO"),
         ("visualglm", "VisualGLM"),
         ("xlm", "XLM"),
+        ("xlm-roberta", "XLMRoberta"),
         ("xlnet", "XLNet"),
         ("yuan", "Yuan"),
     ]

paddlenlp/transformers/auto/modeling.py

@@ -94,6 +94,7 @@
         ("UNIMO", "unimo"),
         ("XLNet", "xlnet"),
         ("XLM", "xlm"),
+        ("XLMRoberta", "xlm_roberta"),
         ("GPT", "gpt"),
         ("GLM", "glm"),
         ("MT5", "mt5"),

paddlenlp/transformers/auto/tokenizer.py

@@ -99,6 +99,7 @@
             ("squeezebert", "SqueezeBertTokenizer"),
             ("t5", "T5Tokenizer"),
             ("xlm", "XLMTokenizer"),
+            ("xlm_roberta", "XLMRobertaTokenizer"),
             ("xlnet", "XLNetTokenizer"),
             ("bert_japanese", "BertJapaneseTokenizer"),
             ("bigbird", "BigBirdTokenizer"),

paddlenlp/transformers/xlm_roberta/__init__.py

@@ -0,0 +1,17 @@
+# Copyright (c) 2025 PaddlePaddle Authors. 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 .configuration import *
+from .modeling import *
+from .tokenizer import *

paddlenlp/transformers/xlm_roberta/configuration.py

@@ -0,0 +1,160 @@
+# coding=utf-8
+# Copyright (c) 2025 PaddlePaddle Authors. All Rights Reserved.
+# Copyright 2018 The Google AI Language Team Authors and The HuggingFace Inc. team.
+# Copyright (c) 2018, NVIDIA CORPORATION.  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.
+""" XLM-RoBERTa configuration"""
+
+from ..model_utils import PretrainedConfig
+
+__all__ = ["PRETRAINED_INIT_CONFIGURATION", "XLMRobertaConfig"]
+
+PRETRAINED_INIT_CONFIGURATION = {
+    "hf-internal-testing/tiny-random-onnx-xlm-roberta": {
+        "attention_probs_dropout_prob": 0.1,
+        "bos_token_id": 0,
+        "classifier_dropout": None,
+        "eos_token_id": 2,
+        "hidden_act": "gelu",
+        "hidden_dropout_prob": 0.1,
+        "hidden_size": 4,
+        "initializer_range": 0.02,
+        "intermediate_size": 37,
+        "layer_norm_eps": 1e-05,
+        "max_position_embeddings": 514,
+        "model_type": "xlm-roberta",
+        "num_attention_heads": 4,
+        "num_hidden_layers": 5,
+        "output_past": True,
+        "pad_token_id": 1,
+        "position_embedding_type": "absolute",
+        "dtype": "float32",
+        "type_vocab_size": 1,
+        "use_cache": True,
+        "vocab_size": 250002,
+    },
+}
+
+
+class XLMRobertaConfig(PretrainedConfig):
+    r"""
+    This is the configuration class to store the configuration of a [`XLMRobertaModel`] or a [`TFXLMRobertaModel`]. It
+    is used to instantiate a XLM-RoBERTa model according to the specified arguments, defining the model architecture.
+    Instantiating a configuration with the defaults will yield a similar configuration to that of the XLMRoBERTa
+    [xlm-roberta-base](https://huggingface.co/xlm-roberta-base) architecture.
+
+    Configuration objects inherit from [`PretrainedConfig`] and can be used to control the model outputs. Read the
+    documentation from [`PretrainedConfig`] for more information.
+
+
+    Args:
+        vocab_size (`int`, *optional*, defaults to 30522):
+            Vocabulary size of the XLM-RoBERTa model. Defines the number of different tokens that can be represented by
+            the `inputs_ids` passed when calling [`XLMRobertaModel`] or [`TFXLMRobertaModel`].
+        hidden_size (`int`, *optional*, defaults to 768):
+            Dimensionality of the encoder layers and the pooler layer.
+        num_hidden_layers (`int`, *optional*, defaults to 12):
+            Number of hidden layers in the Transformer encoder.
+        num_attention_heads (`int`, *optional*, defaults to 12):
+            Number of attention heads for each attention layer in the Transformer encoder.
+        intermediate_size (`int`, *optional*, defaults to 3072):
+            Dimensionality of the "intermediate" (often named feed-forward) layer in the Transformer encoder.
+        hidden_act (`str` or `Callable`, *optional*, defaults to `"gelu"`):
+            The non-linear activation function (function or string) in the encoder and pooler. If string, `"gelu"`,
+            `"relu"`, `"silu"` and `"gelu_new"` are supported.
+        hidden_dropout_prob (`float`, *optional*, defaults to 0.1):
+            The dropout probability for all fully connected layers in the embeddings, encoder, and pooler.
+        attention_probs_dropout_prob (`float`, *optional*, defaults to 0.1):
+            The dropout ratio for the attention probabilities.
+        max_position_embeddings (`int`, *optional*, defaults to 512):
+            The maximum sequence length that this model might ever be used with. Typically set this to something large
+            just in case (e.g., 512 or 1024 or 2048).
+        type_vocab_size (`int`, *optional*, defaults to 2):
+            The vocabulary size of the `token_type_ids` passed when calling [`XLMRobertaModel`] or
+            [`TFXLMRobertaModel`].
+        initializer_range (`float`, *optional*, defaults to 0.02):
+            The standard deviation of the truncated_normal_initializer for initializing all weight matrices.
+        layer_norm_eps (`float`, *optional*, defaults to 1e-12):
+            The epsilon used by the layer normalization layers.
+        position_embedding_type (`str`, *optional*, defaults to `"absolute"`):
+            Type of position embedding. Choose one of `"absolute"`, `"relative_key"`, `"relative_key_query"`. For
+            positional embeddings use `"absolute"`. For more information on `"relative_key"`, please refer to
+            [Self-Attention with Relative Position Representations (Shaw et al.)](https://arxiv.org/abs/1803.02155).
+            For more information on `"relative_key_query"`, please refer to *Method 4* in [Improve Transformer Models
+            with Better Relative Position Embeddings (Huang et al.)](https://arxiv.org/abs/2009.13658).
+        is_decoder (`bool`, *optional*, defaults to `False`):
+            Whether the model is used as a decoder or not. If `False`, the model is used as an encoder.
+        use_cache (`bool`, *optional*, defaults to `True`):
+            Whether or not the model should return the last key/values attentions (not used by all models). Only
+            relevant if `config.is_decoder=True`.
+        classifier_dropout (`float`, *optional*):
+            The dropout ratio for the classification head.
+
+    Examples:
+
+    ```python
+    >>> from paddlenlp.transformers import XLMRobertaConfig, XLMRobertaModel
+
+    >>> # Initializing a XLM-RoBERTa xlm-roberta-base style configuration
+    >>> configuration = XLMRobertaConfig()
+
+    >>> # Initializing a model (with random weights) from the xlm-roberta-base style configuration
+    >>> model = XLMRobertaModel(configuration)
+
+    >>> # Accessing the model configuration
+    >>> configuration = model.config
+    ```"""
+
+    model_type = "xlm-roberta"
+
+    def __init__(
+        self,
+        vocab_size=30522,
+        hidden_size=768,
+        num_hidden_layers=12,
+        num_attention_heads=12,
+        intermediate_size=3072,
+        hidden_act="gelu",
+        hidden_dropout_prob=0.1,
+        attention_probs_dropout_prob=0.1,
+        max_position_embeddings=512,
+        type_vocab_size=2,
+        initializer_range=0.02,
+        layer_norm_eps=1e-12,
+        pad_token_id=1,
+        bos_token_id=0,
+        eos_token_id=2,
+        position_embedding_type="absolute",
+        use_cache=True,
+        classifier_dropout=None,
+        **kwargs,
+    ):
+        kwargs["return_dict"] = kwargs.pop("return_dict", False)
+        super().__init__(pad_token_id=pad_token_id, bos_token_id=bos_token_id, eos_token_id=eos_token_id, **kwargs)
+
+        self.vocab_size = vocab_size
+        self.hidden_size = hidden_size
+        self.num_hidden_layers = num_hidden_layers
+        self.num_attention_heads = num_attention_heads
+        self.hidden_act = hidden_act
+        self.intermediate_size = intermediate_size
+        self.hidden_dropout_prob = hidden_dropout_prob
+        self.attention_probs_dropout_prob = attention_probs_dropout_prob
+        self.max_position_embeddings = max_position_embeddings
+        self.type_vocab_size = type_vocab_size
+        self.initializer_range = initializer_range
+        self.layer_norm_eps = layer_norm_eps
+        self.position_embedding_type = position_embedding_type
+        self.use_cache = use_cache
+        self.classifier_dropout = classifier_dropout