T5gemma2

docs/source/en/_toctree.yml

@@ -716,6 +716,8 @@
         title: T5
       - local: model_doc/t5gemma
         title: T5Gemma
+      - local: model_doc/t5gemma2
+        title: T5Gemma2
       - local: model_doc/t5v1.1
         title: T5v1.1
       - local: model_doc/tapex

docs/source/en/model_doc/t5gemma2.md

@@ -0,0 +1,109 @@
+
+<!--Copyright 2025 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.
+
+-->
+<div style="float: right;">
+    <div class="flex flex-wrap space-x-1">
+        <img alt="PyTorch" src="https://img.shields.io/badge/PyTorch-DE3412?style=flat&logo=pytorch&logoColor=white">
+        <img alt="FlashAttention" src="https://img.shields.io/badge/%E2%9A%A1%EF%B8%8E%20FlashAttention-eae0c8?style=flat">
+        <img alt="SDPA" src="https://img.shields.io/badge/SDPA-DE3412?style=flat&logo=pytorch&logoColor=white">
+    </div>
+</div>
+
+# T5Gemma 2
+
+T5Gemma 2 is a family of pretrained encoder-decoder large language models with strong multilingual, multimodal and long-context capability, available in 270M-270M, 1B-1B and 4B-4B parameters. Following T5Gemma, it is built via model adaptation (based on Gemma 3) using UL2. The architecture is similar to T5Gemma and Gemma 3, enhanced with tied word embeddings and merged self- and cross-attention to save model parameters.
+
+> [!TIP]
+> Click on the T5Gemma 2 models in the right sidebar for more examples of how to apply T5Gemma 2 to different language tasks.
+
+The example below demonstrates how to chat with the model with [`Pipeline`] or the [`AutoModel`] class, and from the command line.
+
+<hfoptions id="usage">
+<hfoption id="Pipeline">
+
+```python
+import torch
+from transformers import pipeline
+
+generator = pipeline(
+    "image-text-to-text",
+    model="google/t5gemma-2-270m-270m",
+    dtype=torch.bfloat16,
+    device_map="auto",
+)
+
+generator(
+    "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/bee.jpg",
+    text="<start_of_image> in this image, there is",
+    generate_kwargs={"do_sample": False, "max_new_tokens": 50},
+)
+```
+
+</hfoption>
+<hfoption id="AutoModel">
+
+```python
+import torch
+import requests
+from PIL import Image
+from transformers import AutoProcessor, AutoModelForSeq2SeqLM
+
+processor = AutoProcessor.from_pretrained("google/t5gemma-2-270m-270m")
+model = AutoModelForSeq2SeqLM.from_pretrained(
+    "google/t5gemma-2-270m-270m",
+    device_map="auto",
+    dtype=torch.bfloat16,
+)
+
+url = "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/bee.jpg"
+image = Image.open(requests.get(url, stream=True).raw)
+prompt = "<start_of_image> in this image, there is"
+
+model_inputs = processor(text=prompt, images=image, return_tensors="pt")
+generation = model.generate(**model_inputs, max_new_tokens=20, do_sample=False)
+print(processor.decode(generation[0]))
+```
+
+</hfoption>
+</hfoptions>
+
+## T5Gemma2Config
+
+[[autodoc]] T5Gemma2Config
+
+## T5Gemma2ModuleConfig
+
+[[autodoc]] T5Gemma2ModuleConfig
+
+## T5Gemma2Model
+
+[[autodoc]] T5Gemma2Model
+    - forward
+
+## T5Gemma2ForConditionalGeneration
+
+[[autodoc]] T5Gemma2ForConditionalGeneration
+    - forward
+
+## T5Gemma2ForSequenceClassification
+
+[[autodoc]] T5Gemma2ForSequenceClassification
+    - forward
+
+## T5Gemma2ForTokenClassification
+
+[[autodoc]] T5Gemma2ForTokenClassification
+    - forward

src/transformers/models/__init__.py

@@ -335,6 +335,7 @@
     from .switch_transformers import *
     from .t5 import *
     from .t5gemma import *
+    from .t5gemma2 import *
     from .table_transformer import *
     from .tapas import *
     from .textnet import *

src/transformers/models/auto/configuration_auto.py

@@ -396,6 +396,7 @@
         ("switch_transformers", "SwitchTransformersConfig"),
         ("t5", "T5Config"),
         ("t5gemma", "T5GemmaConfig"),
+        ("t5gemma2", "T5Gemma2Config"),
         ("table-transformer", "TableTransformerConfig"),
         ("tapas", "TapasConfig"),
         ("textnet", "TextNetConfig"),
@@ -855,6 +856,7 @@
         ("switch_transformers", "SwitchTransformers"),
         ("t5", "T5"),
         ("t5gemma", "T5Gemma"),
+        ("t5gemma2", "T5Gemma2"),
         ("t5v1.1", "T5v1.1"),
         ("table-transformer", "Table Transformer"),
         ("tapas", "TAPAS"),

src/transformers/models/auto/image_processing_auto.py

@@ -176,6 +176,7 @@
             ("swin", ("ViTImageProcessor", "ViTImageProcessorFast")),
             ("swin2sr", ("Swin2SRImageProcessor", "Swin2SRImageProcessorFast")),
             ("swinv2", ("ViTImageProcessor", "ViTImageProcessorFast")),
+            ("t5gemma2", ("Gemma3ImageProcessor", "Gemma3ImageProcessorFast")),
             ("table-transformer", ("DetrImageProcessor", "DetrImageProcessorFast")),
             ("textnet", ("TextNetImageProcessor", "TextNetImageProcessorFast")),
             ("timesformer", ("VideoMAEImageProcessor", None)),

src/transformers/models/auto/modeling_auto.py

@@ -382,6 +382,7 @@ class _BaseModelWithGenerate(PreTrainedModel, GenerationMixin):
         ("switch_transformers", "SwitchTransformersModel"),
         ("t5", "T5Model"),
         ("t5gemma", "T5GemmaModel"),
+        ("t5gemma2", "T5Gemma2Model"),
         ("table-transformer", "TableTransformerModel"),
         ("tapas", "TapasModel"),
         ("textnet", "TextNetModel"),
@@ -513,6 +514,7 @@ class _BaseModelWithGenerate(PreTrainedModel, GenerationMixin):
         ("switch_transformers", "SwitchTransformersForConditionalGeneration"),
         ("t5", "T5ForConditionalGeneration"),
         ("t5gemma", "T5GemmaForConditionalGeneration"),
+        ("t5gemma2", "T5Gemma2ForConditionalGeneration"),
         ("tapas", "TapasForMaskedLM"),
         ("transfo-xl", "TransfoXLLMHeadModel"),
         ("tvlt", "TvltForPreTraining"),
@@ -612,6 +614,7 @@ class _BaseModelWithGenerate(PreTrainedModel, GenerationMixin):
         ("switch_transformers", "SwitchTransformersForConditionalGeneration"),
         ("t5", "T5ForConditionalGeneration"),
         ("t5gemma", "T5GemmaForConditionalGeneration"),
+        ("t5gemma2", "T5Gemma2ForConditionalGeneration"),
         ("tapas", "TapasForMaskedLM"),
         ("transfo-xl", "TransfoXLLMHeadModel"),
         ("wav2vec2", "Wav2Vec2ForMaskedLM"),
@@ -1059,6 +1062,7 @@ class _BaseModelWithGenerate(PreTrainedModel, GenerationMixin):
         ("qwen3_vl_moe", "Qwen3VLMoeForConditionalGeneration"),
         ("shieldgemma2", "Gemma3ForConditionalGeneration"),
         ("smolvlm", "SmolVLMForConditionalGeneration"),
+        ("t5gemma2", "T5Gemma2ForConditionalGeneration"),
         ("udop", "UdopForConditionalGeneration"),
         ("video_llama_3", "VideoLlama3ForConditionalGeneration"),
         ("vipllava", "VipLlavaForConditionalGeneration"),
@@ -1185,6 +1189,7 @@ class _BaseModelWithGenerate(PreTrainedModel, GenerationMixin):
         ("switch_transformers", "SwitchTransformersForConditionalGeneration"),
         ("t5", "T5ForConditionalGeneration"),
         ("t5gemma", "T5GemmaForConditionalGeneration"),
+        ("t5gemma2", "T5Gemma2ForConditionalGeneration"),
         ("umt5", "UMT5ForConditionalGeneration"),
         ("voxtral", "VoxtralForConditionalGeneration"),
         ("xlm-prophetnet", "XLMProphetNetForConditionalGeneration"),
@@ -1314,6 +1319,7 @@ class _BaseModelWithGenerate(PreTrainedModel, GenerationMixin):
         ("starcoder2", "Starcoder2ForSequenceClassification"),
         ("t5", "T5ForSequenceClassification"),
         ("t5gemma", "T5GemmaForSequenceClassification"),
+        ("t5gemma2", "T5Gemma2ForSequenceClassification"),
         ("tapas", "TapasForSequenceClassification"),
         ("transfo-xl", "TransfoXLForSequenceClassification"),
         ("umt5", "UMT5ForSequenceClassification"),
@@ -1521,6 +1527,7 @@ class _BaseModelWithGenerate(PreTrainedModel, GenerationMixin):
         ("starcoder2", "Starcoder2ForTokenClassification"),
         ("t5", "T5ForTokenClassification"),
         ("t5gemma", "T5GemmaForTokenClassification"),
+        ("t5gemma2", "T5Gemma2ForTokenClassification"),
         ("umt5", "UMT5ForTokenClassification"),
         ("xlm", "XLMForTokenClassification"),
         ("xlm-roberta", "XLMRobertaForTokenClassification"),

src/transformers/models/auto/processing_auto.py

@@ -137,6 +137,7 @@
         ("speech_to_text", "Speech2TextProcessor"),
         ("speech_to_text_2", "Speech2Text2Processor"),
         ("speecht5", "SpeechT5Processor"),
+        ("t5gemma2", "Gemma3Processor"),
         ("trocr", "TrOCRProcessor"),
         ("tvlt", "TvltProcessor"),
         ("tvp", "TvpProcessor"),

src/transformers/models/auto/tokenization_auto.py

@@ -689,6 +689,13 @@
                 "GemmaTokenizerFast" if is_tokenizers_available() else None,
             ),
         ),
+        (
+            "t5gemma2",
+            (
+                "GemmaTokenizer" if is_sentencepiece_available() else None,
+                "GemmaTokenizerFast" if is_tokenizers_available() else None,
+            ),
+        ),
         ("tapas", ("TapasTokenizer", None)),
         ("tapex", ("TapexTokenizer", None)),
         ("transfo-xl", ("TransfoXLTokenizer", None)),

src/transformers/models/t5gemma/__init__.py

@@ -18,8 +18,8 @@
 
 
 if TYPE_CHECKING:
-    from .configuration_encdecgemma2 import *
-    from .modeling_encdecgemma2 import *
+    from .configuration_t5gemma import *
+    from .modeling_t5gemma import *
 else:
     import sys
 

src/transformers/models/t5gemma2/__init__.py

@@ -0,0 +1,27 @@
+# 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
+
+
+if TYPE_CHECKING:
+    from .configuration_t5gemma2 import *
+    from .modeling_t5gemma2 import *
+else:
+    import sys
+
+    _file = globals()["__file__"]
+    sys.modules[__name__] = _LazyModule(__name__, _file, define_import_structure(_file), module_spec=__spec__)