support telechat2

docs/source/en/_toctree.yml

@@ -596,6 +596,8 @@
         title: T5v1.1
       - local: model_doc/tapex
         title: TAPEX
+      - local: model_doc/telechat2
+        title: TeleChat2
       - local: model_doc/transfo-xl
         title: Transformer XL
       - local: model_doc/ul2

docs/source/en/model_doc/telechat2.md

@@ -0,0 +1,38 @@
+<!--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.
+
+⚠️ Note that this file is in Markdown but contain specific syntax for our doc-builder (similar to MDX) that may not be
+rendered properly in your Markdown viewer.
+
+-->
+
+# TeleChat2
+
+## Overview
+
+The TeleChat2 model was proposed in [<INSERT PAPER NAME HERE>](<INSERT PAPER LINK HERE>) by <INSERT AUTHORS HERE>.
+<INSERT SHORT SUMMARY HERE>
+
+The abstract from the paper is the following:
+
+*<INSERT PAPER ABSTRACT HERE>*
+
+Tips:
+
+<INSERT TIPS ABOUT MODEL HERE>
+
+This model was contributed by [INSERT YOUR HF USERNAME HERE](https://huggingface.co/<INSERT YOUR HF USERNAME HERE>).
+The original code can be found [here](<INSERT LINK TO GITHUB REPO HERE>).
+
+
+## TeleChat2Config
+
+[[autodoc]] TeleChat2Config

src/transformers/__init__.py

@@ -541,6 +541,7 @@
     "models.levit": ["LevitConfig"],
     "models.lilt": ["LiltConfig"],
     "models.llama": ["LlamaConfig"],
+    "models.telechat2": ["TeleChat2Config"],
     "models.llava": [
         "LlavaConfig",
         "LlavaProcessor",
@@ -1032,6 +1033,7 @@
     _import_structure["models.gpt_sw3"].append("GPTSw3Tokenizer")
     _import_structure["models.layoutxlm"].append("LayoutXLMTokenizer")
     _import_structure["models.llama"].append("LlamaTokenizer")
+    _import_structure["models.telechat2"].append("TeleChat2Tokenizer")
     _import_structure["models.m2m_100"].append("M2M100Tokenizer")
     _import_structure["models.marian"].append("MarianTokenizer")
     _import_structure["models.mbart"].append("MBartTokenizer")
@@ -2643,6 +2645,16 @@
             "LlamaPreTrainedModel",
         ]
     )
+    _import_structure["models.telechat2"].extend(
+        [
+            "TeleChat2ForCausalLM",
+            "TeleChat2ForQuestionAnswering",
+            "TeleChat2ForSequenceClassification",
+            "TeleChat2ForTokenClassification",
+            "TeleChat2Model",
+            "TeleChat2PreTrainedModel",
+        ]
+    )
     _import_structure["models.llava"].extend(
         [
             "LlavaForConditionalGeneration",
@@ -4848,6 +4860,7 @@
     )
     _import_structure["models.gptj"].extend(["FlaxGPTJForCausalLM", "FlaxGPTJModel", "FlaxGPTJPreTrainedModel"])
     _import_structure["models.llama"].extend(["FlaxLlamaForCausalLM", "FlaxLlamaModel", "FlaxLlamaPreTrainedModel"])
+    _import_structure["models.telechat2"].extend(["FlaxTeleChat2ForCausalLM", "FlaxTeleChat2Model", "FlaxTeleChat2PreTrainedModel"])
     _import_structure["models.gemma"].extend(["FlaxGemmaForCausalLM", "FlaxGemmaModel", "FlaxGemmaPreTrainedModel"])
     _import_structure["models.longt5"].extend(
         [
@@ -5504,6 +5517,7 @@
     from .models.levit import LevitConfig
     from .models.lilt import LiltConfig
     from .models.llama import LlamaConfig
+    from .models.telechat2 import TeleChat2Config
     from .models.llava import (
         LlavaConfig,
         LlavaProcessor,
@@ -6051,6 +6065,7 @@
         from .models.gpt_sw3 import GPTSw3Tokenizer
         from .models.layoutxlm import LayoutXLMTokenizer
         from .models.llama import LlamaTokenizer
+        from .models.telechat2 import TeleChat2Tokenizer
         from .models.m2m_100 import M2M100Tokenizer
         from .models.marian import MarianTokenizer
         from .models.mbart import MBartTokenizer
@@ -7402,6 +7417,14 @@
             LlamaModel,
             LlamaPreTrainedModel,
         )
+        from .models.telechat2 import (
+            TeleChat2ForCausalLM,
+            TeleChat2ForQuestionAnswering,
+            TeleChat2ForSequenceClassification,
+            TeleChat2ForTokenClassification,
+            TeleChat2Model,
+            TeleChat2PreTrainedModel,
+        )
         from .models.llava import (
             LlavaForConditionalGeneration,
             LlavaPreTrainedModel,
@@ -9178,6 +9201,11 @@
             FlaxLlamaModel,
             FlaxLlamaPreTrainedModel,
         )
+        from .models.telechat2 import (
+            FlaxTeleChat2ForCausalLM,
+            FlaxTeleChat2Model,
+            FlaxTeleChat2PreTrainedModel,
+        )
         from .models.longt5 import (
             FlaxLongT5ForConditionalGeneration,
             FlaxLongT5Model,

src/transformers/convert_slow_tokenizer.py

@@ -1593,6 +1593,7 @@ def converted(self) -> Tokenizer:
     "SeamlessM4TTokenizer": SeamlessM4TConverter,
     "SqueezeBertTokenizer": BertConverter,
     "T5Tokenizer": T5Converter,
+    "TeleChat2Tokenizer": LlamaConverter,
     "UdopTokenizer": UdopConverter,
     "WhisperTokenizer": WhisperConverter,
     "XLMRobertaTokenizer": XLMRobertaConverter,

src/transformers/models/__init__.py

@@ -137,6 +137,7 @@
     levit,
     lilt,
     llama,
+    telechat2,
     llava,
     llava_next,
     llava_next_video,

src/transformers/models/auto/configuration_auto.py

@@ -157,6 +157,7 @@
         ("levit", "LevitConfig"),
         ("lilt", "LiltConfig"),
         ("llama", "LlamaConfig"),
+        ("telechat2", "TeleChat2Config"),
         ("llava", "LlavaConfig"),
         ("llava_next", "LlavaNextConfig"),
         ("llava_next_video", "LlavaNextVideoConfig"),
@@ -473,6 +474,7 @@
         ("levit", "LeViT"),
         ("lilt", "LiLT"),
         ("llama", "LLaMA"),
+        ("telechat2", "TeleChat2"),
         ("llama2", "Llama2"),
         ("llama3", "Llama3"),
         ("llava", "LLaVa"),

src/transformers/models/auto/modeling_auto.py

@@ -457,6 +457,7 @@
         ("switch_transformers", "SwitchTransformersForConditionalGeneration"),
         ("t5", "T5ForConditionalGeneration"),
         ("tapas", "TapasForMaskedLM"),
+        ("telechat2", "TeleChat2ForCausalLM"),
         ("transfo-xl", "TransfoXLLMHeadModel"),
         ("wav2vec2", "Wav2Vec2ForMaskedLM"),
         ("whisper", "WhisperForConditionalGeneration"),

src/transformers/models/auto/tokenization_auto.py

@@ -261,6 +261,7 @@
                     "LlamaTokenizerFast" if is_tokenizers_available() else None,
                 ),
             ),
+            ("telechat2", ("LlamaTokenizer", "LlamaTokenizerFast" if is_tokenizers_available() else None)),
             ("llava", ("LlamaTokenizer", "LlamaTokenizerFast" if is_tokenizers_available() else None)),
             ("llava_next", ("LlamaTokenizer", "LlamaTokenizerFast" if is_tokenizers_available() else None)),
             ("llava_next_video", ("LlamaTokenizer", "LlamaTokenizerFast" if is_tokenizers_available() else None)),

src/transformers/models/telechat2/__init__.py

@@ -0,0 +1,59 @@
+# Copyright 2024 EleutherAI and The HuggingFace Inc. 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 (
+    OptionalDependencyNotAvailable,
+    _LazyModule,
+    is_torch_available,
+)
+
+
+_import_structure = {
+    "configuration_telechat2": ["TeleChat2Config"],
+}
+
+try:
+    if not is_torch_available():
+        raise OptionalDependencyNotAvailable()
+except OptionalDependencyNotAvailable:
+    pass
+else:
+    _import_structure["modeling_telechat2"] = [
+        "TeleChat2PreTrainedModel",
+        "TeleChat2Model",
+        "TeleChat2ForCausalLM",
+    ]
+
+
+if TYPE_CHECKING:
+    from .configuration_telechat2 import TeleChat2Config
+
+    try:
+        if not is_torch_available():
+            raise OptionalDependencyNotAvailable()
+    except OptionalDependencyNotAvailable:
+        pass
+    else:
+        from .modeling_qwen2 import (
+            TeleChat2PreTrainedModel,
+            TeleChat2Model,
+            TeleChat2ForCausalLM,
+        )
+
+
+else:
+    import sys
+
+    sys.modules[__name__] = _LazyModule(__name__, globals()["__file__"], _import_structure, module_spec=__spec__)