Add T5LA models

docs/source/en/_toctree.yml

@@ -693,6 +693,8 @@
         title: T5
       - local: model_doc/t5gemma
         title: T5Gemma
+      - local: model_doc/t5la
+        title: T5LA
       - local: model_doc/t5v1.1
         title: T5v1.1
       - local: model_doc/tapex

docs/source/en/model_doc/t5la.md

@@ -0,0 +1,59 @@
+<!--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.
+
+-->
+*This model was released on {release_date} and added to Hugging Face Transformers on 2025-08-05.*
+
+
+# T5LA
+
+## Overview
+
+The T5LA 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>).
+
+## Usage examples
+
+<INSERT SOME NICE EXAMPLES HERE>
+
+## T5LaConfig
+
+[[autodoc]] T5LaConfig
+
+## T5LaForConditionalGeneration
+
+[[autodoc]] T5LaForConditionalGeneration
+
+## T5LaPreTrainedModel
+
+[[autodoc]] T5LaPreTrainedModel
+    - forward
+
+## load_tf_weights_in_t5
+
+[[autodoc]] load_tf_weights_in_t5

src/transformers/models/__init__.py

@@ -320,6 +320,7 @@
     from .switch_transformers import *
     from .t5 import *
     from .t5gemma import *
+    from .t5la import *
     from .table_transformer import *
     from .tapas import *
     from .textnet import *

src/transformers/models/auto/configuration_auto.py

@@ -378,6 +378,7 @@
         ("switch_transformers", "SwitchTransformersConfig"),
         ("t5", "T5Config"),
         ("t5gemma", "T5GemmaConfig"),
+        ("t5la", "T5LaConfig"),
         ("table-transformer", "TableTransformerConfig"),
         ("tapas", "TapasConfig"),
         ("textnet", "TextNetConfig"),
@@ -815,6 +816,7 @@
         ("switch_transformers", "SwitchTransformers"),
         ("t5", "T5"),
         ("t5gemma", "T5Gemma"),
+        ("t5la", "T5LA"),
         ("t5v1.1", "T5v1.1"),
         ("table-transformer", "Table Transformer"),
         ("tapas", "TAPAS"),

src/transformers/models/auto/modeling_auto.py

@@ -364,6 +364,7 @@ class _BaseModelWithGenerate(PreTrainedModel, GenerationMixin):
         ("switch_transformers", "SwitchTransformersModel"),
         ("t5", "T5Model"),
         ("t5gemma", "T5GemmaModel"),
+        ("t5la", "T5LaForConditionalGeneration"),
         ("table-transformer", "TableTransformerModel"),
         ("tapas", "TapasModel"),
         ("textnet", "TextNetModel"),
@@ -492,6 +493,7 @@ class _BaseModelWithGenerate(PreTrainedModel, GenerationMixin):
         ("switch_transformers", "SwitchTransformersForConditionalGeneration"),
         ("t5", "T5ForConditionalGeneration"),
         ("t5gemma", "T5GemmaForConditionalGeneration"),
+        ("t5la", "T5LaForConditionalGeneration"),
         ("tapas", "TapasForMaskedLM"),
         ("transfo-xl", "TransfoXLLMHeadModel"),
         ("tvlt", "TvltForPreTraining"),
@@ -1149,6 +1151,7 @@ class _BaseModelWithGenerate(PreTrainedModel, GenerationMixin):
         ("switch_transformers", "SwitchTransformersForConditionalGeneration"),
         ("t5", "T5ForConditionalGeneration"),
         ("t5gemma", "T5GemmaForConditionalGeneration"),
+        ("t5la", "T5LaForConditionalGeneration"),
         ("umt5", "UMT5ForConditionalGeneration"),
         ("voxtral", "VoxtralForConditionalGeneration"),
         ("xlm-prophetnet", "XLMProphetNetForConditionalGeneration"),

src/transformers/models/t5la/__init__.py

@@ -0,0 +1,29 @@
+# coding=utf-8
+# 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.
+
+from typing import TYPE_CHECKING
+
+from ...utils import _LazyModule
+from ...utils.import_utils import define_import_structure
+
+
+if TYPE_CHECKING:
+    from .configuration_t5la import *
+    from .modeling_t5la import *
+else:
+    import sys
+
+    _file = globals()["__file__"]
+    sys.modules[__name__] = _LazyModule(__name__, _file, define_import_structure(_file), module_spec=__spec__)

src/transformers/models/t5la/configuration_t5la.py

@@ -0,0 +1,149 @@
+#                🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨
+#           This file was automatically generated from src/transformers/models/t5la/modular_t5la.py.
+#               Do NOT edit this file manually as any edits will be overwritten by the generation of
+#             the file from the modular. If any change should be done, please apply the change to the
+#                          modular_t5la.py file directly. One of our CI enforces this.
+#                🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨
+# coding=utf-8
+# 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.
+
+from ...configuration_utils import PretrainedConfig
+
+
+class T5LaConfig(PretrainedConfig):
+    r"""
+    This is the configuration class to store the configuration of a [`T5LaModel`] or a [`TFT5Model`]. It is used to
+    instantiate a T5LA 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 T5LA
+    [hrezaei/T5LA](https://huggingface.co/hrezaei/T5LA) architecture.
+
+    Configuration objects inherit from [`PretrainedConfig`] and can be used to control the model outputs. Read the
+    documentation from [`PretrainedConfig`] for more information.
+
+    Arguments:
+        vocab_size (`int`, *optional*, defaults to 32128):
+            Vocabulary size of the T5LA model. Defines the number of different tokens that can be represented by the
+            `inputs_ids` passed when calling [`T5LaModel`] or [`TFT5Model`].
+        d_model (`int`, *optional*, defaults to 512):
+            Size of the encoder layers and the pooler layer.
+        d_kv (`int`, *optional*, defaults to 64):
+            Size of the key, query, value projections per attention head. The `inner_dim` of the projection layer will
+            be defined as `num_heads * d_kv`.
+        d_ff (`int`, *optional*, defaults to 2048):
+            Size of the intermediate feed forward layer in each `T5LaBlock`.
+        num_layers (`int`, *optional*, defaults to 6):
+            Number of hidden layers in the Transformer encoder.
+        num_decoder_layers (`int`, *optional*):
+            Number of hidden layers in the Transformer decoder. Will use the same value as `num_layers` if not set.
+        num_heads (`int`, *optional*, defaults to 8):
+            Number of attention heads for each attention layer in the Transformer encoder.
+        relative_attention_num_buckets (`int`, *optional*, defaults to 32):
+            The number of buckets to use for each attention layer.
+        relative_attention_max_distance (`int`, *optional*, defaults to 128):
+            The maximum distance of the longer sequences for the bucket separation.
+        dropout_rate (`float`, *optional*, defaults to 0.1):
+            The ratio for all dropout layers.
+        classifier_dropout (`float`, *optional*, defaults to 0.0):
+            The dropout ratio for classifier.
+        layer_norm_eps (`float`, *optional*, defaults to 1e-6):
+            The epsilon used by the layer normalization layers.
+        initializer_factor (`float`, *optional*, defaults to 1):
+            A factor for initializing all weight matrices (should be kept to 1, used internally for initialization
+            testing).
+        feed_forward_proj (`string`, *optional*, defaults to `"relu"`):
+            Type of feed forward layer to be used. Should be one of `"relu"` or `"gated-gelu"`. T5Lav1.1 uses the
+            `"gated-gelu"` feed forward projection. Original T5LA uses `"relu"`.
+        use_cache (`bool`, *optional*, defaults to `True`):
+            Whether or not the model should return the last key/values attentions (not used by all models).
+        lookahead_type (`string`, *optional*, defaults to "la"):
+            other options are "laa", "laa2", and "lae". For more details, see the paper:
+            https://aclanthology.org/2025.acl-srw.41/
+        lookahead_size (`int`, *optional*, defaults to 1):
+            Number of future tokens to be predicted after the immediately next token. The K parameter as explained in
+            the paper https://aclanthology.org/2025.acl-srw.41/
+    """
+
+    model_type = "t5la"
+    keys_to_ignore_at_inference = ["past_key_values", "lookahead_logits", "lookahead_loss"]
+    attribute_map = {
+        "hidden_size": "d_model",
+        "num_attention_heads": "num_heads",
+        "num_hidden_layers": "num_layers",
+        "head_dim": "d_kv",
+    }
+
+    def __init__(
+        self,
+        vocab_size=32128,
+        d_model=512,
+        d_kv=64,
+        d_ff=2048,
+        num_layers=6,
+        num_decoder_layers=None,
+        num_heads=8,
+        relative_attention_num_buckets=32,
+        relative_attention_max_distance=128,
+        dropout_rate=0.1,
+        layer_norm_epsilon=1e-6,
+        initializer_factor=1.0,
+        feed_forward_proj="relu",
+        is_encoder_decoder=True,
+        use_cache=True,
+        pad_token_id=0,
+        eos_token_id=1,
+        classifier_dropout=0.0,
+        lookahead_type="la",
+        lookahead_size=1,
+        **kwargs,
+    ):
+        super().__init__(
+            pad_token_id=pad_token_id,
+            eos_token_id=eos_token_id,
+            is_encoder_decoder=is_encoder_decoder,
+            **kwargs,
+        )
+        self.vocab_size = vocab_size
+        self.d_model = d_model
+        self.d_kv = d_kv
+        self.d_ff = d_ff
+        self.num_layers = num_layers
+        self.num_decoder_layers = (
+            num_decoder_layers if num_decoder_layers is not None else self.num_layers
+        )  # default = symmetry
+        self.num_heads = num_heads
+        self.relative_attention_num_buckets = relative_attention_num_buckets
+        self.relative_attention_max_distance = relative_attention_max_distance
+        self.dropout_rate = dropout_rate
+        self.classifier_dropout = classifier_dropout
+        self.layer_norm_epsilon = layer_norm_epsilon
+        self.initializer_factor = initializer_factor
+        self.feed_forward_proj = feed_forward_proj
+        self.use_cache = use_cache
+
+        act_info = self.feed_forward_proj.split("-")
+        self.dense_act_fn = act_info[-1]
+        self.is_gated_act = act_info[0] == "gated"
+
+        if len(act_info) > 1 and act_info[0] != "gated" or len(act_info) > 2:
+            raise ValueError(
+                f"`feed_forward_proj`: {feed_forward_proj} is not a valid activation function of the dense layer. "
+                "Please make sure `feed_forward_proj` is of the format `gated-{ACT_FN}` or `{ACT_FN}`, e.g. "
+                "'gated-gelu' or 'relu'"
+            )
+        self.lookahead_type = lookahead_type
+        self.lookahead_size = lookahead_size
+
+
+__all__ = ["T5LaConfig"]