mlx-beam-vision 0.1.0a5__tar.gz

mlx_beam_vision-0.1.0a5/.gitignore

@@ -0,0 +1,14 @@
+__pycache__/
+*.py[cod]
+*.egg-info/
+.venv/
+dist/
+build/
+.ruff_cache/
+.pytest_cache/
+mlx_beam/_version.py
+packages/*/*/_version.py
+docs/site/node_modules/
+docs/site/dist/
+docs/site/.astro/
+.DS_Store

mlx_beam_vision-0.1.0a5/LICENSE

@@ -0,0 +1,202 @@
+
+                                 Apache License
+                           Version 2.0, January 2004
+                        http://www.apache.org/licenses/
+
+   TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
+
+   1. Definitions.
+
+      "License" shall mean the terms and conditions for use, reproduction,
+      and distribution as defined by Sections 1 through 9 of this document.
+
+      "Licensor" shall mean the copyright owner or entity authorized by
+      the copyright owner that is granting the License.
+
+      "Legal Entity" shall mean the union of the acting entity and all
+      other entities that control, are controlled by, or are under common
+      control with that entity. For the purposes of this definition,
+      "control" means (i) the power, direct or indirect, to cause the
+      direction or management of such entity, whether by contract or
+      otherwise, or (ii) ownership of fifty percent (50%) or more of the
+      outstanding shares, or (iii) beneficial ownership of such entity.
+
+      "You" (or "Your") shall mean an individual or Legal Entity
+      exercising permissions granted by this License.
+
+      "Source" form shall mean the preferred form for making modifications,
+      including but not limited to software source code, documentation
+      source, and configuration files.
+
+      "Object" form shall mean any form resulting from mechanical
+      transformation or translation of a Source form, including but
+      not limited to compiled object code, generated documentation,
+      and conversions to other media types.
+
+      "Work" shall mean the work of authorship, whether in Source or
+      Object form, made available under the License, as indicated by a
+      copyright notice that is included in or attached to the work
+      (an example is provided in the Appendix below).
+
+      "Derivative Works" shall mean any work, whether in Source or Object
+      form, that is based on (or derived from) the Work and for which the
+      editorial revisions, annotations, elaborations, or other modifications
+      represent, as a whole, an original work of authorship. For the purposes
+      of this License, Derivative Works shall not include works that remain
+      separable from, or merely link (or bind by name) to the interfaces of,
+      the Work and Derivative Works thereof.
+
+      "Contribution" shall mean any work of authorship, including
+      the original version of the Work and any modifications or additions
+      to that Work or Derivative Works thereof, that is intentionally
+      submitted to Licensor for inclusion in the Work by the copyright owner
+      or by an individual or Legal Entity authorized to submit on behalf of
+      the copyright owner. For the purposes of this definition, "submitted"
+      means any form of electronic, verbal, or written communication sent
+      to the Licensor or its representatives, including but not limited to
+      communication on electronic mailing lists, source code control systems,
+      and issue tracking systems that are managed by, or on behalf of, the
+      Licensor for the purpose of discussing and improving the Work, but
+      excluding communication that is conspicuously marked or otherwise
+      designated in writing by the copyright owner as "Not a Contribution."
+
+      "Contributor" shall mean Licensor and any individual or Legal Entity
+      on behalf of whom a Contribution has been received by Licensor and
+      subsequently incorporated within the Work.
+
+   2. Grant of Copyright License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      copyright license to reproduce, prepare Derivative Works of,
+      publicly display, publicly perform, sublicense, and distribute the
+      Work and such Derivative Works in Source or Object form.
+
+   3. Grant of Patent License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      (except as stated in this section) patent license to make, have made,
+      use, offer to sell, sell, import, and otherwise transfer the Work,
+      where such license applies only to those patent claims licensable
+      by such Contributor that are necessarily infringed by their
+      Contribution(s) alone or by combination of their Contribution(s)
+      with the Work to which such Contribution(s) was submitted. If You
+      institute patent litigation against any entity (including a
+      cross-claim or counterclaim in a lawsuit) alleging that the Work
+      or a Contribution incorporated within the Work constitutes direct
+      or contributory patent infringement, then any patent licenses
+      granted to You under this License for that Work shall terminate
+      as of the date such litigation is filed.
+
+   4. Redistribution. You may reproduce and distribute copies of the
+      Work or Derivative Works thereof in any medium, with or without
+      modifications, and in Source or Object form, provided that You
+      meet the following conditions:
+
+      (a) You must give any other recipients of the Work or
+          Derivative Works a copy of this License; and
+
+      (b) You must cause any modified files to carry prominent notices
+          stating that You changed the files; and
+
+      (c) You must retain, in the Source form of any Derivative Works
+          that You distribute, all copyright, patent, trademark, and
+          attribution notices from the Source form of the Work,
+          excluding those notices that do not pertain to any part of
+          the Derivative Works; and
+
+      (d) If the Work includes a "NOTICE" text file as part of its
+          distribution, then any Derivative Works that You distribute must
+          include a readable copy of the attribution notices contained
+          within such NOTICE file, excluding those notices that do not
+          pertain to any part of the Derivative Works, in at least one
+          of the following places: within a NOTICE text file distributed
+          as part of the Derivative Works; within the Source form or
+          documentation, if provided along with the Derivative Works; or,
+          within a display generated by the Derivative Works, if and
+          wherever such third-party notices normally appear. The contents
+          of the NOTICE file are for informational purposes only and
+          do not modify the License. You may add Your own attribution
+          notices within Derivative Works that You distribute, alongside
+          or as an addendum to the NOTICE text from the Work, provided
+          that such additional attribution notices cannot be construed
+          as modifying the License.
+
+      You may add Your own copyright statement to Your modifications and
+      may provide additional or different license terms and conditions
+      for use, reproduction, or distribution of Your modifications, or
+      for any such Derivative Works as a whole, provided Your use,
+      reproduction, and distribution of the Work otherwise complies with
+      the conditions stated in this License.
+
+   5. Submission of Contributions. Unless You explicitly state otherwise,
+      any Contribution intentionally submitted for inclusion in the Work
+      by You to the Licensor shall be under the terms and conditions of
+      this License, without any additional terms or conditions.
+      Notwithstanding the above, nothing herein shall supersede or modify
+      the terms of any separate license agreement you may have executed
+      with Licensor regarding such Contributions.
+
+   6. Trademarks. This License does not grant permission to use the trade
+      names, trademarks, service marks, or product names of the Licensor,
+      except as required for reasonable and customary use in describing the
+      origin of the Work and reproducing the content of the NOTICE file.
+
+   7. Disclaimer of Warranty. Unless required by applicable law or
+      agreed to in writing, Licensor provides the Work (and each
+      Contributor provides its Contributions) on an "AS IS" BASIS,
+      WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
+      implied, including, without limitation, any warranties or conditions
+      of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
+      PARTICULAR PURPOSE. You are solely responsible for determining the
+      appropriateness of using or redistributing the Work and assume any
+      risks associated with Your exercise of permissions under this License.
+
+   8. Limitation of Liability. In no event and under no legal theory,
+      whether in tort (including negligence), contract, or otherwise,
+      unless required by applicable law (such as deliberate and grossly
+      negligent acts) or agreed to in writing, shall any Contributor be
+      liable to You for damages, including any direct, indirect, special,
+      incidental, or consequential damages of any character arising as a
+      result of this License or out of the use or inability to use the
+      Work (including but not limited to damages for loss of goodwill,
+      work stoppage, computer failure or malfunction, or any and all
+      other commercial damages or losses), even if such Contributor
+      has been advised of the possibility of such damages.
+
+   9. Accepting Warranty or Additional Liability. While redistributing
+      the Work or Derivative Works thereof, You may choose to offer,
+      and charge a fee for, acceptance of support, warranty, indemnity,
+      or other liability obligations and/or rights consistent with this
+      License. However, in accepting such obligations, You may act only
+      on Your own behalf and on Your sole responsibility, not on behalf
+      of any other Contributor, and only if You agree to indemnify,
+      defend, and hold each Contributor harmless for any liability
+      incurred by, or claims asserted against, such Contributor by reason
+      of your accepting any such warranty or additional liability.
+
+   END OF TERMS AND CONDITIONS
+
+   APPENDIX: How to apply the Apache License to your work.
+
+      To apply the Apache License to your work, attach the following
+      boilerplate notice, with the fields enclosed by brackets "[]"
+      replaced with your own identifying information. (Don't include
+      the brackets!)  The text should be enclosed in the appropriate
+      comment syntax for the file format. We also recommend that a
+      file or class name and description of purpose be included on the
+      same "printed page" as the copyright notice for easier
+      identification within third-party archives.
+
+   Copyright [yyyy] [name of copyright owner]
+
+   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.

mlx_beam_vision-0.1.0a5/NOTICE

@@ -0,0 +1,6 @@
+mlx-beam-vision
+Copyright 2026 the mlx-beam contributors
+
+This product includes software developed as part of mlx-vlm
+(https://github.com/Blaizzy/mlx-vlm), MIT License; see
+mlx_beam_vision/_vendor/VENDORED.md and mlx_beam_vision/_vendor/mlx_vlm/LICENSE.

mlx_beam_vision-0.1.0a5/PKG-INFO

@@ -0,0 +1,60 @@
+Metadata-Version: 2.5
+Name: mlx-beam-vision
+Version: 0.1.0a5
+Summary: Vision towers for mlx-beam: images in the prompt, served by the checkpoint's own tower.
+Project-URL: Homepage, https://p4ik.github.io/mlx-beam/
+Project-URL: Repository, https://github.com/p4ik/mlx-beam
+Author: the mlx-beam contributors
+License-Expression: Apache-2.0
+License-File: LICENSE
+License-File: NOTICE
+Keywords: apple-silicon,inference,mlx,vision,vlm
+Classifier: Development Status :: 2 - Pre-Alpha
+Classifier: License :: OSI Approved :: Apache Software License
+Classifier: Operating System :: MacOS
+Classifier: Programming Language :: Python :: 3 :: Only
+Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
+Requires-Python: >=3.11
+Requires-Dist: mlx-beam
+Requires-Dist: mlx>=0.32; sys_platform == 'darwin'
+Requires-Dist: numpy
+Requires-Dist: pillow>=10
+Requires-Dist: transformers>=5.15
+Provides-Extra: dev
+Requires-Dist: mlx[cpu]>=0.32; (sys_platform == 'linux') and extra == 'dev'
+Requires-Dist: pytest>=8; extra == 'dev'
+Description-Content-Type: text/markdown
+
+# mlx-beam-vision
+
+The towers that let [mlx-beam](https://github.com/p4ik/mlx-beam) see. Install
+it next to the engine (`pip install mlx-beam[vision]`, or this package by
+name) and a checkpoint that ships a vision tower serves images: an
+`image_url` part with a `data:` URL in a chat request, the checkpoint's own
+processor renders the template, the tower encodes the image, and the
+engine's prefill takes the image's features in place of the placeholder
+tokens - beside text-only requests in the same batch, with the prefix cache
+keyed on the image's digest. Without this package the engine answers image
+input with a 400 that names the `vision` extra.
+
+Towers vendored from [mlx-vlm](https://github.com/Blaizzy/mlx-vlm) (MIT,
+see `mlx_beam_vision/_vendor/VENDORED.md`):
+
+| Family | Models | Notes |
+|---|---|---|
+| `qwen3_vl` | Qwen3-VL, Qwen3.5, Qwen3.8 | DeepStack applied: the tower's intermediate features are added after the text model's first layers |
+| `mistral3` | Mistral Small 3.x (Pixtral tower) | the image's rows are separate spans (`[IMG_BREAK]` between them) |
+| `gemma4` | Gemma 4 E2B, E4B | image only; audio is not served yet. The 12B (`gemma4_unified`, encoder-free, its image tokens attending bidirectionally in the text model) is refused with that reason until the text trunk takes such a mask |
+| `muse_glimmer` | Muse Glimmer | the projector's norm matches the text model's input norm |
+| `granite4_vision` | Granite Vision 4.1 | AnyRes tiles, window Q-Former projectors; nothing at the embedding, every projector's features added ahead of its text layer |
+
+Not yet: audio and video input, quantized towers (the towers load at model
+precision; a package that ships a quantized tower is refused with a
+message). An image above 32 megapixels is refused before it is decoded.
+
+Every family is exercised in the tests on a tiny random tower against a
+direct forward; the real towers and processors are validated on Apple
+silicon with the checkpoints themselves.
+
+`/health.vision` reports the family, the tower, the processor and the
+feature cache (encoder outputs by image digest, bounded in bytes).

mlx_beam_vision-0.1.0a5/README.md

@@ -0,0 +1,33 @@
+# mlx-beam-vision
+
+The towers that let [mlx-beam](https://github.com/p4ik/mlx-beam) see. Install
+it next to the engine (`pip install mlx-beam[vision]`, or this package by
+name) and a checkpoint that ships a vision tower serves images: an
+`image_url` part with a `data:` URL in a chat request, the checkpoint's own
+processor renders the template, the tower encodes the image, and the
+engine's prefill takes the image's features in place of the placeholder
+tokens - beside text-only requests in the same batch, with the prefix cache
+keyed on the image's digest. Without this package the engine answers image
+input with a 400 that names the `vision` extra.
+
+Towers vendored from [mlx-vlm](https://github.com/Blaizzy/mlx-vlm) (MIT,
+see `mlx_beam_vision/_vendor/VENDORED.md`):
+
+| Family | Models | Notes |
+|---|---|---|
+| `qwen3_vl` | Qwen3-VL, Qwen3.5, Qwen3.8 | DeepStack applied: the tower's intermediate features are added after the text model's first layers |
+| `mistral3` | Mistral Small 3.x (Pixtral tower) | the image's rows are separate spans (`[IMG_BREAK]` between them) |
+| `gemma4` | Gemma 4 E2B, E4B | image only; audio is not served yet. The 12B (`gemma4_unified`, encoder-free, its image tokens attending bidirectionally in the text model) is refused with that reason until the text trunk takes such a mask |
+| `muse_glimmer` | Muse Glimmer | the projector's norm matches the text model's input norm |
+| `granite4_vision` | Granite Vision 4.1 | AnyRes tiles, window Q-Former projectors; nothing at the embedding, every projector's features added ahead of its text layer |
+
+Not yet: audio and video input, quantized towers (the towers load at model
+precision; a package that ships a quantized tower is refused with a
+message). An image above 32 megapixels is refused before it is decoded.
+
+Every family is exercised in the tests on a tiny random tower against a
+direct forward; the real towers and processors are validated on Apple
+silicon with the checkpoints themselves.
+
+`/health.vision` reports the family, the tower, the processor and the
+feature cache (encoder outputs by image digest, bounded in bytes).

mlx_beam_vision-0.1.0a5/mlx_beam_vision/__init__.py

@@ -0,0 +1,8 @@
+"""mlx-beam-vision: the towers that let mlx-beam see."""
+
+try:
+    from mlx_beam_vision._version import __version__
+except ImportError:  # pragma: no cover - source tree without a build
+    __version__ = "0.0.0+unknown"
+
+__all__ = ["__version__"]

mlx_beam_vision-0.1.0a5/mlx_beam_vision/_vendor/VENDORED.md

@@ -0,0 +1,23 @@
+# Vendored in mlx-beam-vision
+
+The vision towers come from [mlx-vlm](https://github.com/Blaizzy/mlx-vlm)
+(MIT; `mlx_vlm/LICENSE`), wheel 0.7.1 on PyPI. Each tower is the model
+family's `vision.py` and `config.py` under `mlx_vlm/models/<family>/`,
+copied unchanged and served from `mlx_vlm/<family>/` here; `base.py` holds
+the two pieces they import (`BaseModelConfig`, `ensure_fused_sdpa`), copied
+from `mlx_vlm/models/base.py`. What mlx-vlm does around a tower - the
+processor, the merge of image features into the text embeddings, the
+language model - is not vendored: the processor is the checkpoint's own
+through transformers, the merge is mlx-beam's prefill (image spans), the
+language model is mlx-beam's.
+
+| Family here | Upstream directory | Serves |
+|---|---|---|
+| `qwen3_vl` | `mlx_vlm/models/qwen3_vl/` (`vision.py`, `config.py`) | Qwen3-VL, Qwen3.5, Qwen3.8 (the same tower; DeepStack features returned alongside) |
+| `pixtral` | `mlx_vlm/models/pixtral/` (`vision.py`, `config.py`) | Mistral 3 (the projector is ours, after upstream's `mistral3.py`) |
+| `gemma4` | `mlx_vlm/models/gemma4/` (`vision.py`, `config.py`) | Gemma 4 E-series (the embedder is ours, after upstream's `gemma4.py`); upstream's `gemma4_unified/` (the encoder-free 12B) is not taken - its image tokens need a bidirectional mask in the text model |
+| `muse_glimmer` | `mlx_vlm/models/muse_glimmer/` (`vision.py`, `config.py`) | Muse Glimmer (adapter and projection ours, after upstream's `muse_glimmer.py`) |
+| `granite4_vision` | `mlx_vlm/models/granite4_vision/` (`vision.py`, `config.py`, `downsampling.py`, `qformer.py`) | Granite Vision 4.1 (the packing and per-layer routing ours, after upstream's `granite4_vision.py`) |
+
+`tools/vendor_diff.py` in the repository root compares these files against
+the wheel (`tools/vendor.toml`, part `mlx-vlm-vision`).

mlx_beam_vision-0.1.0a5/mlx_beam_vision/_vendor/mlx_vlm/LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright © 2025 Prince Canuma
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

mlx_beam_vision-0.1.0a5/mlx_beam_vision/_vendor/mlx_vlm/base.py

@@ -0,0 +1,39 @@
+# Copyright © 2024 Apple Inc. and the mlx-vlm contributors. MIT License.
+#
+# The two pieces the vendored towers import from mlx_vlm/models/base.py,
+# copied verbatim; the rest of that module (processors, features) stays out.
+
+import inspect
+from dataclasses import dataclass
+
+import mlx.core as mx
+
+
+@dataclass
+class BaseModelConfig:
+    @classmethod
+    def from_dict(cls, params):
+        if not params:
+            return cls()
+        return cls(
+            **{
+                k: v
+                for k, v in params.items()
+                if k in inspect.signature(cls).parameters
+            }
+        )
+
+    def to_dict(self):
+        return {k: v for k, v in self.__dict__.items() if v is not None}
+
+
+def ensure_fused_sdpa(q, k, v, scale, mask=None):
+    fused_dims = (64, 80, 128)  # supported by MLX's fused SDPA kernel
+    d = q.shape[-1]
+    target = next((t for t in fused_dims if d <= t), d)
+    if target != d:
+        pad = [(0, 0)] * (q.ndim - 1) + [(0, target - d)]
+        q, k, v = mx.pad(q, pad), mx.pad(k, pad), mx.pad(v, pad)
+    return mx.fast.scaled_dot_product_attention(q, k, v, scale=scale, mask=mask)[
+        ..., :d
+    ]

mlx_beam_vision-0.1.0a5/mlx_beam_vision/_vendor/mlx_vlm/gemma4/config.py

@@ -0,0 +1,146 @@
+from dataclasses import dataclass, field
+from typing import Dict, List, Optional
+
+from ..base import BaseModelConfig
+
+
+@dataclass
+class AudioConfig(BaseModelConfig):
+    hidden_size: int = 1024
+    num_hidden_layers: int = 12
+    num_attention_heads: int = 8
+    hidden_act: str = "silu"
+    subsampling_conv_channels: tuple = (128, 32)
+    conv_kernel_size: int = 5
+    residual_weight: float = 0.5
+    attention_chunk_size: int = 12
+    attention_context_left: int = 13
+    attention_context_right: int = 0
+    attention_logit_cap: float = 50.0
+    attention_invalid_logits_value: float = -1e9
+    use_clipped_linears: bool = True
+    rms_norm_eps: float = 1e-6
+    gradient_clipping: float = 10000000000.0
+    output_proj_dims: Optional[int] = 1536
+
+
+@dataclass
+class VisionConfig(BaseModelConfig):
+    model_type: str = "gemma4_vision"
+    hidden_size: int = 768
+    intermediate_size: int = 3072
+    num_hidden_layers: int = 16
+    num_attention_heads: int = 12
+    num_key_value_heads: int = 12
+    head_dim: int = 64
+    global_head_dim: int = 64
+    hidden_activation: str = "gelu_pytorch_tanh"
+    rms_norm_eps: float = 1e-6
+    max_position_embeddings: int = 131072
+    attention_bias: bool = False
+    attention_dropout: float = 0.0
+    use_bidirectional_attention: str = "vision"
+    layer_types: Optional[List[str]] = None
+    rope_parameters: Optional[Dict] = None
+    default_output_length: int = 280
+    patch_size: int = 16
+    position_embedding_size: int = 10240
+    pooling_kernel_size: int = 3
+    use_clipped_linears: bool = False
+    standardize: bool = False
+
+    def __post_init__(self):
+        if self.layer_types is None:
+            self.layer_types = ["full_attention"] * self.num_hidden_layers
+        if self.rope_parameters is None:
+            self.rope_parameters = {"rope_theta": 100.0, "rope_type": "default"}
+
+
+@dataclass
+class TextConfig(BaseModelConfig):
+    model_type: str = "gemma4_text"
+    hidden_size: int = 1536
+    num_hidden_layers: int = 35
+    intermediate_size: int = 6144
+    num_attention_heads: int = 8
+    head_dim: int = 256
+    global_head_dim: int = 512
+    global_partial_rotary_factor: float = 0.25
+    rms_norm_eps: float = 1e-6
+    vocab_size: int = 262144
+    vocab_size_per_layer_input: int = 262144
+    num_key_value_heads: int = 1
+    num_global_key_value_heads: Optional[int] = None
+    num_kv_shared_layers: int = 20
+    pad_token_id: int = 0
+    hidden_activation: str = "gelu_pytorch_tanh"
+    hidden_size_per_layer_input: int = 256
+    rope_traditional: bool = False
+    partial_rotary_factor: float = 1.0
+    rope_parameters: Optional[Dict] = None
+    sliding_window: int = 512
+    sliding_window_pattern: int = 5
+    _sliding_window_pattern: int = 5
+    max_position_embeddings: int = 131072
+    attention_bias: bool = False
+    attention_dropout: float = 0.0
+    attention_k_eq_v: bool = False
+    use_bidirectional_attention: Optional[str] = None
+    final_logit_softcapping: float = 30.0
+    use_double_wide_mlp: bool = True
+    enable_moe_block: bool = False
+    use_second_mlp_block: bool = False
+    exact_speculative_verify: bool = False
+    num_experts: Optional[int] = None
+    top_k_experts: Optional[int] = None
+    moe_intermediate_size: Optional[int] = None
+    layer_types: Optional[List[str]] = None
+    tie_word_embeddings: bool = True
+
+    def __post_init__(self):
+        if self.rope_parameters is None:
+            self.rope_parameters = {
+                "full_attention": {
+                    "partial_rotary_factor": 1.0,
+                    "rope_theta": 1000000.0,
+                    "rope_type": "proportional",
+                },
+                "sliding_attention": {
+                    "partial_rotary_factor": 1.0,
+                    "rope_theta": 10000.0,
+                    "rope_type": "default",
+                },
+            }
+        if self.layer_types is None:
+            pattern = ["sliding_attention"] * (self.sliding_window_pattern - 1) + [
+                "full_attention"
+            ]
+            self.layer_types = (pattern * (self.num_hidden_layers // len(pattern) + 1))[
+                : self.num_hidden_layers
+            ]
+
+
+@dataclass
+class ModelConfig(BaseModelConfig):
+    text_config: TextConfig = field(default_factory=TextConfig)
+    vision_config: VisionConfig = field(default_factory=VisionConfig)
+    audio_config: Optional[AudioConfig] = None
+    model_type: str = "gemma4"
+    vocab_size: int = 262144
+    ignore_index: int = -100
+    image_token_id: int = 258880
+    audio_token_id: int = 258881
+    video_token_id: Optional[int] = None
+    boi_token_id: int = 255999
+    eoi_token_id: int = 258882
+    boa_token_id: int = 256000
+    eoa_token_id: int = 258883
+    hidden_size: int = 1536
+    pad_token_id: int = 0
+    vision_soft_tokens_per_image: int = 280
+    vision_soft_tokens_per_video_frame: int = 70
+    audio_soft_tokens_per_image: int = 750
+    audio_ms_per_token: int = 40
+    eos_token_id: Optional[List[int]] = None
+    tie_word_embeddings: bool = True
+    initializer_range: float = 0.02