Add camembert model (#1034)

Author: liuchuting

mindone/transformers/__init__.py

@@ -120,6 +120,16 @@
     Blip2QFormerModel,
     Blip2VisionModel,
 )
+from .models.camembert import (
+    CamembertForCausalLM,
+    CamembertForMaskedLM,
+    CamembertForMultipleChoice,
+    CamembertForQuestionAnswering,
+    CamembertForSequenceClassification,
+    CamembertForTokenClassification,
+    CamembertModel,
+    CamembertPreTrainedModel,
+)
 from .models.chameleon import (
     ChameleonForConditionalGeneration,
     ChameleonModel,

mindone/transformers/models/__init__.py

@@ -28,6 +28,7 @@
     bit,
     blip,
     blip_2,
+    camembert,
     clap,
     clip,
     dpt,

mindone/transformers/models/auto/configuration_auto.py

@@ -60,6 +60,7 @@
         ("granitemoeshared", "GraniteMoeSharedConfig"),
         ("helium", "HeliumConfig"),
         ("hiera", "HieraConfig"),
+        ("camembert", "CamembertConfig"),
         ("idefics", "IdeficsConfig"),
         ("idefics2", "Idefics2Config"),
         ("idefics3", "Idefics3Config"),
@@ -111,6 +112,7 @@
         ("aria_text", "AriaText"),
         ("bert", "BERT"),
         ("bart", "BART"),
+        ("camembert", "CamemBERT"),
         ("bit", "BiT"),
         ("blip", "BLIP"),
         ("blip-2", "BLIP-2"),

mindone/transformers/models/auto/modeling_auto.py

@@ -37,6 +37,7 @@
         ("aria_text", "AriaTextModel"),
         ("bert", "BertModel"),
         ("bart", "BartModel"),
+        ("camembert", "CamembertModel"),
         ("bit", "BitModel"),
         ("blip", "BlipModel"),
         ("blip-2", "Blip2Model"),
@@ -106,6 +107,7 @@
         # Model for pre-training mapping
         ("albert", "AlbertForPreTraining"),
         ("bart", "BartForConditionalGeneration"),
+        ("camembert", "CamembertForMaskedLM"),
         ("bert", "BertForPreTraining"),
         ("gpt2", "GPT2LMHeadModel"),
         ("gemma3", "Gemma3ForConditionalGeneration"),
@@ -137,6 +139,7 @@
         ("deberta-v2", "DebertaV2ForMaskedLM"),
         ("gpt2", "GPT2LMHeadModel"),
         ("led", "LEDForConditionalGeneration"),
+        ("camembert", "CamembertForMaskedLM"),
         ("roberta", "RobertaForMaskedLM"),
         ("megatron-bert", "MegatronBertForCausalLM"),
         ("mobilebert", "MobileBertForMaskedLM"),
@@ -155,6 +158,7 @@
         # Model for Causal LM mapping
         ("aria_text", "AriaTextForCausalLM"),
         ("bart", "BartForCausalLM"),
+        ("camembert", "CamembertForCausalLM"),
         ("bert", "BertLMHeadModel"),
         ("bert-generation", "BertGenerationDecoder"),
         ("gemma", "GemmaForCausalLM"),
@@ -274,6 +278,7 @@
         ("bart", "BartForConditionalGeneration"),
         ("bert", "BertForMaskedLM"),
         ("roberta", "RobertaForMaskedLM"),
+        ("camembert", "CamembertForMaskedLM"),
         ("deberta", "DebertaForMaskedLM"),
         ("deberta-v2", "DebertaV2ForMaskedLM"),
         ("mobilebert", "MobileBertForMaskedLM"),
@@ -340,6 +345,7 @@
         # Model for Sequence Classification mapping
         ("albert", "AlbertForSequenceClassification"),
         ("bart", "BartForSequenceClassification"),
+        ("camembert", "CamembertForSequenceClassification"),
         ("bert", "BertForSequenceClassification"),
         ("roberta", "RobertaForSequenceClassification"),
         ("deberta", "DebertaForSequenceClassification"),
@@ -373,6 +379,7 @@
         ("albert", "AlbertForQuestionAnswering"),
         ("bart", "BartForQuestionAnswering"),
         ("bert", "BertForQuestionAnswering"),
+        ("camembert", "CamembertForQuestionAnswering"),
         ("roberta", "RobertaForQuestionAnswering"),
         ("deberta", "DebertaForQuestionAnswering"),
         ("deberta-v2", "DebertaV2ForQuestionAnswering"),
@@ -407,6 +414,7 @@
         # Model for Token Classification mapping
         ("albert", "AlbertForTokenClassification"),
         ("bert", "BertForTokenClassification"),
+        ("camembert", "CamembertForTokenClassification"),
         ("deberta", "DebertaForTokenClassification"),
         ("deberta-v2", "DebertaV2ForTokenClassification"),
         ("glm", "GlmForTokenClassification"),
@@ -433,6 +441,7 @@
 MODEL_FOR_MULTIPLE_CHOICE_MAPPING_NAMES = OrderedDict(
     [
         # Model for Multiple Choice mapping
+        ("camembert", "CamembertForMultipleChoice"),
         ("albert", "AlbertForMultipleChoice"),
         ("bert", "BertForMultipleChoice"),
         ("roberta", "RobertaForMultipleChoice"),

mindone/transformers/models/camembert/__init__.py

@@ -0,0 +1,26 @@
+# Copyright 2024 The HuggingFace Team. All rights reserved.
+#
+# This code is adapted from https://github.com/huggingface/transformers
+# with modifications to run transformers on mindspore.
+#
+# 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 .modeling_camembert import (
+    CamembertForCausalLM,
+    CamembertForMaskedLM,
+    CamembertForMultipleChoice,
+    CamembertForQuestionAnswering,
+    CamembertForSequenceClassification,
+    CamembertForTokenClassification,
+    CamembertModel,
+    CamembertPreTrainedModel,
+)