Strip gemma-4-26B-A4B-it-OnlyText to text-only (vision/audio removed)

.gitattributes

@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+tokenizer.json filter=lfs diff=lfs merge=lfs -text

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.

README.md

@@ -0,0 +1,42 @@
+---
+license: apache-2.0
+base_model: google/gemma-4-26B-A4B-it
+library_name: transformers
+tags:
+  - text-generation
+  - conversational
+  - text-only
+pipeline_tag: text-generation
+---
+
+# gemma-4-26B-A4B-it-OnlyText
+
+Text-only causal language model derived from [google/gemma-4-26B-A4B-it](https://huggingface.co/google/gemma-4-26B-A4B-it)
+by removing the vision and audio stacks (their towers/embedders and projector weights)
+and the multimodal special tokens. The text backbone and LM head are preserved.
+
+## Details
+
+- Base model: [google/gemma-4-26B-A4B-it](https://huggingface.co/google/gemma-4-26B-A4B-it)
+- Architecture: `Gemma4ForCausalLM`
+- Parameters: 25.23B
+- Layers: 30 · Hidden size: 2816
+- MTP head: not present
+- Weights: `bfloat16`
+
+## Attribution
+
+This model is a derivative of [google/gemma-4-26B-A4B-it](https://huggingface.co/google/gemma-4-26B-A4B-it) by the
+[Google team](https://huggingface.co/google), released under the
+[apache-2.0 license](https://www.apache.org/licenses/LICENSE-2.0).
+All credit for the underlying weights and capabilities belongs to the original
+authors; this repository only removes modalities, it does not add new training.
+
+## Usage
+
+```python
+from transformers import AutoModelForCausalLM, AutoTokenizer
+
+model = AutoModelForCausalLM.from_pretrained("OnlyTextLLMs/gemma-4-26B-A4B-it-OnlyText")
+tokenizer = AutoTokenizer.from_pretrained("OnlyTextLLMs/gemma-4-26B-A4B-it-OnlyText")
+```

chat_template.jinja

@@ -0,0 +1,390 @@
+{#
+  Template: Google Gemma 4 Canonical Chat Template
+  Author: Google Gemma Engineering Team
+  Published: 2026-07-09
+  Context: Fixed tool-calling loops, turn closures, and thinking content-ordering.
+#}
+{%- macro format_parameters(properties, required, filter_keys=false) -%}
+    {%- set standard_keys = ['description', 'type', 'properties', 'required', 'nullable'] -%}
+    {%- set ns = namespace(found_first=false) -%}
+    {%- for key, value in properties | dictsort -%}
+        {%- set add_comma = false -%}
+        {%- if not filter_keys or key not in standard_keys -%}
+            {%- if ns.found_first %},{% endif -%}
+            {%- set ns.found_first = true -%}
+            {{ key }}:{
+            {%- if value['description'] -%}
+                description:<|"|>{{ value['description'] }}<|"|>
+                {%- set add_comma = true -%}
+            {%- endif -%}
+            {%- if value['type'] | upper == 'STRING' -%}
+                {%- if value['enum'] -%}
+                    {%- if add_comma %},{%- else -%} {%- set add_comma = true -%} {% endif -%}
+                    enum:{{ format_argument(value['enum']) }}
+                {%- endif -%}
+            {%- elif value['type'] | upper == 'ARRAY' -%}
+                {%- if value['items'] is mapping and value['items'] -%}
+                    {%- if add_comma %},{%- else -%} {%- set add_comma = true -%} {% endif -%}
+                    items:{
+                    {%- set ns_items = namespace(found_first=false) -%}
+                    {%- for item_key, item_value in value['items'] | dictsort -%}
+                        {%- if item_value is not none -%}
+                            {%- if ns_items.found_first %},{% endif -%}
+                            {%- set ns_items.found_first = true -%}
+                            {%- if item_key == 'properties' -%}
+                                properties:{
+                                {%- if item_value is mapping -%}
+                                    {{- format_parameters(item_value, value['items']['required'] | default([])) -}}
+                                {%- endif -%}
+                                }
+                            {%- elif item_key == 'required' -%}
+                                required:[
+                                {%- for req_item in item_value -%}
+                                    <|"|>{{- req_item -}}<|"|>
+                                    {%- if not loop.last %},{% endif -%}
+                                {%- endfor -%}
+                                ]
+                            {%- elif item_key == 'type' -%}
+                                {%- if item_value is string -%}
+                                    type:{{ format_argument(item_value | upper) }}
+                                {%- else -%}
+                                    type:{{ format_argument(item_value | map('upper') | list) }}
+                                {%- endif -%}
+                            {%- else -%}
+                                {{ item_key }}:{{ format_argument(item_value) }}
+                            {%- endif -%}
+                        {%- endif -%}
+                    {%- endfor -%}
+                    }
+                {%- endif -%}
+            {%- endif -%}
+            {%- if value['nullable'] %}
+                {%- if add_comma %},{%- else -%} {%- set add_comma = true -%} {% endif -%}
+                nullable:true
+            {%- endif -%}
+            {%- if value['type'] | upper == 'OBJECT' -%}
+                {%- if value['properties'] is defined and value['properties'] is mapping -%}
+                    {%- if add_comma %},{%- else -%} {%- set add_comma = true -%} {% endif -%}
+                    properties:{
+                    {{- format_parameters(value['properties'], value['required'] | default([])) -}}
+                    }
+                {%- elif value is mapping -%}
+                    {%- if add_comma %},{%- else -%} {%- set add_comma = true -%} {% endif -%}
+                    properties:{
+                    {{- format_parameters(value, value['required'] | default([]), filter_keys=true) -}}
+                    }
+                {%- endif -%}
+                {%- if value['required'] -%}
+                    {%- if add_comma %},{%- else -%} {%- set add_comma = true -%} {% endif -%}
+                    required:[
+                    {%- for item in value['required'] | default([]) -%}
+                        <|"|>{{- item -}}<|"|>
+                        {%- if not loop.last %},{% endif -%}
+                    {%- endfor -%}
+                    ]
+                {%- endif -%}
+            {%- endif -%}
+            {%- if add_comma %},{%- else -%} {%- set add_comma = true -%} {% endif -%}
+            type:<|"|>{{ value['type'] | upper }}<|"|>}
+        {%- endif -%}
+    {%- endfor -%}
+{%- endmacro -%}
+{%- macro format_function_declaration(tool_data) -%}
+    declaration:{{- tool_data['function']['name'] -}}{description:<|"|>{{- tool_data['function']['description'] -}}<|"|>
+    {%- set params = tool_data['function']['parameters'] -%}
+    {%- if params -%}
+        ,parameters:{
+        {%- if params['properties'] -%}
+            properties:{ {{- format_parameters(params['properties'], params['required']) -}} },
+        {%- endif -%}
+        {%- if params['required'] -%}
+            required:[
+            {%- for item in params['required'] -%}
+                <|"|>{{- item -}}<|"|>
+                {{- ',' if not loop.last -}}
+            {%- endfor -%}
+            ],
+        {%- endif -%}
+        {%- if params['type'] -%}
+            type:<|"|>{{- params['type'] | upper -}}<|"|>}
+        {%- endif -%}
+    {%- endif -%}
+    {%- if 'response' in tool_data['function'] -%}
+        {%- set response_declaration = tool_data['function']['response'] -%}
+        ,response:{
+        {%- if response_declaration['description'] -%}
+            description:<|"|>{{- response_declaration['description'] -}}<|"|>,
+        {%- endif -%}
+        {%- if response_declaration['type'] | upper == 'OBJECT' -%}
+            type:<|"|>{{- response_declaration['type'] | upper -}}<|"|>}
+        {%- endif -%}
+    {%- endif -%}
+    }
+{%- endmacro -%}
+{%- macro format_argument(argument, escape_keys=True) -%}
+    {%- if argument is none -%}
+        {{- 'null' -}}
+    {%- elif argument is string -%}
+        {{- '<|"|>' + argument + '<|"|>' -}}
+    {%- elif argument is boolean -%}
+        {{- 'true' if argument else 'false' -}}
+    {%- elif argument is mapping -%}
+        {{- '{' -}}
+        {%- set ns = namespace(found_first=false) -%}
+        {%- for key, value in argument | dictsort -%}
+            {%- if ns.found_first %},{% endif -%}
+            {%- set ns.found_first = true -%}
+            {%- if escape_keys -%}
+                {{- '<|"|>' + key + '<|"|>' -}}
+            {%- else -%}
+                {{- key -}}
+            {%- endif -%}
+            :{{- format_argument(value, escape_keys=escape_keys) -}}
+        {%- endfor -%}
+        {{- '}' -}}
+    {%- elif argument is sequence -%}
+        {{- '[' -}}
+        {%- for item in argument -%}
+            {{- format_argument(item, escape_keys=escape_keys) -}}
+            {%- if not loop.last %},{% endif -%}
+        {%- endfor -%}
+        {{- ']' -}}
+    {%- else -%}
+        {{- argument -}}
+    {%- endif -%}
+{%- endmacro -%}
+{%- macro strip_thinking(text) -%}
+    {%- set ns = namespace(result='') -%}
+    {%- for part in text.split('<channel|>') -%}
+        {%- if '<|channel>' in part -%}
+            {%- set ns.result = ns.result + part.split('<|channel>')[0] -%}
+        {%- else -%}
+            {%- set ns.result = ns.result + part -%}
+        {%- endif -%}
+    {%- endfor -%}
+    {{- ns.result | trim -}}
+{%- endmacro -%}
+
+{%- macro format_tool_response_block(tool_name, response) -%}
+    {{- '<|tool_response>' -}}
+    {%- if response is mapping -%}
+        {{- 'response:' + tool_name + '{' -}}
+        {%- for key, value in response | dictsort -%}
+            {{- key -}}:{{- format_argument(value, escape_keys=False) -}}
+            {%- if not loop.last %},{% endif -%}
+        {%- endfor -%}
+        {{- '}' -}}
+    {%- else -%}
+        {{- 'response:' + tool_name + '{value:' + format_argument(response, escape_keys=False) + '}' -}}
+    {%- endif -%}
+    {{- '<tool_response|>' -}}
+{%- endmacro -%}
+
+{#- ===== SETUP ===== -#}
+{%- set ns = namespace(prev_message_type=None, prev_non_tool_role=None) -%}
+{%- set loop_messages = messages -%}
+{%- set enable_thinking = enable_thinking | default(false) -%}
+{%- set preserve_thinking = preserve_thinking | default(false) -%}
+{{- bos_token -}}
+{#- Handle System/Tool Definitions Block -#}
+{%- if enable_thinking or tools or (messages and messages[0]['role'] in ['system', 'developer']) -%}
+    {{- '<|turn>system\n' -}}
+    {#- Inject Thinking token at the very top of the FIRST system turn -#}
+    {%- if enable_thinking -%}
+        {{- '<|think|>\n' -}}
+        {%- set ns.prev_message_type = 'think' -%}
+    {%- endif -%}
+    {%- if messages and messages[0]['role'] in ['system', 'developer'] -%}
+        {%- if messages[0]['content'] is string -%}
+            {{- messages[0]['content'] | trim -}}
+        {%- elif messages[0]['content'] is sequence -%}
+            {%- for item in messages[0]['content'] -%}
+                {{- item['text'] | trim + ' '-}}
+            {%- endfor -%}
+        {%- endif -%}
+        {%- set loop_messages = messages[1:] -%}
+    {%- endif -%}
+    {%- if tools -%}
+        {%- for tool in tools %}
+            {{- '<|tool>' -}}
+            {{- format_function_declaration(tool) | trim -}}
+            {{- '<tool|>' -}}
+        {%- endfor %}
+        {%- set ns.prev_message_type = 'tool' -%}
+    {%- endif -%}
+    {{- '<turn|>\n' -}}
+{%- endif %}
+
+{#- Pre-scan: find last user message index for reasoning guard -#}
+{%- set ns_turn = namespace(last_user_idx=-1) -%}
+{%- for i in range(loop_messages | length) -%}
+    {%- if loop_messages[i]['role'] == 'user' -%}
+        {%- set ns_turn.last_user_idx = i -%}
+    {%- endif -%}
+{%- endfor -%}
+
+{#- Loop through messages -#}
+{%- for message in loop_messages -%}
+    {%- if message['role'] != 'tool' -%}
+    {%- set ns.prev_message_type = None -%}
+    {%- set role = 'model' if message['role'] == 'assistant' else message['role'] -%}
+    {#- Detect continuation using tracked state — O(1) instead of O(n) backward scan -#}
+    {%- set continue_same_model_turn = (role == 'model' and ns.prev_non_tool_role == 'assistant') -%}
+    {%- if not continue_same_model_turn -%}
+        {{- '<|turn>' + role + '\n' }}
+
+    {%- endif -%}
+
+    {#- Render reasoning/reasoning_content as thinking channel -#}
+    {%- set thinking_text = message.get('reasoning') or message.get('reasoning_content') -%}
+    {%- set thinking_gate = (loop.index0 > ns_turn.last_user_idx) or (preserve_thinking and message.get('tool_calls')) -%}
+    {%- if thinking_text and thinking_gate -%}
+        {{- '<|channel>thought\n' + thinking_text + '\n<channel|>' -}}
+    {%- endif -%}
+
+            {%- if message.get('tool_calls') -%}
+                {%- for tool_call in message.get('tool_calls') -%}
+                    {%- set function = tool_call['function'] -%}
+                    {{- '<|tool_call>call:' + function['name'] + '{' -}}
+                    {%- if function['arguments'] is mapping -%}
+                        {%- set ns_args = namespace(found_first=false) -%}
+                        {%- for key, value in function['arguments'] | dictsort -%}
+                            {%- if ns_args.found_first %},{% endif -%}
+                            {%- set ns_args.found_first = true -%}
+                            {{- key -}}:{{- format_argument(value, escape_keys=False) -}}
+                        {%- endfor -%}
+                    {%- elif function['arguments'] is none -%}
+                    {%- else -%}
+                        {{- raise_exception(
+                            "chat_template: tool_calls[].function.arguments must be a "
+                            "JSON object (mapping), not a string. Deserialize arguments "
+                            "before passing to the template."
+                        ) -}}
+                    {%- endif -%}
+                    {{- '}<tool_call|>' -}}
+                {%- endfor -%}
+                {%- set ns.prev_message_type = 'tool_call' -%}
+            {%- endif -%}
+
+            {%- set ns_tr_out = namespace(flag=false) -%}
+            {%- if message.get('tool_responses') -%}
+                {#- Legacy: tool_responses embedded on the assistant message (Google/Gemma native) -#}
+                {%- for tool_response in message.get('tool_responses') -%}
+                    {{- format_tool_response_block(tool_response['name'] | default('unknown', true), tool_response['response']) -}}
+                    {%- set ns_tr_out.flag = true -%}
+                    {%- set ns.prev_message_type = 'tool_response' -%}
+                {%- endfor -%}
+            {%- elif message.get('tool_calls') -%}
+                {#- OpenAI Chat Completions: forward-scan consecutive role:tool messages -#}
+                {%- set ns_tool_scan = namespace(stopped=false) -%}
+                {%- for k in range(loop.index0 + 1, loop_messages | length) -%}
+                    {%- if ns_tool_scan.stopped -%}
+                    {%- elif loop_messages[k]['role'] != 'tool' -%}
+                        {%- set ns_tool_scan.stopped = true -%}
+                    {%- else -%}
+                        {%- set follow = loop_messages[k] -%}
+                        {#- Resolve tool_call_id to function name -#}
+                        {%- set ns_tname = namespace(name=follow.get('name') or 'unknown') -%}
+                        {%- for tc in message.get('tool_calls') -%}
+                            {%- if tc.get('id') == follow.get('tool_call_id') -%}
+                                {%- set ns_tname.name = tc['function']['name'] -%}
+                            {%- endif -%}
+                        {%- endfor -%}
+                        {#- Handle content as string or content-parts array -#}
+                        {%- set tool_body = follow.get('content') -%}
+                        {%- if tool_body is string -%}
+                            {{- format_tool_response_block(ns_tname.name, tool_body) -}}
+                        {%- elif tool_body is sequence and tool_body is not string -%}
+                            {%- set ns_txt = namespace(s='') -%}
+                            {%- for part in tool_body -%}
+                                {%- if part.get('type') == 'text' -%}
+                                    {%- set ns_txt.s = ns_txt.s + (part.get('text') | default('')) -%}
+                                {%- endif -%}
+                            {%- endfor -%}
+                            {{- format_tool_response_block(ns_tname.name, ns_txt.s) -}}
+                            {%- for part in tool_body -%}
+                                {%- if part.get('type') in ['image', 'image_url'] -%}
+                                    {{- '<|image|>' -}}
+                                {%- elif part.get('type') in ['audio', 'input_audio'] -%}
+                                    {{- '<|audio|>' -}}
+                                {%- elif part.get('type') == 'video' -%}
+                                    {{- '<|video|>' -}}
+                                {%- endif -%}
+                            {%- endfor -%}
+                        {%- else -%}
+                            {{- format_tool_response_block(ns_tname.name, tool_body) -}}
+                        {%- endif -%}
+                        {%- set ns_tr_out.flag = true -%}
+                        {%- set ns.prev_message_type = 'tool_response' -%}
+                    {%- endif -%}
+                {%- endfor -%}
+            {%- endif -%}
+
+            {%- set captured_content -%}
+            {%- if message.get('content') is string -%}
+                {%- if role == 'model' -%}
+                    {{- strip_thinking(message['content']) -}}
+                {%- else -%}
+                    {{- message['content'] | trim -}}
+                {%- endif -%}
+            {%- elif message.get('content') is sequence -%}
+                {%- for item in message['content'] -%}
+                    {%- if item.get('type') == 'text' -%}
+                        {%- if role == 'model' -%}
+                            {{- strip_thinking(item['text']) -}}
+                        {%- else -%}
+                            {{- item['text'] | trim -}}
+                        {%- endif -%}
+                    {%- elif item.get('type') in ['image', 'image_url'] -%}
+                        {{- '<|image|>' -}}
+                    {%- elif item.get('type') in ['audio', 'input_audio'] -%}
+                        {{- '<|audio|>' -}}
+                    {%- elif item.get('type') == 'video' -%}
+                        {{- '<|video|>' -}}
+                    {%- endif -%}
+                {%- endfor -%}
+            {%- endif -%}
+            {%- endset -%}
+
+            {{- captured_content -}}
+            {%- set has_content = captured_content | trim | length > 0 -%}
+
+        {#- Forward-scan: find next non-tool message role for continuation detection -#}
+        {%- set next_nt = namespace(role=None, found=false) -%}
+        {%- for j in range(loop.index0 + 1, loop_messages | length) -%}
+            {%- if not next_nt.found -%}
+                {%- if loop_messages[j]['role'] != 'tool' -%}
+                    {%- set next_nt.role = loop_messages[j]['role'] -%}
+                    {%- set next_nt.found = true -%}
+                {%- endif -%}
+            {%- endif -%}
+        {%- endfor -%}
+
+        {%- set continues_into_next = (
+            role == 'model'
+            and next_nt.role == 'assistant'
+            and (not message.get('tool_calls') or ns_tr_out.flag)
+        ) -%}
+
+        {%- if ns.prev_message_type == 'tool_call' and not ns_tr_out.flag -%}
+            {{- '<|tool_response>' -}}
+        {%- elif continues_into_next -%}
+        {%- elif not (ns_tr_out.flag and not has_content and not next_nt.found) -%}
+            {{- '<turn|>\n' -}}
+        {%- endif -%}
+
+    {#- Track previous non-tool role for next iteration (avoids O(n) backward scan) -#}
+    {%- set ns.prev_non_tool_role = message['role'] -%}
+    {%- endif -%}
+{%- endfor -%}
+
+{%- if add_generation_prompt -%}
+    {%- if ns.prev_message_type != 'tool_response' and ns.prev_message_type != 'tool_call' -%}
+        {{- '<|turn>model\n' -}}
+        {%- if not enable_thinking -%}
+            {{- '<|channel>thought\n<channel|>' -}}
+        {%- endif -%}
+    {%- elif ns.prev_message_type == 'tool_response' and enable_thinking -%}
+        {{- '<|channel>thought\n' -}}
+    {%- endif -%}
+{%- endif -%}

config.json

@@ -0,0 +1,116 @@
+{
+  "transformers_version": "5.15.0",
+  "architectures": [
+    "Gemma4ForCausalLM"
+  ],
+  "output_hidden_states": false,
+  "return_dict": true,
+  "dtype": "bfloat16",
+  "chunk_size_feed_forward": 0,
+  "is_encoder_decoder": false,
+  "id2label": {
+    "0": "LABEL_0",
+    "1": "LABEL_1"
+  },
+  "label2id": {
+    "LABEL_0": 0,
+    "LABEL_1": 1
+  },
+  "vocab_size": 262137,
+  "hidden_size": 2816,
+  "intermediate_size": 2112,
+  "num_hidden_layers": 30,
+  "num_attention_heads": 16,
+  "num_key_value_heads": 8,
+  "head_dim": 256,
+  "hidden_activation": "gelu_pytorch_tanh",
+  "max_position_embeddings": 262144,
+  "initializer_range": 0.02,
+  "rms_norm_eps": 1e-06,
+  "use_cache": true,
+  "pad_token_id": 0,
+  "eos_token_id": 1,
+  "bos_token_id": 2,
+  "tie_word_embeddings": true,
+  "rope_parameters": {
+    "full_attention": {
+      "partial_rotary_factor": 0.25,
+      "rope_theta": 1000000.0,
+      "rope_type": "proportional"
+    },
+    "sliding_attention": {
+      "rope_theta": 10000.0,
+      "rope_type": "default"
+    }
+  },
+  "attention_bias": false,
+  "attention_dropout": 0.0,
+  "sliding_window": 1024,
+  "layer_types": [
+    "sliding_attention",
+    "sliding_attention",
+    "sliding_attention",
+    "sliding_attention",
+    "sliding_attention",
+    "full_attention",
+    "sliding_attention",
+    "sliding_attention",
+    "sliding_attention",
+    "sliding_attention",
+    "sliding_attention",
+    "full_attention",
+    "sliding_attention",
+    "sliding_attention",
+    "sliding_attention",
+    "sliding_attention",
+    "sliding_attention",
+    "full_attention",
+    "sliding_attention",
+    "sliding_attention",
+    "sliding_attention",
+    "sliding_attention",
+    "sliding_attention",
+    "full_attention",
+    "sliding_attention",
+    "sliding_attention",
+    "sliding_attention",
+    "sliding_attention",
+    "sliding_attention",
+    "full_attention"
+  ],
+  "final_logit_softcapping": 30.0,
+  "use_bidirectional_attention": "vision",
+  "vocab_size_per_layer_input": 262137,
+  "hidden_size_per_layer_input": 0,
+  "attention_k_eq_v": true,
+  "num_kv_shared_layers": 0,
+  "enable_moe_block": true,
+  "use_double_wide_mlp": false,
+  "num_experts": 128,
+  "top_k_experts": 8,
+  "moe_intermediate_size": 704,
+  "model_type": "gemma4_text",
+  "output_attentions": false,
+  "per_layer_config": {
+    "05": {
+      "head_dim": 512,
+      "num_key_value_heads": 2
+    },
+    "11": {
+      "head_dim": 512,
+      "num_key_value_heads": 2
+    },
+    "17": {
+      "head_dim": 512,
+      "num_key_value_heads": 2
+    },
+    "23": {
+      "head_dim": 512,
+      "num_key_value_heads": 2
+    },
+    "29": {
+      "head_dim": 512,
+      "num_key_value_heads": 2
+    }
+  }
+}

generation_config.json

@@ -0,0 +1,14 @@
+{
+  "bos_token_id": 2,
+  "do_sample": true,
+  "eos_token_id": [
+    1,
+    106,
+    50
+  ],
+  "pad_token_id": 0,
+  "temperature": 1.0,
+  "top_k": 64,
+  "top_p": 0.95,
+  "transformers_version": "5.5.0.dev0"
+}

model-00001.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3f8294e204ae4763c449eda2c9cbff97a4843475f691572ec4e09fd9c45c8e94
+size 4627080762

model-00002.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6ec2f098f3e205a7f2aaf206a398fd0c643ae0a54f7a2861a4e28c0c84ba2918
+size 4876968166

model-00003.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cce8bb40806ca50646f7a388d29213c55b46d3f76827c39d5c239a27f52ff932
+size 4884577422

model-00004.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3d9c81ed4c1629ec45131b99c95b4dc71b4d2771b06dea63e7d9199aab9edc57
+size 4369457286

model-00005.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:018451c84425299ae7a585205bc49dc3f42521e1544316d8a3e1aa99de08b0b4
+size 4884577398

model-00006.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5a4acb012f19cf6e7b8ceb651ed711f23713c54b35b630b16daca9568b00e42a
+size 4377066518

model-00007.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e007af4f8f6052674bc552f6b82b4568a9e9c6c978b557087b93207dbb5b8b70
+size 4876968206

model-00008.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:43436a76f355b90d1e662671850082449d49ce9dbecf9bcdc9c080428b77363e
+size 4377066518

model-00009.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:687e34a22f6953c1f47a381cc12cb2f4601b7efc804c7d490ef66820dafea6d6
+size 4876968158

model-00010.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:18da381ff6aa891a121ad397e80c31834ec59f2dba8408a7e8ea9f1c668c2e23
+size 4876968150

model-00011.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:272d9755ec78b351da36cc1b082df1fd2fc437919baeb3b23871ba89b9cebeb2
+size 3438620612

model.safetensors.index.json

@@ -0,0 +1,664 @@
+{
+  "metadata": {
+    "total_size": 50466244156
+  },
+  "weight_map": {
+    "model.embed_tokens.weight": "model-00001.safetensors",
+    "model.layers.0.experts.down_proj": "model-00001.safetensors",
+    "model.layers.0.experts.gate_up_proj": "model-00001.safetensors",
+    "model.layers.0.input_layernorm.weight": "model-00001.safetensors",
+    "model.layers.0.layer_scalar": "model-00001.safetensors",
+    "model.layers.0.mlp.down_proj.weight": "model-00001.safetensors",
+    "model.layers.0.mlp.gate_proj.weight": "model-00001.safetensors",
+    "model.layers.0.mlp.up_proj.weight": "model-00001.safetensors",
+    "model.layers.0.post_attention_layernorm.weight": "model-00001.safetensors",
+    "model.layers.0.post_feedforward_layernorm.weight": "model-00001.safetensors",
+    "model.layers.0.post_feedforward_layernorm_1.weight": "model-00001.safetensors",
+    "model.layers.0.post_feedforward_layernorm_2.weight": "model-00001.safetensors",
+    "model.layers.0.pre_feedforward_layernorm.weight": "model-00001.safetensors",
+    "model.layers.0.pre_feedforward_layernorm_2.weight": "model-00001.safetensors",
+    "model.layers.0.router.per_expert_scale": "model-00001.safetensors",
+    "model.layers.0.router.proj.weight": "model-00001.safetensors",
+    "model.layers.0.router.scale": "model-00001.safetensors",
+    "model.layers.0.self_attn.k_norm.weight": "model-00001.safetensors",
+    "model.layers.0.self_attn.k_proj.weight": "model-00001.safetensors",
+    "model.layers.0.self_attn.o_proj.weight": "model-00001.safetensors",
+    "model.layers.0.self_attn.q_norm.weight": "model-00001.safetensors",
+    "model.layers.0.self_attn.q_proj.weight": "model-00001.safetensors",
+    "model.layers.0.self_attn.v_proj.weight": "model-00001.safetensors",
+    "model.layers.1.experts.down_proj": "model-00001.safetensors",
+    "model.layers.1.experts.gate_up_proj": "model-00001.safetensors",
+    "model.layers.1.input_layernorm.weight": "model-00002.safetensors",
+    "model.layers.1.layer_scalar": "model-00002.safetensors",
+    "model.layers.1.mlp.down_proj.weight": "model-00002.safetensors",
+    "model.layers.1.mlp.gate_proj.weight": "model-00002.safetensors",
+    "model.layers.1.mlp.up_proj.weight": "model-00002.safetensors",
+    "model.layers.1.post_attention_layernorm.weight": "model-00002.safetensors",
+    "model.layers.1.post_feedforward_layernorm.weight": "model-00002.safetensors",
+    "model.layers.1.post_feedforward_layernorm_1.weight": "model-00002.safetensors",
+    "model.layers.1.post_feedforward_layernorm_2.weight": "model-00002.safetensors",
+    "model.layers.1.pre_feedforward_layernorm.weight": "model-00002.safetensors",
+    "model.layers.1.pre_feedforward_layernorm_2.weight": "model-00002.safetensors",
+    "model.layers.1.router.per_expert_scale": "model-00002.safetensors",
+    "model.layers.1.router.proj.weight": "model-00002.safetensors",
+    "model.layers.1.router.scale": "model-00002.safetensors",
+    "model.layers.1.self_attn.k_norm.weight": "model-00002.safetensors",
+    "model.layers.1.self_attn.k_proj.weight": "model-00002.safetensors",
+    "model.layers.1.self_attn.o_proj.weight": "model-00002.safetensors",
+    "model.layers.1.self_attn.q_norm.weight": "model-00002.safetensors",
+    "model.layers.1.self_attn.q_proj.weight": "model-00002.safetensors",
+    "model.layers.1.self_attn.v_proj.weight": "model-00002.safetensors",
+    "model.layers.10.experts.down_proj": "model-00002.safetensors",
+    "model.layers.10.experts.gate_up_proj": "model-00002.safetensors",
+    "model.layers.10.input_layernorm.weight": "model-00002.safetensors",
+    "model.layers.10.layer_scalar": "model-00002.safetensors",
+    "model.layers.10.mlp.down_proj.weight": "model-00002.safetensors",
+    "model.layers.10.mlp.gate_proj.weight": "model-00002.safetensors",
+    "model.layers.10.mlp.up_proj.weight": "model-00002.safetensors",
+    "model.layers.10.post_attention_layernorm.weight": "model-00002.safetensors",
+    "model.layers.10.post_feedforward_layernorm.weight": "model-00002.safetensors",
+    "model.layers.10.post_feedforward_layernorm_1.weight": "model-00002.safetensors",
+    "model.layers.10.post_feedforward_layernorm_2.weight": "model-00002.safetensors",
+    "model.layers.10.pre_feedforward_layernorm.weight": "model-00002.safetensors",
+    "model.layers.10.pre_feedforward_layernorm_2.weight": "model-00002.safetensors",
+    "model.layers.10.router.per_expert_scale": "model-00002.safetensors",
+    "model.layers.10.router.proj.weight": "model-00002.safetensors",
+    "model.layers.10.router.scale": "model-00002.safetensors",
+    "model.layers.10.self_attn.k_norm.weight": "model-00002.safetensors",
+    "model.layers.10.self_attn.k_proj.weight": "model-00002.safetensors",
+    "model.layers.10.self_attn.o_proj.weight": "model-00002.safetensors",
+    "model.layers.10.self_attn.q_norm.weight": "model-00002.safetensors",
+    "model.layers.10.self_attn.q_proj.weight": "model-00002.safetensors",
+    "model.layers.10.self_attn.v_proj.weight": "model-00002.safetensors",
+    "model.layers.11.experts.down_proj": "model-00002.safetensors",
+    "model.layers.11.experts.gate_up_proj": "model-00002.safetensors",
+    "model.layers.11.layer_scalar": "model-00002.safetensors",
+    "model.layers.11.router.per_expert_scale": "model-00002.safetensors",
+    "model.layers.11.router.scale": "model-00002.safetensors",
+    "model.layers.11.self_attn.k_norm.weight": "model-00002.safetensors",
+    "model.layers.11.self_attn.k_proj.weight": "model-00002.safetensors",
+    "model.layers.11.self_attn.o_proj.weight": "model-00002.safetensors",
+    "model.layers.11.self_attn.q_norm.weight": "model-00002.safetensors",
+    "model.layers.11.self_attn.q_proj.weight": "model-00002.safetensors",
+    "model.layers.12.experts.down_proj": "model-00002.safetensors",
+    "model.layers.12.experts.gate_up_proj": "model-00002.safetensors",
+    "model.layers.12.input_layernorm.weight": "model-00003.safetensors",
+    "model.layers.12.layer_scalar": "model-00003.safetensors",
+    "model.layers.12.mlp.down_proj.weight": "model-00003.safetensors",
+    "model.layers.12.mlp.gate_proj.weight": "model-00003.safetensors",
+    "model.layers.12.mlp.up_proj.weight": "model-00003.safetensors",
+    "model.layers.12.post_attention_layernorm.weight": "model-00003.safetensors",
+    "model.layers.12.post_feedforward_layernorm.weight": "model-00003.safetensors",
+    "model.layers.12.post_feedforward_layernorm_1.weight": "model-00003.safetensors",
+    "model.layers.12.post_feedforward_layernorm_2.weight": "model-00003.safetensors",
+    "model.layers.12.pre_feedforward_layernorm.weight": "model-00003.safetensors",
+    "model.layers.12.pre_feedforward_layernorm_2.weight": "model-00003.safetensors",
+    "model.layers.12.router.per_expert_scale": "model-00003.safetensors",
+    "model.layers.12.router.proj.weight": "model-00003.safetensors",
+    "model.layers.12.router.scale": "model-00003.safetensors",
+    "model.layers.12.self_attn.k_norm.weight": "model-00003.safetensors",
+    "model.layers.12.self_attn.k_proj.weight": "model-00003.safetensors",
+    "model.layers.12.self_attn.o_proj.weight": "model-00003.safetensors",
+    "model.layers.12.self_attn.q_norm.weight": "model-00003.safetensors",
+    "model.layers.12.self_attn.q_proj.weight": "model-00003.safetensors",
+    "model.layers.12.self_attn.v_proj.weight": "model-00003.safetensors",
+    "model.layers.13.experts.down_proj": "model-00003.safetensors",
+    "model.layers.13.experts.gate_up_proj": "model-00003.safetensors",
+    "model.layers.13.input_layernorm.weight": "model-00003.safetensors",
+    "model.layers.13.layer_scalar": "model-00003.safetensors",
+    "model.layers.13.mlp.down_proj.weight": "model-00003.safetensors",
+    "model.layers.13.mlp.gate_proj.weight": "model-00003.safetensors",
+    "model.layers.13.mlp.up_proj.weight": "model-00003.safetensors",
+    "model.layers.13.post_attention_layernorm.weight": "model-00003.safetensors",
+    "model.layers.13.post_feedforward_layernorm.weight": "model-00003.safetensors",
+    "model.layers.13.post_feedforward_layernorm_1.weight": "model-00003.safetensors",
+    "model.layers.13.post_feedforward_layernorm_2.weight": "model-00003.safetensors",
+    "model.layers.13.pre_feedforward_layernorm.weight": "model-00003.safetensors",
+    "model.layers.13.pre_feedforward_layernorm_2.weight": "model-00003.safetensors",
+    "model.layers.13.router.per_expert_scale": "model-00003.safetensors",
+    "model.layers.13.router.proj.weight": "model-00003.safetensors",
+    "model.layers.13.router.scale": "model-00003.safetensors",
+    "model.layers.13.self_attn.k_norm.weight": "model-00003.safetensors",
+    "model.layers.13.self_attn.k_proj.weight": "model-00003.safetensors",
+    "model.layers.13.self_attn.o_proj.weight": "model-00003.safetensors",
+    "model.layers.13.self_attn.q_norm.weight": "model-00003.safetensors",
+    "model.layers.13.self_attn.q_proj.weight": "model-00003.safetensors",
+    "model.layers.13.self_attn.v_proj.weight": "model-00003.safetensors",
+    "model.layers.14.experts.down_proj": "model-00003.safetensors",
+    "model.layers.14.experts.gate_up_proj": "model-00003.safetensors",
+    "model.layers.14.input_layernorm.weight": "model-00003.safetensors",
+    "model.layers.14.layer_scalar": "model-00003.safetensors",
+    "model.layers.14.mlp.down_proj.weight": "model-00003.safetensors",
+    "model.layers.14.mlp.gate_proj.weight": "model-00003.safetensors",
+    "model.layers.14.mlp.up_proj.weight": "model-00003.safetensors",
+    "model.layers.14.post_attention_layernorm.weight": "model-00003.safetensors",
+    "model.layers.14.post_feedforward_layernorm.weight": "model-00003.safetensors",
+    "model.layers.14.post_feedforward_layernorm_1.weight": "model-00003.safetensors",
+    "model.layers.14.post_feedforward_layernorm_2.weight": "model-00003.safetensors",
+    "model.layers.14.pre_feedforward_layernorm.weight": "model-00003.safetensors",
+    "model.layers.14.pre_feedforward_layernorm_2.weight": "model-00003.safetensors",
+    "model.layers.14.router.per_expert_scale": "model-00003.safetensors",
+    "model.layers.14.router.proj.weight": "model-00003.safetensors",
+    "model.layers.14.router.scale": "model-00003.safetensors",
+    "model.layers.14.self_attn.k_norm.weight": "model-00003.safetensors",
+    "model.layers.14.self_attn.k_proj.weight": "model-00003.safetensors",
+    "model.layers.14.self_attn.o_proj.weight": "model-00003.safetensors",
+    "model.layers.14.self_attn.q_norm.weight": "model-00003.safetensors",
+    "model.layers.14.self_attn.q_proj.weight": "model-00003.safetensors",
+    "model.layers.14.self_attn.v_proj.weight": "model-00003.safetensors",
+    "model.layers.15.experts.down_proj": "model-00003.safetensors",
+    "model.layers.15.experts.gate_up_proj": "model-00003.safetensors",
+    "model.layers.15.input_layernorm.weight": "model-00004.safetensors",
+    "model.layers.15.layer_scalar": "model-00004.safetensors",
+    "model.layers.15.mlp.down_proj.weight": "model-00004.safetensors",
+    "model.layers.15.mlp.gate_proj.weight": "model-00004.safetensors",
+    "model.layers.15.mlp.up_proj.weight": "model-00004.safetensors",
+    "model.layers.15.post_attention_layernorm.weight": "model-00004.safetensors",
+    "model.layers.15.post_feedforward_layernorm.weight": "model-00004.safetensors",
+    "model.layers.15.post_feedforward_layernorm_1.weight": "model-00004.safetensors",
+    "model.layers.15.post_feedforward_layernorm_2.weight": "model-00004.safetensors",
+    "model.layers.15.pre_feedforward_layernorm.weight": "model-00004.safetensors",
+    "model.layers.15.pre_feedforward_layernorm_2.weight": "model-00004.safetensors",
+    "model.layers.15.router.per_expert_scale": "model-00004.safetensors",
+    "model.layers.15.router.proj.weight": "model-00004.safetensors",
+    "model.layers.15.router.scale": "model-00004.safetensors",
+    "model.layers.15.self_attn.k_norm.weight": "model-00004.safetensors",
+    "model.layers.15.self_attn.k_proj.weight": "model-00004.safetensors",
+    "model.layers.15.self_attn.o_proj.weight": "model-00004.safetensors",
+    "model.layers.15.self_attn.q_norm.weight": "model-00004.safetensors",
+    "model.layers.15.self_attn.q_proj.weight": "model-00004.safetensors",
+    "model.layers.15.self_attn.v_proj.weight": "model-00004.safetensors",
+    "model.layers.16.experts.down_proj": "model-00004.safetensors",
+    "model.layers.16.experts.gate_up_proj": "model-00004.safetensors",
+    "model.layers.16.input_layernorm.weight": "model-00004.safetensors",
+    "model.layers.16.layer_scalar": "model-00004.safetensors",
+    "model.layers.16.mlp.down_proj.weight": "model-00004.safetensors",
+    "model.layers.16.mlp.gate_proj.weight": "model-00004.safetensors",
+    "model.layers.16.mlp.up_proj.weight": "model-00004.safetensors",
+    "model.layers.16.post_attention_layernorm.weight": "model-00004.safetensors",
+    "model.layers.16.post_feedforward_layernorm.weight": "model-00004.safetensors",
+    "model.layers.16.post_feedforward_layernorm_1.weight": "model-00004.safetensors",
+    "model.layers.16.post_feedforward_layernorm_2.weight": "model-00004.safetensors",
+    "model.layers.16.pre_feedforward_layernorm.weight": "model-00004.safetensors",
+    "model.layers.16.pre_feedforward_layernorm_2.weight": "model-00004.safetensors",
+    "model.layers.16.router.per_expert_scale": "model-00004.safetensors",
+    "model.layers.16.router.proj.weight": "model-00004.safetensors",
+    "model.layers.16.router.scale": "model-00004.safetensors",
+    "model.layers.16.self_attn.k_norm.weight": "model-00004.safetensors",
+    "model.layers.16.self_attn.k_proj.weight": "model-00004.safetensors",
+    "model.layers.16.self_attn.o_proj.weight": "model-00004.safetensors",
+    "model.layers.16.self_attn.q_norm.weight": "model-00004.safetensors",
+    "model.layers.16.self_attn.q_proj.weight": "model-00004.safetensors",
+    "model.layers.16.self_attn.v_proj.weight": "model-00004.safetensors",
+    "model.layers.17.experts.gate_up_proj": "model-00004.safetensors",
+    "model.layers.17.layer_scalar": "model-00004.safetensors",
+    "model.layers.17.router.per_expert_scale": "model-00004.safetensors",
+    "model.layers.17.router.scale": "model-00004.safetensors",
+    "model.layers.17.self_attn.k_norm.weight": "model-00004.safetensors",
+    "model.layers.17.self_attn.k_proj.weight": "model-00004.safetensors",
+    "model.layers.17.self_attn.o_proj.weight": "model-00004.safetensors",
+    "model.layers.17.self_attn.q_norm.weight": "model-00004.safetensors",
+    "model.layers.17.self_attn.q_proj.weight": "model-00004.safetensors",
+    "model.layers.18.experts.down_proj": "model-00004.safetensors",
+    "model.layers.18.experts.gate_up_proj": "model-00004.safetensors",
+    "model.layers.18.input_layernorm.weight": "model-00005.safetensors",
+    "model.layers.18.layer_scalar": "model-00005.safetensors",
+    "model.layers.18.mlp.down_proj.weight": "model-00005.safetensors",
+    "model.layers.18.mlp.gate_proj.weight": "model-00005.safetensors",
+    "model.layers.18.mlp.up_proj.weight": "model-00005.safetensors",
+    "model.layers.18.post_attention_layernorm.weight": "model-00005.safetensors",
+    "model.layers.18.post_feedforward_layernorm.weight": "model-00005.safetensors",
+    "model.layers.18.post_feedforward_layernorm_1.weight": "model-00005.safetensors",
+    "model.layers.18.post_feedforward_layernorm_2.weight": "model-00005.safetensors",
+    "model.layers.18.pre_feedforward_layernorm.weight": "model-00005.safetensors",
+    "model.layers.18.pre_feedforward_layernorm_2.weight": "model-00005.safetensors",
+    "model.layers.18.router.per_expert_scale": "model-00005.safetensors",
+    "model.layers.18.router.proj.weight": "model-00005.safetensors",
+    "model.layers.18.router.scale": "model-00005.safetensors",
+    "model.layers.18.self_attn.k_norm.weight": "model-00005.safetensors",
+    "model.layers.18.self_attn.k_proj.weight": "model-00005.safetensors",
+    "model.layers.18.self_attn.o_proj.weight": "model-00005.safetensors",
+    "model.layers.18.self_attn.q_norm.weight": "model-00005.safetensors",
+    "model.layers.18.self_attn.q_proj.weight": "model-00005.safetensors",
+    "model.layers.18.self_attn.v_proj.weight": "model-00005.safetensors",
+    "model.layers.19.experts.down_proj": "model-00005.safetensors",
+    "model.layers.19.experts.gate_up_proj": "model-00005.safetensors",
+    "model.layers.19.input_layernorm.weight": "model-00005.safetensors",
+    "model.layers.19.layer_scalar": "model-00005.safetensors",
+    "model.layers.19.mlp.down_proj.weight": "model-00005.safetensors",
+    "model.layers.19.mlp.gate_proj.weight": "model-00005.safetensors",
+    "model.layers.19.mlp.up_proj.weight": "model-00005.safetensors",
+    "model.layers.19.post_attention_layernorm.weight": "model-00005.safetensors",
+    "model.layers.19.post_feedforward_layernorm.weight": "model-00005.safetensors",
+    "model.layers.19.post_feedforward_layernorm_1.weight": "model-00005.safetensors",
+    "model.layers.19.post_feedforward_layernorm_2.weight": "model-00005.safetensors",
+    "model.layers.19.pre_feedforward_layernorm.weight": "model-00005.safetensors",
+    "model.layers.19.pre_feedforward_layernorm_2.weight": "model-00005.safetensors",
+    "model.layers.19.router.per_expert_scale": "model-00005.safetensors",
+    "model.layers.19.router.proj.weight": "model-00005.safetensors",
+    "model.layers.19.router.scale": "model-00005.safetensors",
+    "model.layers.19.self_attn.k_norm.weight": "model-00005.safetensors",
+    "model.layers.19.self_attn.k_proj.weight": "model-00005.safetensors",
+    "model.layers.19.self_attn.o_proj.weight": "model-00005.safetensors",
+    "model.layers.19.self_attn.q_norm.weight": "model-00005.safetensors",
+    "model.layers.19.self_attn.q_proj.weight": "model-00005.safetensors",
+    "model.layers.19.self_attn.v_proj.weight": "model-00005.safetensors",
+    "model.layers.2.experts.down_proj": "model-00005.safetensors",
+    "model.layers.2.experts.gate_up_proj": "model-00005.safetensors",
+    "model.layers.2.input_layernorm.weight": "model-00005.safetensors",
+    "model.layers.2.layer_scalar": "model-00005.safetensors",
+    "model.layers.2.mlp.down_proj.weight": "model-00005.safetensors",
+    "model.layers.2.mlp.gate_proj.weight": "model-00005.safetensors",
+    "model.layers.2.mlp.up_proj.weight": "model-00005.safetensors",
+    "model.layers.2.post_attention_layernorm.weight": "model-00005.safetensors",
+    "model.layers.2.post_feedforward_layernorm.weight": "model-00005.safetensors",
+    "model.layers.2.post_feedforward_layernorm_1.weight": "model-00005.safetensors",
+    "model.layers.2.post_feedforward_layernorm_2.weight": "model-00005.safetensors",
+    "model.layers.2.pre_feedforward_layernorm.weight": "model-00005.safetensors",
+    "model.layers.2.pre_feedforward_layernorm_2.weight": "model-00005.safetensors",
+    "model.layers.2.router.per_expert_scale": "model-00005.safetensors",
+    "model.layers.2.router.proj.weight": "model-00005.safetensors",
+    "model.layers.2.router.scale": "model-00005.safetensors",
+    "model.layers.2.self_attn.k_norm.weight": "model-00005.safetensors",
+    "model.layers.2.self_attn.k_proj.weight": "model-00005.safetensors",
+    "model.layers.2.self_attn.o_proj.weight": "model-00005.safetensors",
+    "model.layers.2.self_attn.q_norm.weight": "model-00005.safetensors",
+    "model.layers.2.self_attn.q_proj.weight": "model-00005.safetensors",
+    "model.layers.2.self_attn.v_proj.weight": "model-00005.safetensors",
+    "model.layers.20.experts.down_proj": "model-00005.safetensors",
+    "model.layers.20.experts.gate_up_proj": "model-00005.safetensors",
+    "model.layers.20.input_layernorm.weight": "model-00006.safetensors",
+    "model.layers.20.layer_scalar": "model-00006.safetensors",
+    "model.layers.20.mlp.down_proj.weight": "model-00006.safetensors",
+    "model.layers.20.mlp.gate_proj.weight": "model-00006.safetensors",
+    "model.layers.20.mlp.up_proj.weight": "model-00006.safetensors",
+    "model.layers.20.post_attention_layernorm.weight": "model-00006.safetensors",
+    "model.layers.20.post_feedforward_layernorm.weight": "model-00006.safetensors",
+    "model.layers.20.post_feedforward_layernorm_1.weight": "model-00006.safetensors",
+    "model.layers.20.post_feedforward_layernorm_2.weight": "model-00006.safetensors",
+    "model.layers.20.pre_feedforward_layernorm.weight": "model-00006.safetensors",
+    "model.layers.20.pre_feedforward_layernorm_2.weight": "model-00006.safetensors",
+    "model.layers.20.router.per_expert_scale": "model-00006.safetensors",
+    "model.layers.20.router.proj.weight": "model-00006.safetensors",
+    "model.layers.20.router.scale": "model-00006.safetensors",
+    "model.layers.20.self_attn.k_norm.weight": "model-00006.safetensors",
+    "model.layers.20.self_attn.k_proj.weight": "model-00006.safetensors",
+    "model.layers.20.self_attn.o_proj.weight": "model-00006.safetensors",
+    "model.layers.20.self_attn.q_norm.weight": "model-00006.safetensors",
+    "model.layers.20.self_attn.q_proj.weight": "model-00006.safetensors",
+    "model.layers.20.self_attn.v_proj.weight": "model-00006.safetensors",
+    "model.layers.21.experts.down_proj": "model-00006.safetensors",
+    "model.layers.21.experts.gate_up_proj": "model-00006.safetensors",
+    "model.layers.21.input_layernorm.weight": "model-00006.safetensors",
+    "model.layers.21.layer_scalar": "model-00006.safetensors",
+    "model.layers.21.mlp.down_proj.weight": "model-00006.safetensors",
+    "model.layers.21.mlp.gate_proj.weight": "model-00006.safetensors",
+    "model.layers.21.mlp.up_proj.weight": "model-00006.safetensors",
+    "model.layers.21.post_attention_layernorm.weight": "model-00006.safetensors",
+    "model.layers.21.post_feedforward_layernorm.weight": "model-00006.safetensors",
+    "model.layers.21.post_feedforward_layernorm_1.weight": "model-00006.safetensors",
+    "model.layers.21.post_feedforward_layernorm_2.weight": "model-00006.safetensors",
+    "model.layers.21.pre_feedforward_layernorm.weight": "model-00006.safetensors",
+    "model.layers.21.pre_feedforward_layernorm_2.weight": "model-00006.safetensors",
+    "model.layers.21.router.per_expert_scale": "model-00006.safetensors",
+    "model.layers.21.router.proj.weight": "model-00006.safetensors",
+    "model.layers.21.router.scale": "model-00006.safetensors",
+    "model.layers.21.self_attn.k_norm.weight": "model-00006.safetensors",
+    "model.layers.21.self_attn.k_proj.weight": "model-00006.safetensors",
+    "model.layers.21.self_attn.o_proj.weight": "model-00006.safetensors",
+    "model.layers.21.self_attn.q_norm.weight": "model-00006.safetensors",
+    "model.layers.21.self_attn.q_proj.weight": "model-00006.safetensors",
+    "model.layers.21.self_attn.v_proj.weight": "model-00006.safetensors",
+    "model.layers.22.experts.down_proj": "model-00006.safetensors",
+    "model.layers.22.experts.gate_up_proj": "model-00006.safetensors",
+    "model.layers.22.input_layernorm.weight": "model-00006.safetensors",
+    "model.layers.22.layer_scalar": "model-00006.safetensors",
+    "model.layers.22.mlp.down_proj.weight": "model-00006.safetensors",
+    "model.layers.22.mlp.gate_proj.weight": "model-00006.safetensors",
+    "model.layers.22.mlp.up_proj.weight": "model-00006.safetensors",
+    "model.layers.22.post_attention_layernorm.weight": "model-00006.safetensors",
+    "model.layers.22.post_feedforward_layernorm.weight": "model-00006.safetensors",
+    "model.layers.22.post_feedforward_layernorm_1.weight": "model-00006.safetensors",
+    "model.layers.22.post_feedforward_layernorm_2.weight": "model-00006.safetensors",
+    "model.layers.22.pre_feedforward_layernorm.weight": "model-00006.safetensors",
+    "model.layers.22.pre_feedforward_layernorm_2.weight": "model-00006.safetensors",
+    "model.layers.22.router.per_expert_scale": "model-00006.safetensors",
+    "model.layers.22.router.proj.weight": "model-00006.safetensors",
+    "model.layers.22.router.scale": "model-00006.safetensors",
+    "model.layers.22.self_attn.k_norm.weight": "model-00006.safetensors",
+    "model.layers.22.self_attn.k_proj.weight": "model-00006.safetensors",
+    "model.layers.22.self_attn.o_proj.weight": "model-00006.safetensors",
+    "model.layers.22.self_attn.q_norm.weight": "model-00006.safetensors",
+    "model.layers.22.self_attn.q_proj.weight": "model-00006.safetensors",
+    "model.layers.22.self_attn.v_proj.weight": "model-00006.safetensors",
+    "model.layers.23.experts.gate_up_proj": "model-00006.safetensors",
+    "model.layers.23.layer_scalar": "model-00007.safetensors",
+    "model.layers.23.router.per_expert_scale": "model-00007.safetensors",
+    "model.layers.23.router.scale": "model-00007.safetensors",
+    "model.layers.23.self_attn.k_norm.weight": "model-00007.safetensors",
+    "model.layers.23.self_attn.k_proj.weight": "model-00007.safetensors",
+    "model.layers.23.self_attn.o_proj.weight": "model-00007.safetensors",
+    "model.layers.23.self_attn.q_norm.weight": "model-00007.safetensors",
+    "model.layers.23.self_attn.q_proj.weight": "model-00007.safetensors",
+    "model.layers.24.experts.down_proj": "model-00007.safetensors",
+    "model.layers.24.experts.gate_up_proj": "model-00007.safetensors",
+    "model.layers.24.input_layernorm.weight": "model-00007.safetensors",
+    "model.layers.24.layer_scalar": "model-00007.safetensors",
+    "model.layers.24.mlp.down_proj.weight": "model-00007.safetensors",
+    "model.layers.24.mlp.gate_proj.weight": "model-00007.safetensors",
+    "model.layers.24.mlp.up_proj.weight": "model-00007.safetensors",
+    "model.layers.24.post_attention_layernorm.weight": "model-00007.safetensors",
+    "model.layers.24.post_feedforward_layernorm.weight": "model-00007.safetensors",
+    "model.layers.24.post_feedforward_layernorm_1.weight": "model-00007.safetensors",
+    "model.layers.24.post_feedforward_layernorm_2.weight": "model-00007.safetensors",
+    "model.layers.24.pre_feedforward_layernorm.weight": "model-00007.safetensors",
+    "model.layers.24.pre_feedforward_layernorm_2.weight": "model-00007.safetensors",
+    "model.layers.24.router.per_expert_scale": "model-00007.safetensors",
+    "model.layers.24.router.proj.weight": "model-00007.safetensors",
+    "model.layers.24.router.scale": "model-00007.safetensors",
+    "model.layers.24.self_attn.k_norm.weight": "model-00007.safetensors",
+    "model.layers.24.self_attn.k_proj.weight": "model-00007.safetensors",
+    "model.layers.24.self_attn.o_proj.weight": "model-00007.safetensors",
+    "model.layers.24.self_attn.q_norm.weight": "model-00007.safetensors",
+    "model.layers.24.self_attn.q_proj.weight": "model-00007.safetensors",
+    "model.layers.24.self_attn.v_proj.weight": "model-00007.safetensors",
+    "model.layers.25.experts.down_proj": "model-00007.safetensors",
+    "model.layers.25.experts.gate_up_proj": "model-00007.safetensors",
+    "model.layers.25.input_layernorm.weight": "model-00007.safetensors",
+    "model.layers.25.layer_scalar": "model-00007.safetensors",
+    "model.layers.25.mlp.down_proj.weight": "model-00007.safetensors",
+    "model.layers.25.mlp.gate_proj.weight": "model-00007.safetensors",
+    "model.layers.25.mlp.up_proj.weight": "model-00007.safetensors",
+    "model.layers.25.post_attention_layernorm.weight": "model-00007.safetensors",
+    "model.layers.25.post_feedforward_layernorm.weight": "model-00007.safetensors",
+    "model.layers.25.post_feedforward_layernorm_1.weight": "model-00007.safetensors",
+    "model.layers.25.post_feedforward_layernorm_2.weight": "model-00007.safetensors",
+    "model.layers.25.pre_feedforward_layernorm.weight": "model-00007.safetensors",
+    "model.layers.25.pre_feedforward_layernorm_2.weight": "model-00007.safetensors",
+    "model.layers.25.router.per_expert_scale": "model-00007.safetensors",
+    "model.layers.25.router.proj.weight": "model-00007.safetensors",
+    "model.layers.25.router.scale": "model-00007.safetensors",
+    "model.layers.25.self_attn.k_norm.weight": "model-00007.safetensors",
+    "model.layers.25.self_attn.k_proj.weight": "model-00007.safetensors",
+    "model.layers.25.self_attn.o_proj.weight": "model-00007.safetensors",
+    "model.layers.25.self_attn.q_norm.weight": "model-00007.safetensors",
+    "model.layers.25.self_attn.q_proj.weight": "model-00007.safetensors",
+    "model.layers.25.self_attn.v_proj.weight": "model-00007.safetensors",
+    "model.layers.26.experts.down_proj": "model-00007.safetensors",
+    "model.layers.26.experts.gate_up_proj": "model-00007.safetensors",
+    "model.layers.26.input_layernorm.weight": "model-00008.safetensors",
+    "model.layers.26.layer_scalar": "model-00008.safetensors",
+    "model.layers.26.mlp.down_proj.weight": "model-00008.safetensors",
+    "model.layers.26.mlp.gate_proj.weight": "model-00008.safetensors",
+    "model.layers.26.mlp.up_proj.weight": "model-00008.safetensors",
+    "model.layers.26.post_attention_layernorm.weight": "model-00008.safetensors",
+    "model.layers.26.post_feedforward_layernorm.weight": "model-00008.safetensors",
+    "model.layers.26.post_feedforward_layernorm_1.weight": "model-00008.safetensors",
+    "model.layers.26.post_feedforward_layernorm_2.weight": "model-00008.safetensors",
+    "model.layers.26.pre_feedforward_layernorm.weight": "model-00008.safetensors",
+    "model.layers.26.pre_feedforward_layernorm_2.weight": "model-00008.safetensors",
+    "model.layers.26.router.per_expert_scale": "model-00008.safetensors",
+    "model.layers.26.router.proj.weight": "model-00008.safetensors",
+    "model.layers.26.router.scale": "model-00008.safetensors",
+    "model.layers.26.self_attn.k_norm.weight": "model-00008.safetensors",
+    "model.layers.26.self_attn.k_proj.weight": "model-00008.safetensors",
+    "model.layers.26.self_attn.o_proj.weight": "model-00008.safetensors",
+    "model.layers.26.self_attn.q_norm.weight": "model-00008.safetensors",
+    "model.layers.26.self_attn.q_proj.weight": "model-00008.safetensors",
+    "model.layers.26.self_attn.v_proj.weight": "model-00008.safetensors",
+    "model.layers.27.experts.down_proj": "model-00008.safetensors",
+    "model.layers.27.experts.gate_up_proj": "model-00008.safetensors",
+    "model.layers.27.input_layernorm.weight": "model-00008.safetensors",
+    "model.layers.27.layer_scalar": "model-00008.safetensors",
+    "model.layers.27.mlp.down_proj.weight": "model-00008.safetensors",
+    "model.layers.27.mlp.gate_proj.weight": "model-00008.safetensors",
+    "model.layers.27.mlp.up_proj.weight": "model-00008.safetensors",
+    "model.layers.27.post_attention_layernorm.weight": "model-00008.safetensors",
+    "model.layers.27.post_feedforward_layernorm.weight": "model-00008.safetensors",
+    "model.layers.27.post_feedforward_layernorm_1.weight": "model-00008.safetensors",
+    "model.layers.27.post_feedforward_layernorm_2.weight": "model-00008.safetensors",
+    "model.layers.27.pre_feedforward_layernorm.weight": "model-00008.safetensors",
+    "model.layers.27.pre_feedforward_layernorm_2.weight": "model-00008.safetensors",
+    "model.layers.27.router.per_expert_scale": "model-00008.safetensors",
+    "model.layers.27.router.proj.weight": "model-00008.safetensors",
+    "model.layers.27.router.scale": "model-00008.safetensors",
+    "model.layers.27.self_attn.k_norm.weight": "model-00008.safetensors",
+    "model.layers.27.self_attn.k_proj.weight": "model-00008.safetensors",
+    "model.layers.27.self_attn.o_proj.weight": "model-00008.safetensors",
+    "model.layers.27.self_attn.q_norm.weight": "model-00008.safetensors",
+    "model.layers.27.self_attn.q_proj.weight": "model-00008.safetensors",
+    "model.layers.27.self_attn.v_proj.weight": "model-00008.safetensors",
+    "model.layers.28.experts.down_proj": "model-00008.safetensors",
+    "model.layers.28.experts.gate_up_proj": "model-00008.safetensors",
+    "model.layers.28.input_layernorm.weight": "model-00008.safetensors",
+    "model.layers.28.layer_scalar": "model-00008.safetensors",
+    "model.layers.28.mlp.down_proj.weight": "model-00008.safetensors",
+    "model.layers.28.mlp.gate_proj.weight": "model-00008.safetensors",
+    "model.layers.28.mlp.up_proj.weight": "model-00008.safetensors",
+    "model.layers.28.post_attention_layernorm.weight": "model-00008.safetensors",
+    "model.layers.28.post_feedforward_layernorm.weight": "model-00008.safetensors",
+    "model.layers.28.post_feedforward_layernorm_1.weight": "model-00008.safetensors",
+    "model.layers.28.post_feedforward_layernorm_2.weight": "model-00008.safetensors",
+    "model.layers.28.pre_feedforward_layernorm.weight": "model-00008.safetensors",
+    "model.layers.28.pre_feedforward_layernorm_2.weight": "model-00008.safetensors",
+    "model.layers.28.router.per_expert_scale": "model-00008.safetensors",
+    "model.layers.28.router.proj.weight": "model-00008.safetensors",
+    "model.layers.28.router.scale": "model-00008.safetensors",
+    "model.layers.28.self_attn.k_norm.weight": "model-00008.safetensors",
+    "model.layers.28.self_attn.k_proj.weight": "model-00008.safetensors",
+    "model.layers.28.self_attn.o_proj.weight": "model-00008.safetensors",
+    "model.layers.28.self_attn.q_norm.weight": "model-00008.safetensors",
+    "model.layers.28.self_attn.q_proj.weight": "model-00008.safetensors",
+    "model.layers.28.self_attn.v_proj.weight": "model-00008.safetensors",
+    "model.layers.29.experts.gate_up_proj": "model-00008.safetensors",
+    "model.layers.29.layer_scalar": "model-00009.safetensors",
+    "model.layers.29.router.per_expert_scale": "model-00009.safetensors",
+    "model.layers.29.router.scale": "model-00009.safetensors",
+    "model.layers.29.self_attn.k_norm.weight": "model-00009.safetensors",
+    "model.layers.29.self_attn.k_proj.weight": "model-00009.safetensors",
+    "model.layers.29.self_attn.o_proj.weight": "model-00009.safetensors",
+    "model.layers.29.self_attn.q_norm.weight": "model-00009.safetensors",
+    "model.layers.29.self_attn.q_proj.weight": "model-00009.safetensors",
+    "model.layers.3.experts.down_proj": "model-00009.safetensors",
+    "model.layers.3.experts.gate_up_proj": "model-00009.safetensors",
+    "model.layers.3.input_layernorm.weight": "model-00009.safetensors",
+    "model.layers.3.layer_scalar": "model-00009.safetensors",
+    "model.layers.3.mlp.down_proj.weight": "model-00009.safetensors",
+    "model.layers.3.mlp.gate_proj.weight": "model-00009.safetensors",
+    "model.layers.3.mlp.up_proj.weight": "model-00009.safetensors",
+    "model.layers.3.post_attention_layernorm.weight": "model-00009.safetensors",
+    "model.layers.3.post_feedforward_layernorm.weight": "model-00009.safetensors",
+    "model.layers.3.post_feedforward_layernorm_1.weight": "model-00009.safetensors",
+    "model.layers.3.post_feedforward_layernorm_2.weight": "model-00009.safetensors",
+    "model.layers.3.pre_feedforward_layernorm.weight": "model-00009.safetensors",
+    "model.layers.3.pre_feedforward_layernorm_2.weight": "model-00009.safetensors",
+    "model.layers.3.router.per_expert_scale": "model-00009.safetensors",
+    "model.layers.3.router.proj.weight": "model-00009.safetensors",
+    "model.layers.3.router.scale": "model-00009.safetensors",
+    "model.layers.3.self_attn.k_norm.weight": "model-00009.safetensors",
+    "model.layers.3.self_attn.k_proj.weight": "model-00009.safetensors",
+    "model.layers.3.self_attn.o_proj.weight": "model-00009.safetensors",
+    "model.layers.3.self_attn.q_norm.weight": "model-00009.safetensors",
+    "model.layers.3.self_attn.q_proj.weight": "model-00009.safetensors",
+    "model.layers.3.self_attn.v_proj.weight": "model-00009.safetensors",
+    "model.layers.4.experts.down_proj": "model-00009.safetensors",
+    "model.layers.4.experts.gate_up_proj": "model-00009.safetensors",
+    "model.layers.4.input_layernorm.weight": "model-00009.safetensors",
+    "model.layers.4.layer_scalar": "model-00009.safetensors",
+    "model.layers.4.mlp.down_proj.weight": "model-00009.safetensors",
+    "model.layers.4.mlp.gate_proj.weight": "model-00009.safetensors",
+    "model.layers.4.mlp.up_proj.weight": "model-00009.safetensors",
+    "model.layers.4.post_attention_layernorm.weight": "model-00009.safetensors",
+    "model.layers.4.post_feedforward_layernorm.weight": "model-00009.safetensors",
+    "model.layers.4.post_feedforward_layernorm_1.weight": "model-00009.safetensors",
+    "model.layers.4.post_feedforward_layernorm_2.weight": "model-00009.safetensors",
+    "model.layers.4.pre_feedforward_layernorm.weight": "model-00009.safetensors",
+    "model.layers.4.pre_feedforward_layernorm_2.weight": "model-00009.safetensors",
+    "model.layers.4.router.per_expert_scale": "model-00009.safetensors",
+    "model.layers.4.router.proj.weight": "model-00009.safetensors",
+    "model.layers.4.router.scale": "model-00009.safetensors",
+    "model.layers.4.self_attn.k_norm.weight": "model-00009.safetensors",
+    "model.layers.4.self_attn.k_proj.weight": "model-00009.safetensors",
+    "model.layers.4.self_attn.o_proj.weight": "model-00009.safetensors",
+    "model.layers.4.self_attn.q_norm.weight": "model-00009.safetensors",
+    "model.layers.4.self_attn.q_proj.weight": "model-00009.safetensors",
+    "model.layers.4.self_attn.v_proj.weight": "model-00009.safetensors",
+    "model.layers.5.experts.down_proj": "model-00009.safetensors",
+    "model.layers.5.experts.gate_up_proj": "model-00009.safetensors",
+    "model.layers.5.layer_scalar": "model-00010.safetensors",
+    "model.layers.5.router.per_expert_scale": "model-00010.safetensors",
+    "model.layers.5.router.scale": "model-00010.safetensors",
+    "model.layers.5.self_attn.k_norm.weight": "model-00010.safetensors",
+    "model.layers.5.self_attn.k_proj.weight": "model-00010.safetensors",
+    "model.layers.5.self_attn.o_proj.weight": "model-00010.safetensors",
+    "model.layers.5.self_attn.q_norm.weight": "model-00010.safetensors",
+    "model.layers.5.self_attn.q_proj.weight": "model-00010.safetensors",
+    "model.layers.6.experts.down_proj": "model-00010.safetensors",
+    "model.layers.6.experts.gate_up_proj": "model-00010.safetensors",
+    "model.layers.6.input_layernorm.weight": "model-00010.safetensors",
+    "model.layers.6.layer_scalar": "model-00010.safetensors",
+    "model.layers.6.mlp.down_proj.weight": "model-00010.safetensors",
+    "model.layers.6.mlp.gate_proj.weight": "model-00010.safetensors",
+    "model.layers.6.mlp.up_proj.weight": "model-00010.safetensors",
+    "model.layers.6.post_attention_layernorm.weight": "model-00010.safetensors",
+    "model.layers.6.post_feedforward_layernorm.weight": "model-00010.safetensors",
+    "model.layers.6.post_feedforward_layernorm_1.weight": "model-00010.safetensors",
+    "model.layers.6.post_feedforward_layernorm_2.weight": "model-00010.safetensors",
+    "model.layers.6.pre_feedforward_layernorm.weight": "model-00010.safetensors",
+    "model.layers.6.pre_feedforward_layernorm_2.weight": "model-00010.safetensors",
+    "model.layers.6.router.per_expert_scale": "model-00010.safetensors",
+    "model.layers.6.router.proj.weight": "model-00010.safetensors",
+    "model.layers.6.router.scale": "model-00010.safetensors",
+    "model.layers.6.self_attn.k_norm.weight": "model-00010.safetensors",
+    "model.layers.6.self_attn.k_proj.weight": "model-00010.safetensors",
+    "model.layers.6.self_attn.o_proj.weight": "model-00010.safetensors",
+    "model.layers.6.self_attn.q_norm.weight": "model-00010.safetensors",
+    "model.layers.6.self_attn.q_proj.weight": "model-00010.safetensors",
+    "model.layers.6.self_attn.v_proj.weight": "model-00010.safetensors",
+    "model.layers.7.experts.down_proj": "model-00010.safetensors",
+    "model.layers.7.experts.gate_up_proj": "model-00010.safetensors",
+    "model.layers.7.input_layernorm.weight": "model-00010.safetensors",
+    "model.layers.7.layer_scalar": "model-00010.safetensors",
+    "model.layers.7.mlp.down_proj.weight": "model-00010.safetensors",
+    "model.layers.7.mlp.gate_proj.weight": "model-00010.safetensors",
+    "model.layers.7.mlp.up_proj.weight": "model-00010.safetensors",
+    "model.layers.7.post_attention_layernorm.weight": "model-00010.safetensors",
+    "model.layers.7.post_feedforward_layernorm.weight": "model-00010.safetensors",
+    "model.layers.7.post_feedforward_layernorm_1.weight": "model-00010.safetensors",
+    "model.layers.7.post_feedforward_layernorm_2.weight": "model-00010.safetensors",
+    "model.layers.7.pre_feedforward_layernorm.weight": "model-00010.safetensors",
+    "model.layers.7.pre_feedforward_layernorm_2.weight": "model-00010.safetensors",
+    "model.layers.7.router.per_expert_scale": "model-00010.safetensors",
+    "model.layers.7.router.proj.weight": "model-00010.safetensors",
+    "model.layers.7.router.scale": "model-00010.safetensors",
+    "model.layers.7.self_attn.k_norm.weight": "model-00010.safetensors",
+    "model.layers.7.self_attn.k_proj.weight": "model-00010.safetensors",
+    "model.layers.7.self_attn.o_proj.weight": "model-00010.safetensors",
+    "model.layers.7.self_attn.q_norm.weight": "model-00010.safetensors",
+    "model.layers.7.self_attn.q_proj.weight": "model-00010.safetensors",
+    "model.layers.7.self_attn.v_proj.weight": "model-00010.safetensors",
+    "model.layers.8.experts.down_proj": "model-00010.safetensors",
+    "model.layers.8.experts.gate_up_proj": "model-00010.safetensors",
+    "model.layers.8.input_layernorm.weight": "model-00011.safetensors",
+    "model.layers.8.layer_scalar": "model-00011.safetensors",
+    "model.layers.8.mlp.down_proj.weight": "model-00011.safetensors",
+    "model.layers.8.mlp.gate_proj.weight": "model-00011.safetensors",
+    "model.layers.8.mlp.up_proj.weight": "model-00011.safetensors",
+    "model.layers.8.post_attention_layernorm.weight": "model-00011.safetensors",
+    "model.layers.8.post_feedforward_layernorm.weight": "model-00011.safetensors",
+    "model.layers.8.post_feedforward_layernorm_1.weight": "model-00011.safetensors",
+    "model.layers.8.post_feedforward_layernorm_2.weight": "model-00011.safetensors",
+    "model.layers.8.pre_feedforward_layernorm.weight": "model-00011.safetensors",
+    "model.layers.8.pre_feedforward_layernorm_2.weight": "model-00011.safetensors",
+    "model.layers.8.router.per_expert_scale": "model-00011.safetensors",
+    "model.layers.8.router.proj.weight": "model-00011.safetensors",
+    "model.layers.8.router.scale": "model-00011.safetensors",
+    "model.layers.8.self_attn.k_norm.weight": "model-00011.safetensors",
+    "model.layers.8.self_attn.k_proj.weight": "model-00011.safetensors",
+    "model.layers.8.self_attn.o_proj.weight": "model-00011.safetensors",
+    "model.layers.8.self_attn.q_norm.weight": "model-00011.safetensors",
+    "model.layers.8.self_attn.q_proj.weight": "model-00011.safetensors",
+    "model.layers.8.self_attn.v_proj.weight": "model-00011.safetensors",
+    "model.layers.9.experts.down_proj": "model-00011.safetensors",
+    "model.layers.9.experts.gate_up_proj": "model-00011.safetensors",
+    "model.layers.9.input_layernorm.weight": "model-00011.safetensors",
+    "model.layers.9.layer_scalar": "model-00011.safetensors",
+    "model.layers.9.mlp.down_proj.weight": "model-00011.safetensors",
+    "model.layers.9.mlp.gate_proj.weight": "model-00011.safetensors",
+    "model.layers.9.mlp.up_proj.weight": "model-00011.safetensors",
+    "model.layers.9.post_attention_layernorm.weight": "model-00011.safetensors",
+    "model.layers.9.post_feedforward_layernorm.weight": "model-00011.safetensors",
+    "model.layers.9.post_feedforward_layernorm_1.weight": "model-00011.safetensors",
+    "model.layers.9.post_feedforward_layernorm_2.weight": "model-00011.safetensors",
+    "model.layers.9.pre_feedforward_layernorm.weight": "model-00011.safetensors",
+    "model.layers.9.pre_feedforward_layernorm_2.weight": "model-00011.safetensors",
+    "model.layers.9.router.per_expert_scale": "model-00011.safetensors",
+    "model.layers.9.router.proj.weight": "model-00011.safetensors",
+    "model.layers.9.router.scale": "model-00011.safetensors",
+    "model.layers.9.self_attn.k_norm.weight": "model-00011.safetensors",
+    "model.layers.9.self_attn.k_proj.weight": "model-00011.safetensors",
+    "model.layers.9.self_attn.o_proj.weight": "model-00011.safetensors",
+    "model.layers.9.self_attn.q_norm.weight": "model-00011.safetensors",
+    "model.layers.9.self_attn.q_proj.weight": "model-00011.safetensors",
+    "model.layers.9.self_attn.v_proj.weight": "model-00011.safetensors",
+    "model.norm.weight": "model-00011.safetensors",
+    "model.layers.11.input_layernorm.weight": "model-00011.safetensors",
+    "model.layers.11.mlp.down_proj.weight": "model-00011.safetensors",
+    "model.layers.11.mlp.gate_proj.weight": "model-00011.safetensors",
+    "model.layers.11.mlp.up_proj.weight": "model-00011.safetensors",
+    "model.layers.11.post_attention_layernorm.weight": "model-00011.safetensors",
+    "model.layers.11.post_feedforward_layernorm.weight": "model-00011.safetensors",
+    "model.layers.11.post_feedforward_layernorm_1.weight": "model-00011.safetensors",
+    "model.layers.11.post_feedforward_layernorm_2.weight": "model-00011.safetensors",
+    "model.layers.11.pre_feedforward_layernorm.weight": "model-00011.safetensors",
+    "model.layers.11.pre_feedforward_layernorm_2.weight": "model-00011.safetensors",
+    "model.layers.11.router.proj.weight": "model-00011.safetensors",
+    "model.layers.17.experts.down_proj": "model-00011.safetensors",
+    "model.layers.17.input_layernorm.weight": "model-00011.safetensors",
+    "model.layers.17.mlp.down_proj.weight": "model-00011.safetensors",
+    "model.layers.17.mlp.gate_proj.weight": "model-00011.safetensors",
+    "model.layers.17.mlp.up_proj.weight": "model-00011.safetensors",
+    "model.layers.17.post_attention_layernorm.weight": "model-00011.safetensors",
+    "model.layers.17.post_feedforward_layernorm.weight": "model-00011.safetensors",
+    "model.layers.17.post_feedforward_layernorm_1.weight": "model-00011.safetensors",
+    "model.layers.17.post_feedforward_layernorm_2.weight": "model-00011.safetensors",
+    "model.layers.17.pre_feedforward_layernorm.weight": "model-00011.safetensors",
+    "model.layers.17.pre_feedforward_layernorm_2.weight": "model-00011.safetensors",
+    "model.layers.17.router.proj.weight": "model-00011.safetensors",
+    "model.layers.23.experts.down_proj": "model-00011.safetensors",
+    "model.layers.23.input_layernorm.weight": "model-00011.safetensors",
+    "model.layers.23.mlp.down_proj.weight": "model-00011.safetensors",
+    "model.layers.23.mlp.gate_proj.weight": "model-00011.safetensors",
+    "model.layers.23.mlp.up_proj.weight": "model-00011.safetensors",
+    "model.layers.23.post_attention_layernorm.weight": "model-00011.safetensors",
+    "model.layers.23.post_feedforward_layernorm.weight": "model-00011.safetensors",
+    "model.layers.23.post_feedforward_layernorm_1.weight": "model-00011.safetensors",
+    "model.layers.23.post_feedforward_layernorm_2.weight": "model-00011.safetensors",
+    "model.layers.23.pre_feedforward_layernorm.weight": "model-00011.safetensors",
+    "model.layers.23.pre_feedforward_layernorm_2.weight": "model-00011.safetensors",
+    "model.layers.23.router.proj.weight": "model-00011.safetensors",
+    "model.layers.29.experts.down_proj": "model-00011.safetensors",
+    "model.layers.29.input_layernorm.weight": "model-00011.safetensors",
+    "model.layers.29.mlp.down_proj.weight": "model-00011.safetensors",
+    "model.layers.29.mlp.gate_proj.weight": "model-00011.safetensors",
+    "model.layers.29.mlp.up_proj.weight": "model-00011.safetensors",
+    "model.layers.29.post_attention_layernorm.weight": "model-00011.safetensors",
+    "model.layers.29.post_feedforward_layernorm.weight": "model-00011.safetensors",
+    "model.layers.29.post_feedforward_layernorm_1.weight": "model-00011.safetensors",
+    "model.layers.29.post_feedforward_layernorm_2.weight": "model-00011.safetensors",
+    "model.layers.29.pre_feedforward_layernorm.weight": "model-00011.safetensors",
+    "model.layers.29.pre_feedforward_layernorm_2.weight": "model-00011.safetensors",
+    "model.layers.29.router.proj.weight": "model-00011.safetensors",
+    "model.layers.5.input_layernorm.weight": "model-00011.safetensors",
+    "model.layers.5.mlp.down_proj.weight": "model-00011.safetensors",
+    "model.layers.5.mlp.gate_proj.weight": "model-00011.safetensors",
+    "model.layers.5.mlp.up_proj.weight": "model-00011.safetensors",
+    "model.layers.5.post_attention_layernorm.weight": "model-00011.safetensors",
+    "model.layers.5.post_feedforward_layernorm.weight": "model-00011.safetensors",
+    "model.layers.5.post_feedforward_layernorm_1.weight": "model-00011.safetensors",
+    "model.layers.5.post_feedforward_layernorm_2.weight": "model-00011.safetensors",
+    "model.layers.5.pre_feedforward_layernorm.weight": "model-00011.safetensors",
+    "model.layers.5.pre_feedforward_layernorm_2.weight": "model-00011.safetensors",
+    "model.layers.5.router.proj.weight": "model-00011.safetensors"
+  }
+}

tokenizer.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:95e99f5b3d0df745980cc851d270a89b4dbeb6ea75ab27cff723f836b764c517
+size 13592405

tokenizer_config.json

@@ -0,0 +1,112 @@
+{
+  "backend": "tokenizers",
+  "bos_token": "<bos>",
+  "eoc_token": "<channel|>",
+  "eos_token": "<eos>",
+  "eot_token": "<turn|>",
+  "escape_token": "<|\"|>",
+  "etc_token": "<tool_call|>",
+  "etd_token": "<tool|>",
+  "etr_token": "<tool_response|>",
+  "mask_token": "<mask>",
+  "model_max_length": 1000000000000000019884624838656,
+  "pad_token": "<pad>",
+  "padding_side": "left",
+  "processor_class": "Gemma4Processor",
+  "response_schema": {
+    "type": "object",
+    "properties": {
+      "role": {
+        "const": "assistant"
+      },
+      "thinking": {
+        "type": "string"
+      },
+      "content": {
+        "type": "string"
+      },
+      "tool_calls": {
+        "x-regex-iterator": "<\\|tool_call>(.*?)<tool_call\\|>",
+        "type": "array",
+        "items": {
+          "type": "object",
+          "properties": {
+            "type": {
+              "const": "function"
+            },
+            "function": {
+              "type": "object",
+              "x-regex": "call\\:(?P<name>\\w+)(?P<arguments>\\{.*\\})",
+              "properties": {
+                "name": {
+                  "type": "string"
+                },
+                "arguments": {
+                  "type": "object",
+                  "x-parser": "gemma4-tool-call",
+                  "additionalProperties": {}
+                }
+              }
+            }
+          }
+        }
+      }
+    },
+    "x-regex": "(\\<\\|channel\\>thought\\n(?P<thinking>.*?)\\<channel\\|\\>)?(?P<tool_calls>\\<\\|tool_call\\>.*\\<tool_call\\|\\>)?(?P<content>(?:(?!\\<turn\\|\\>)(?!\\<\\|tool_response\\>).)+)?(?:\\<turn\\|\\>|\\<\\|tool_response\\>)?"
+  },
+  "response_template": {
+    "defaults": {
+      "role": "assistant"
+    },
+    "fields": {
+      "content": {
+        "close": [
+          "<turn|>",
+          "<|tool_response>",
+          "<eos>"
+        ],
+        "content": "text"
+      },
+      "thinking": {
+        "close": "<channel|>",
+        "content": "text",
+        "open": "<|channel>thought\n"
+      },
+      "tool_calls": {
+        "close": "<tool_call|>",
+        "content": "json",
+        "content_args": {
+          "string_delims": [
+            [
+              "<|\"|>",
+              "<|\"|>"
+            ]
+          ],
+          "unquoted_keys": true
+        },
+        "open_pattern": "<\\|tool_call>call:(?P<name>\\w+)",
+        "repeats": true,
+        "transform": {
+          "function": {
+            "arguments": "{content}",
+            "name": "{name}"
+          },
+          "type": "function"
+        }
+      }
+    },
+    "start_anchor": [
+      "<|turn>model\n",
+      "<tool_response|>"
+    ]
+  },
+  "soc_token": "<|channel>",
+  "sot_token": "<|turn>",
+  "stc_token": "<|tool_call>",
+  "std_token": "<|tool>",
+  "str_token": "<|tool_response>",
+  "think_token": "<|think|>",
+  "tokenizer_class": "GemmaTokenizer",
+  "unk_token": "<unk>",
+  "added_tokens_decoder": {}
+}