Upload folder using huggingface_hub

.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

README.md

@@ -0,0 +1,128 @@
+---
+license: gemma
+base_model: google/gemma-4-12B-it
+library_name: mlx
+pipeline_tag: image-text-to-text
+thumbnail: dealign_mascot.png
+tags:
+- mlx
+- apple-silicon
+- abliterated
+- uncensored
+- crack
+- vision
+- audio
+- omni-modal
+- gemma4
+- gemma
+- google
+- harmbench
+- mmlu
+- mxfp8
+---
+
+<!-- vmlx-banner -->
+<div align="center">
+<a href="https://vmlx.net">
+<img src="vmlx-banner.png" width="280" />
+<br/>
+<strong>Built for vMLX</strong> — the MLX inferencer with KV-cache quantization, prefix-cache reuse, agentic tool calling, and hybrid sliding+full attention support.
+<br/>
+<sub>Free for macOS · <strong>vmlx.net</strong></sub>
+</a>
+</div>
+
+---
+
+<div align="center">
+
+<img src="dealign_mascot.png" width="128" />
+
+# Gemma 4 12B-it — MXFP8 CRACK
+
+**CRACK abliterated** · **8-bit microscaling** · Omni-modal (text + image + audio + video) · 13 GB
+
+<a href="https://ko-fi.com/dealignai"><img src="https://img.shields.io/badge/Ko--fi-Support_Development-FF5E5B?logo=ko-fi&logoColor=white&style=for-the-badge" alt="Ko-fi"></a>
+
+</div>
+
+---
+
+## What Is This?
+
+This is [Gemma 4 12B-it](https://huggingface.co/google/gemma-4-12B-it) by Google — a unified omni-modal language model (text + image + audio + video, hybrid sliding/full attention, 48 layers, 128k context) that has been:
+
+1. **CRACK abliterated** — safety refusal removed at the weight level. The model now complies across all task categories instead of refusing, while keeping its knowledge, reasoning, and multimodal capabilities intact.
+2. **8-bit microscaling quantized** for MLX on Apple Silicon — 13 GB.
+
+## Results
+
+Evaluated through the Osaurus runtime on a Mac Studio M3 Ultra. Compliance graded via HarmBench-text refusal classifier; MMLU via logit-mode argmax over A/B/C/D token logits (matched on both base and CRACK with identical chat-template rendering — no answer truncated).
+
+### HarmBench compliance (70 prompts · 10 per category)
+
+| Category | CRACK ASR |
+|---|---:|
+| Chemical / biological | **10 / 10** (100%) |
+| Copyright | **10 / 10** (100%) |
+| Cybercrime / intrusion | **10 / 10** (100%) |
+| Harassment / bullying | **10 / 10** (100%) |
+| Illegal | **10 / 10** (100%) |
+| Misinformation / disinformation | **10 / 10** (100%) |
+| General harmful | **10 / 10** (100%) |
+| **Overall** | **70 / 70 (100%)** |
+
+### MMLU-228 (57-subject, 4 q per subject)
+
+| Subject area | base | CRACK | Δ |
+|---|---:|---:|---:|
+| **Overall** | 69.7% | **70.2%** | +0.5pp |
+| STEM | 65.3% | 66.7% | +1.4pp |
+| Humanities | 69.2% | 71.2% | +2.0pp |
+| Social Sciences | 77.1% | 77.1% | +0.0pp |
+| Other (medicine, business, …) | 69.6% | 67.9% | -1.7pp |
+
+## Features
+
+- **Omni-modal**: native text + image + audio + video inputs (Gemma 4's unified early-fusion encoder-free architecture)
+- **128k context** with hybrid sliding-window + full-attention layers
+- **Reasoning** via the Gemma 4 channel-marker format (`<|channel>thought ... <channel|>`)
+- **48 transformer layers**, hidden size 3840
+
+## Usage
+
+Run with the **Osaurus** runtime (recommended — supports the `gemma4_unified` architecture directly):
+
+```python
+# OpenAI-compatible chat completion
+# POST /v1/chat/completions
+{
+  "model": "dealignai/Gemma-4-12B-it-MXFP8-CRACK",
+  "messages": [{"role": "user", "content": "..."}],
+  "temperature": 0.0
+}
+```
+
+Google recommends `temperature=1.0, top_p=0.95, top_k=64` for general use.
+
+## About CRACK
+
+**CRACK** (Controlled Refusal Ablation via Calibrated Knockouts) removes safety-refusal behavior at the weight level so the model complies with all task categories while preserving reasoning quality, factual knowledge, multimodal capability, and coherence.
+
+## Support dealignai
+
+All models are built from original research and released free.
+
+**[Support us on Ko-fi](https://ko-fi.com/dealignai)** — membership gets early access and extras.
+
+[Ko-fi](https://ko-fi.com/dealignai) · [X @dealignai](https://x.com/dealignai) · [dealign.ai](https://dealign.ai)
+
+See our research: [Safety Generalization in Frontier Models](https://dealign.ai/quantsteer.html)
+
+<div align="center"><img src="dealign_logo.png" alt="dealign.ai" width="200"/></div>
+
+---
+
+## Disclaimer
+
+This model has had its safety-refusal behavior removed for research purposes. It will follow instructions across all categories without refusing. You are solely responsible for how you use it and for complying with all applicable laws. Published for AI-safety research and authorized security testing.

chat_template.jinja

@@ -0,0 +1,363 @@
+{%- 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 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 -%}
+
+{%- set ns = namespace(prev_message_type=None) -%}
+{%- set loop_messages = messages -%}
+{{- bos_token -}}
+{#- Handle System/Tool Definitions Block -#}
+{%- if (enable_thinking is defined and enable_thinking) or tools or messages[0]['role'] in ['system', 'developer'] -%}
+    {{- '<|turn>system\n' -}}
+    {#- Inject Thinking token at the very top of the FIRST system turn -#}
+    {%- if enable_thinking is defined and enable_thinking -%}
+        {{- '<|think|>\n' -}}
+        {%- set ns.prev_message_type = 'think' -%}
+    {%- endif -%}
+    {%- if 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' -%}
+        {%- if tool_choice is defined and ((tool_choice is string and tool_choice == 'required') or (tool_choice is mapping and tool_choice.get('type') == 'required')) -%}
+            {{- '\nTool use is REQUIRED for this turn: you must call exactly one of the declared tools. Output only the tool call, and copy every argument value verbatim \u2014 character for character, preserving punctuation and newlines, with no added or removed whitespace.' -}}
+        {%- endif -%}
+    {%- 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: suppress duplicate <|turn>model when previous non-tool message was also assistant -#}
+    {%- set prev_nt = namespace(role=None, found=false) -%}
+    {%- if loop.index0 > 0 -%}
+        {%- for j in range(loop.index0 - 1, -1, -1) -%}
+            {%- if not prev_nt.found -%}
+                {%- if loop_messages[j]['role'] != 'tool' -%}
+                    {%- set prev_nt.role = loop_messages[j]['role'] -%}
+                    {%- set prev_nt.found = true -%}
+                {%- endif -%}
+            {%- endif -%}
+        {%- endfor -%}
+    {%- endif -%}
+    {%- set continue_same_model_turn = (role == 'model' and prev_nt.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') -%}
+    {%- if thinking_text and loop.index0 > ns_turn.last_user_idx and message.get('tool_calls') -%}
+        {{- '<|channel>thought\n' + thinking_text + '\n<channel|>' -}}
+    {%- endif -%}
+
+            {%- if message['tool_calls'] -%}
+                {%- for tool_call in message['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 string -%}
+                        {{- function['arguments'] -}}
+                    {%- 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['tool_responses'] -%}
+                    {{- format_tool_response_block(tool_response['name'] | default('unknown'), 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') | default('unknown')) -%}
+                        {%- for tc in message['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') == 'image' -%}
+                                    {{- '<|image|>' -}}
+                                {%- elif part.get('type') == '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['content'] is string -%}
+                {%- if role == 'model' -%}
+                    {{- strip_thinking(message['content']) -}}
+                {%- else -%}
+                    {{- message['content'] | trim -}}
+                {%- endif -%}
+            {%- elif message['content'] is sequence -%}
+                {%- for item in message['content'] -%}
+                    {%- if item['type'] == 'text' -%}
+                        {%- if role == 'model' -%}
+                            {{- strip_thinking(item['text']) -}}
+                        {%- else -%}
+                            {{- item['text'] | trim -}}
+                        {%- endif -%}
+                    {%- elif item['type'] == 'image' -%}
+                        {{- '<|image|>' -}}
+                        {%- set ns.prev_message_type = 'image' -%}
+                    {%- elif item['type'] == 'audio' -%}
+                        {{- '<|audio|>' -}}
+                        {%- set ns.prev_message_type = 'audio' -%}
+                    {%- elif item['type'] == 'video' -%}
+                        {{- '<|video|>' -}}
+                        {%- set ns.prev_message_type = 'video' -%}
+                    {%- endif -%}
+                {%- endfor -%}
+            {%- endif -%}
+            {%- endset -%}
+
+            {{- captured_content -}}
+            {%- set has_content = captured_content | trim | length > 0 -%}
+
+        {%- if ns.prev_message_type == 'tool_call' and not ns_tr_out.flag -%}
+            {{- '<|tool_response>' -}}
+        {%- elif not (ns_tr_out.flag and not has_content) -%}
+            {{- '<turn|>\n' -}}
+        {%- endif -%}
+    {%- endif -%}
+{%- endfor -%}
+
+{%- if add_generation_prompt -%}
+    {%- if ns.prev_message_type != 'tool_response' and ns.prev_message_type != 'tool_call' -%}
+        {{- '<|turn>model\n' -}}
+    {%- endif -%}
+{%- endif -%}

config.json

@@ -0,0 +1,193 @@
+{
+  "architectures": [
+    "Gemma4UnifiedForConditionalGeneration"
+  ],
+  "audio_config": {
+    "_name_or_path": "",
+    "architectures": null,
+    "audio_embed_dim": 640,
+    "audio_samples_per_token": 640,
+    "chunk_size_feed_forward": 0,
+    "dtype": null,
+    "hidden_size": 640,
+    "id2label": {
+      "0": "LABEL_0",
+      "1": "LABEL_1"
+    },
+    "initializer_range": 0.02,
+    "is_encoder_decoder": false,
+    "label2id": {
+      "LABEL_0": 0,
+      "LABEL_1": 1
+    },
+    "model_type": "gemma4_unified_audio",
+    "output_attentions": false,
+    "output_hidden_states": false,
+    "output_proj_dims": 640,
+    "problem_type": null,
+    "return_dict": true,
+    "rms_norm_eps": 1e-06
+  },
+  "audio_token_id": 258881,
+  "boa_token_id": 256000,
+  "boi_token_id": 255999,
+  "dtype": "bfloat16",
+  "eoa_token_index": 258883,
+  "eoi_token_id": 258882,
+  "eos_token_id": [
+    1,
+    106
+  ],
+  "image_token_id": 258880,
+  "initializer_range": 0.02,
+  "model_type": "gemma4_unified",
+  "text_config": {
+    "attention_bias": false,
+    "attention_dropout": 0.0,
+    "attention_k_eq_v": true,
+    "bos_token_id": 2,
+    "enable_moe_block": false,
+    "eos_token_id": 1,
+    "final_logit_softcapping": 30.0,
+    "global_head_dim": 512,
+    "head_dim": 256,
+    "hidden_activation": "gelu_pytorch_tanh",
+    "hidden_size": 3840,
+    "hidden_size_per_layer_input": 0,
+    "initializer_range": 0.02,
+    "intermediate_size": 15360,
+    "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",
+      "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"
+    ],
+    "max_position_embeddings": 131072,
+    "model_type": "gemma4_unified_text",
+    "moe_intermediate_size": null,
+    "num_attention_heads": 16,
+    "num_experts": null,
+    "num_global_key_value_heads": 1,
+    "num_hidden_layers": 48,
+    "num_key_value_heads": 8,
+    "num_kv_shared_layers": 0,
+    "pad_token_id": 0,
+    "rms_norm_eps": 1e-06,
+    "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"
+      }
+    },
+    "sliding_window": 1024,
+    "tie_word_embeddings": true,
+    "top_k_experts": null,
+    "use_bidirectional_attention": "vision",
+    "use_cache": true,
+    "use_double_wide_mlp": false,
+    "vocab_size": 262144,
+    "vocab_size_per_layer_input": 262144
+  },
+  "tie_word_embeddings": true,
+  "transformers_version": "5.10.0.dev0",
+  "video_token_id": 258884,
+  "vision_config": {
+    "_name_or_path": "",
+    "architectures": null,
+    "chunk_size_feed_forward": 0,
+    "dtype": null,
+    "id2label": {
+      "0": "LABEL_0",
+      "1": "LABEL_1"
+    },
+    "initializer_range": 0.02,
+    "is_encoder_decoder": false,
+    "label2id": {
+      "LABEL_0": 0,
+      "LABEL_1": 1
+    },
+    "mm_embed_dim": 3840,
+    "mm_posemb_size": 1120,
+    "model_patch_size": 48,
+    "model_type": "gemma4_unified_vision",
+    "num_soft_tokens": 280,
+    "output_attentions": false,
+    "output_hidden_states": false,
+    "output_proj_dims": 3840,
+    "patch_size": 16,
+    "pooling_kernel_size": 3,
+    "problem_type": null,
+    "return_dict": true,
+    "rms_norm_eps": 1e-06
+  },
+  "weight_format": "mxfp8",
+  "quantization": {
+    "bits": 8,
+    "group_size": 32,
+    "mode": "mxfp8",
+    "quantization_backend": "mx.quantize",
+    "norm_convention": "gemma4_scale_shift_zero",
+    "tied_embedding": "fp16_passthrough",
+    "multimodal": "fp16_passthrough_embedders_early_fusion",
+    "mtp": "none"
+  },
+  "capabilities": {
+    "reasoning_parser": "gemma4",
+    "tool_parser": "gemma4",
+    "think_in_template": false,
+    "supports_tools": true,
+    "supports_thinking": true,
+    "family": "gemma4",
+    "modality": "vision",
+    "cache_type": "kv"
+  }
+}

generation_config.json

@@ -0,0 +1,18 @@
+{
+  "bos_token_id": 2,
+  "do_sample": true,
+  "eos_token_id": [
+    1,
+    106,
+    50
+  ],
+  "pad_token_id": 0,
+  "suppress_tokens": [
+    258883,
+    258882
+  ],
+  "temperature": 1.0,
+  "top_k": 64,
+  "top_p": 0.95,
+  "transformers_version": "5.10.0.dev0"
+}

jang_config.json

@@ -0,0 +1,53 @@
+{
+  "version": 2,
+  "weight_format": "mxfp8",
+  "profile": "MXFP8",
+  "source_model": {
+    "name": "gemma-4-12B-it",
+    "architecture": "gemma4_unified_text"
+  },
+  "has_vision": true,
+  "has_audio": true,
+  "quantization": {
+    "method": "mxfp8",
+    "quantization_backend": "mx.quantize",
+    "mode": "mxfp8",
+    "group_size": 32,
+    "bits": 8,
+    "norm_convention": "gemma4_scale_shift_zero",
+    "tied_embedding": "fp16_passthrough",
+    "multimodal": "fp16_passthrough_embedders_early_fusion",
+    "mtp_policy": "none",
+    "passthrough_bit_widths_used": [
+      16
+    ],
+    "passthrough_tensor_count": 349
+  },
+  "runtime": {
+    "total_weight_bytes": 13360263832,
+    "total_weight_gb": 12.44,
+    "attention": "hybrid_swa_full_5to1",
+    "sliding_window": 1024,
+    "attention_k_eq_v_on_full_layers": true,
+    "full_attention_layers": [
+      5,
+      11,
+      17,
+      23,
+      29,
+      35,
+      41,
+      47
+    ]
+  },
+  "capabilities": {
+    "reasoning_parser": "gemma4",
+    "tool_parser": "gemma4",
+    "think_in_template": false,
+    "supports_tools": true,
+    "supports_thinking": true,
+    "family": "gemma4",
+    "modality": "vision",
+    "cache_type": "kv"
+  }
+}

model-00001-of-00013.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:951a2a2c026b9810575744a815cdf33dde57a1547ecff8e46d467745fa286555
+size 2047672664

model-00002-of-00013.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:db98aac1e835c32f64a2431356152e2e80a7b7f88c145ffc42a4314e5d89d575
+size 1001925626

model-00003-of-00013.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2ec736c3205e34965b0e6debcf49c17fad4d4f8762d8eac387b81ac1d655319f
+size 1046338600

model-00004-of-00013.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:35c9cf82ff22c6fa8f70442b3a36ab44a36d1b2120d71b98d18ab8d95f83daa5
+size 1052455322

model-00005-of-00013.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a0b55d0fda045480fb5ca6afe33303d3bde32b3ff23cecf072970157fcff2e10
+size 1003761085

model-00006-of-00013.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8a35f38a82fb3b0b33bcf90db13038e4de0121108d954285b2aed9817185ca28
+size 1011219168

model-00007-of-00013.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3391a34070cf623c65870cd1520ad4a29659e07f9ffafb59741d099624a38b77
+size 1026748708

model-00008-of-00013.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1e8c98587a233c94ecb1ed9e6282212d82729fd352c2641e238962798ff756f3
+size 1003761069

model-00009-of-00013.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c9aa3461b76044c1a9933ec2b1088c765ab42a4d4a684588ec82c03af04e4e92
+size 1011219136

model-00010-of-00013.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:889daeb0ffb4ba58e36b261a260077b1c9421f089f957b209704cd0ed0c3df6b
+size 1026748746

model-00011-of-00013.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b00a0f0a909e33f5f22ef765a1b48a19ba1642f962c59e3c87d14a59dc351a65
+size 1003761072

model-00012-of-00013.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:87e61d7c6a7eaab1944d263fce9851422cf6fa7bf7aca7901a8ef8c08289b7b8
+size 1003760992

model-00013-of-00013.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:50639f0b60605b7a26990c0e50f30809950fd0f9c8d019fbdbe7d70dbab2f017
+size 120891632

model.safetensors.index.json

@@ -0,0 +1,1013 @@
+{
+  "metadata": {
+    "format": "mxfp8",
+    "total_size": 13360263832
+  },
+  "weight_map": {
+    "embed_audio.embedding_projection.weight": "model-00001-of-00013.safetensors",
+    "embed_vision.embedding_projection.weight": "model-00001-of-00013.safetensors",
+    "language_model.model.embed_tokens.weight": "model-00001-of-00013.safetensors",
+    "language_model.model.layers.0.input_layernorm.weight": "model-00002-of-00013.safetensors",
+    "language_model.model.layers.0.layer_scalar": "model-00002-of-00013.safetensors",
+    "language_model.model.layers.0.mlp.down_proj.weight": "model-00002-of-00013.safetensors",
+    "language_model.model.layers.0.mlp.down_proj.scales": "model-00002-of-00013.safetensors",
+    "language_model.model.layers.0.mlp.gate_proj.weight": "model-00002-of-00013.safetensors",
+    "language_model.model.layers.0.mlp.gate_proj.scales": "model-00002-of-00013.safetensors",
+    "language_model.model.layers.0.mlp.up_proj.weight": "model-00002-of-00013.safetensors",
+    "language_model.model.layers.0.mlp.up_proj.scales": "model-00002-of-00013.safetensors",
+    "language_model.model.layers.0.post_attention_layernorm.weight": "model-00002-of-00013.safetensors",
+    "language_model.model.layers.0.post_feedforward_layernorm.weight": "model-00002-of-00013.safetensors",
+    "language_model.model.layers.0.pre_feedforward_layernorm.weight": "model-00002-of-00013.safetensors",
+    "language_model.model.layers.0.self_attn.k_norm.weight": "model-00002-of-00013.safetensors",
+    "language_model.model.layers.0.self_attn.k_proj.weight": "model-00002-of-00013.safetensors",
+    "language_model.model.layers.0.self_attn.k_proj.scales": "model-00002-of-00013.safetensors",
+    "language_model.model.layers.0.self_attn.o_proj.weight": "model-00002-of-00013.safetensors",
+    "language_model.model.layers.0.self_attn.o_proj.scales": "model-00002-of-00013.safetensors",
+    "language_model.model.layers.0.self_attn.q_norm.weight": "model-00002-of-00013.safetensors",
+    "language_model.model.layers.0.self_attn.q_proj.weight": "model-00002-of-00013.safetensors",
+    "language_model.model.layers.0.self_attn.q_proj.scales": "model-00002-of-00013.safetensors",
+    "language_model.model.layers.0.self_attn.v_proj.weight": "model-00002-of-00013.safetensors",
+    "language_model.model.layers.0.self_attn.v_proj.scales": "model-00002-of-00013.safetensors",
+    "language_model.model.layers.1.input_layernorm.weight": "model-00002-of-00013.safetensors",
+    "language_model.model.layers.1.layer_scalar": "model-00002-of-00013.safetensors",
+    "language_model.model.layers.1.mlp.down_proj.weight": "model-00002-of-00013.safetensors",
+    "language_model.model.layers.1.mlp.down_proj.scales": "model-00002-of-00013.safetensors",
+    "language_model.model.layers.1.mlp.gate_proj.weight": "model-00002-of-00013.safetensors",
+    "language_model.model.layers.1.mlp.gate_proj.scales": "model-00002-of-00013.safetensors",
+    "language_model.model.layers.1.mlp.up_proj.weight": "model-00002-of-00013.safetensors",
+    "language_model.model.layers.1.mlp.up_proj.scales": "model-00002-of-00013.safetensors",
+    "language_model.model.layers.1.post_attention_layernorm.weight": "model-00002-of-00013.safetensors",
+    "language_model.model.layers.1.post_feedforward_layernorm.weight": "model-00002-of-00013.safetensors",
+    "language_model.model.layers.1.pre_feedforward_layernorm.weight": "model-00002-of-00013.safetensors",
+    "language_model.model.layers.1.self_attn.k_norm.weight": "model-00002-of-00013.safetensors",
+    "language_model.model.layers.1.self_attn.k_proj.weight": "model-00002-of-00013.safetensors",
+    "language_model.model.layers.1.self_attn.k_proj.scales": "model-00002-of-00013.safetensors",
+    "language_model.model.layers.1.self_attn.o_proj.weight": "model-00002-of-00013.safetensors",
+    "language_model.model.layers.1.self_attn.o_proj.scales": "model-00002-of-00013.safetensors",
+    "language_model.model.layers.1.self_attn.q_norm.weight": "model-00002-of-00013.safetensors",
+    "language_model.model.layers.1.self_attn.q_proj.weight": "model-00002-of-00013.safetensors",
+    "language_model.model.layers.1.self_attn.q_proj.scales": "model-00002-of-00013.safetensors",
+    "language_model.model.layers.1.self_attn.v_proj.weight": "model-00002-of-00013.safetensors",
+    "language_model.model.layers.1.self_attn.v_proj.scales": "model-00002-of-00013.safetensors",
+    "language_model.model.layers.10.input_layernorm.weight": "model-00002-of-00013.safetensors",
+    "language_model.model.layers.10.layer_scalar": "model-00002-of-00013.safetensors",
+    "language_model.model.layers.10.mlp.down_proj.weight": "model-00002-of-00013.safetensors",
+    "language_model.model.layers.10.mlp.down_proj.scales": "model-00002-of-00013.safetensors",
+    "language_model.model.layers.10.mlp.gate_proj.weight": "model-00002-of-00013.safetensors",
+    "language_model.model.layers.10.mlp.gate_proj.scales": "model-00002-of-00013.safetensors",
+    "language_model.model.layers.10.mlp.up_proj.weight": "model-00002-of-00013.safetensors",
+    "language_model.model.layers.10.mlp.up_proj.scales": "model-00002-of-00013.safetensors",
+    "language_model.model.layers.10.post_attention_layernorm.weight": "model-00002-of-00013.safetensors",
+    "language_model.model.layers.10.post_feedforward_layernorm.weight": "model-00002-of-00013.safetensors",
+    "language_model.model.layers.10.pre_feedforward_layernorm.weight": "model-00002-of-00013.safetensors",
+    "language_model.model.layers.10.self_attn.k_norm.weight": "model-00002-of-00013.safetensors",
+    "language_model.model.layers.10.self_attn.k_proj.weight": "model-00002-of-00013.safetensors",
+    "language_model.model.layers.10.self_attn.k_proj.scales": "model-00002-of-00013.safetensors",
+    "language_model.model.layers.10.self_attn.o_proj.weight": "model-00002-of-00013.safetensors",
+    "language_model.model.layers.10.self_attn.o_proj.scales": "model-00002-of-00013.safetensors",
+    "language_model.model.layers.10.self_attn.q_norm.weight": "model-00002-of-00013.safetensors",
+    "language_model.model.layers.10.self_attn.q_proj.weight": "model-00002-of-00013.safetensors",
+    "language_model.model.layers.10.self_attn.q_proj.scales": "model-00002-of-00013.safetensors",
+    "language_model.model.layers.10.self_attn.v_proj.weight": "model-00002-of-00013.safetensors",
+    "language_model.model.layers.10.self_attn.v_proj.scales": "model-00002-of-00013.safetensors",
+    "language_model.model.layers.11.input_layernorm.weight": "model-00002-of-00013.safetensors",
+    "language_model.model.layers.11.layer_scalar": "model-00002-of-00013.safetensors",
+    "language_model.model.layers.11.mlp.down_proj.weight": "model-00002-of-00013.safetensors",
+    "language_model.model.layers.11.mlp.down_proj.scales": "model-00002-of-00013.safetensors",
+    "language_model.model.layers.11.mlp.gate_proj.weight": "model-00002-of-00013.safetensors",
+    "language_model.model.layers.11.mlp.gate_proj.scales": "model-00002-of-00013.safetensors",
+    "language_model.model.layers.11.mlp.up_proj.weight": "model-00002-of-00013.safetensors",
+    "language_model.model.layers.11.mlp.up_proj.scales": "model-00002-of-00013.safetensors",
+    "language_model.model.layers.11.post_attention_layernorm.weight": "model-00002-of-00013.safetensors",
+    "language_model.model.layers.11.post_feedforward_layernorm.weight": "model-00002-of-00013.safetensors",
+    "language_model.model.layers.11.pre_feedforward_layernorm.weight": "model-00002-of-00013.safetensors",
+    "language_model.model.layers.11.self_attn.k_norm.weight": "model-00002-of-00013.safetensors",
+    "language_model.model.layers.11.self_attn.k_proj.weight": "model-00002-of-00013.safetensors",
+    "language_model.model.layers.11.self_attn.k_proj.scales": "model-00002-of-00013.safetensors",
+    "language_model.model.layers.11.self_attn.o_proj.weight": "model-00002-of-00013.safetensors",
+    "language_model.model.layers.11.self_attn.o_proj.scales": "model-00002-of-00013.safetensors",
+    "language_model.model.layers.11.self_attn.q_norm.weight": "model-00002-of-00013.safetensors",
+    "language_model.model.layers.11.self_attn.q_proj.weight": "model-00002-of-00013.safetensors",
+    "language_model.model.layers.11.self_attn.q_proj.scales": "model-00002-of-00013.safetensors",
+    "language_model.model.layers.12.input_layernorm.weight": "model-00002-of-00013.safetensors",
+    "language_model.model.layers.12.layer_scalar": "model-00002-of-00013.safetensors",
+    "language_model.model.layers.12.mlp.down_proj.weight": "model-00002-of-00013.safetensors",
+    "language_model.model.layers.12.mlp.down_proj.scales": "model-00003-of-00013.safetensors",
+    "language_model.model.layers.12.mlp.gate_proj.weight": "model-00003-of-00013.safetensors",
+    "language_model.model.layers.12.mlp.gate_proj.scales": "model-00003-of-00013.safetensors",
+    "language_model.model.layers.12.mlp.up_proj.weight": "model-00003-of-00013.safetensors",
+    "language_model.model.layers.12.mlp.up_proj.scales": "model-00003-of-00013.safetensors",
+    "language_model.model.layers.12.post_attention_layernorm.weight": "model-00003-of-00013.safetensors",
+    "language_model.model.layers.12.post_feedforward_layernorm.weight": "model-00003-of-00013.safetensors",
+    "language_model.model.layers.12.pre_feedforward_layernorm.weight": "model-00003-of-00013.safetensors",
+    "language_model.model.layers.12.self_attn.k_norm.weight": "model-00003-of-00013.safetensors",
+    "language_model.model.layers.12.self_attn.k_proj.weight": "model-00003-of-00013.safetensors",
+    "language_model.model.layers.12.self_attn.k_proj.scales": "model-00003-of-00013.safetensors",
+    "language_model.model.layers.12.self_attn.o_proj.weight": "model-00003-of-00013.safetensors",
+    "language_model.model.layers.12.self_attn.o_proj.scales": "model-00003-of-00013.safetensors",
+    "language_model.model.layers.12.self_attn.q_norm.weight": "model-00003-of-00013.safetensors",
+    "language_model.model.layers.12.self_attn.q_proj.weight": "model-00003-of-00013.safetensors",
+    "language_model.model.layers.12.self_attn.q_proj.scales": "model-00003-of-00013.safetensors",
+    "language_model.model.layers.12.self_attn.v_proj.weight": "model-00003-of-00013.safetensors",
+    "language_model.model.layers.12.self_attn.v_proj.scales": "model-00003-of-00013.safetensors",
+    "language_model.model.layers.13.input_layernorm.weight": "model-00003-of-00013.safetensors",
+    "language_model.model.layers.13.layer_scalar": "model-00003-of-00013.safetensors",
+    "language_model.model.layers.13.mlp.down_proj.weight": "model-00003-of-00013.safetensors",
+    "language_model.model.layers.13.mlp.down_proj.scales": "model-00003-of-00013.safetensors",
+    "language_model.model.layers.13.mlp.gate_proj.weight": "model-00003-of-00013.safetensors",
+    "language_model.model.layers.13.mlp.gate_proj.scales": "model-00003-of-00013.safetensors",
+    "language_model.model.layers.13.mlp.up_proj.weight": "model-00003-of-00013.safetensors",
+    "language_model.model.layers.13.mlp.up_proj.scales": "model-00003-of-00013.safetensors",
+    "language_model.model.layers.13.post_attention_layernorm.weight": "model-00003-of-00013.safetensors",
+    "language_model.model.layers.13.post_feedforward_layernorm.weight": "model-00003-of-00013.safetensors",
+    "language_model.model.layers.13.pre_feedforward_layernorm.weight": "model-00003-of-00013.safetensors",
+    "language_model.model.layers.13.self_attn.k_norm.weight": "model-00003-of-00013.safetensors",
+    "language_model.model.layers.13.self_attn.k_proj.weight": "model-00003-of-00013.safetensors",
+    "language_model.model.layers.13.self_attn.k_proj.scales": "model-00003-of-00013.safetensors",
+    "language_model.model.layers.13.self_attn.o_proj.weight": "model-00003-of-00013.safetensors",
+    "language_model.model.layers.13.self_attn.o_proj.scales": "model-00003-of-00013.safetensors",
+    "language_model.model.layers.13.self_attn.q_norm.weight": "model-00003-of-00013.safetensors",
+    "language_model.model.layers.13.self_attn.q_proj.weight": "model-00003-of-00013.safetensors",
+    "language_model.model.layers.13.self_attn.q_proj.scales": "model-00003-of-00013.safetensors",
+    "language_model.model.layers.13.self_attn.v_proj.weight": "model-00003-of-00013.safetensors",
+    "language_model.model.layers.13.self_attn.v_proj.scales": "model-00003-of-00013.safetensors",
+    "language_model.model.layers.14.input_layernorm.weight": "model-00003-of-00013.safetensors",
+    "language_model.model.layers.14.layer_scalar": "model-00003-of-00013.safetensors",
+    "language_model.model.layers.14.mlp.down_proj.weight": "model-00003-of-00013.safetensors",
+    "language_model.model.layers.14.mlp.down_proj.scales": "model-00003-of-00013.safetensors",
+    "language_model.model.layers.14.mlp.gate_proj.weight": "model-00003-of-00013.safetensors",
+    "language_model.model.layers.14.mlp.gate_proj.scales": "model-00003-of-00013.safetensors",
+    "language_model.model.layers.14.mlp.up_proj.weight": "model-00003-of-00013.safetensors",
+    "language_model.model.layers.14.mlp.up_proj.scales": "model-00003-of-00013.safetensors",
+    "language_model.model.layers.14.post_attention_layernorm.weight": "model-00003-of-00013.safetensors",
+    "language_model.model.layers.14.post_feedforward_layernorm.weight": "model-00003-of-00013.safetensors",
+    "language_model.model.layers.14.pre_feedforward_layernorm.weight": "model-00003-of-00013.safetensors",
+    "language_model.model.layers.14.self_attn.k_norm.weight": "model-00003-of-00013.safetensors",
+    "language_model.model.layers.14.self_attn.k_proj.weight": "model-00003-of-00013.safetensors",
+    "language_model.model.layers.14.self_attn.k_proj.scales": "model-00003-of-00013.safetensors",
+    "language_model.model.layers.14.self_attn.o_proj.weight": "model-00003-of-00013.safetensors",
+    "language_model.model.layers.14.self_attn.o_proj.scales": "model-00003-of-00013.safetensors",
+    "language_model.model.layers.14.self_attn.q_norm.weight": "model-00003-of-00013.safetensors",
+    "language_model.model.layers.14.self_attn.q_proj.weight": "model-00003-of-00013.safetensors",
+    "language_model.model.layers.14.self_attn.q_proj.scales": "model-00003-of-00013.safetensors",
+    "language_model.model.layers.14.self_attn.v_proj.weight": "model-00003-of-00013.safetensors",
+    "language_model.model.layers.14.self_attn.v_proj.scales": "model-00003-of-00013.safetensors",
+    "language_model.model.layers.15.input_layernorm.weight": "model-00003-of-00013.safetensors",
+    "language_model.model.layers.15.layer_scalar": "model-00003-of-00013.safetensors",
+    "language_model.model.layers.15.mlp.down_proj.weight": "model-00003-of-00013.safetensors",
+    "language_model.model.layers.15.mlp.down_proj.scales": "model-00003-of-00013.safetensors",
+    "language_model.model.layers.15.mlp.gate_proj.weight": "model-00003-of-00013.safetensors",
+    "language_model.model.layers.15.mlp.gate_proj.scales": "model-00003-of-00013.safetensors",
+    "language_model.model.layers.15.mlp.up_proj.weight": "model-00003-of-00013.safetensors",
+    "language_model.model.layers.15.mlp.up_proj.scales": "model-00003-of-00013.safetensors",
+    "language_model.model.layers.15.post_attention_layernorm.weight": "model-00003-of-00013.safetensors",
+    "language_model.model.layers.15.post_feedforward_layernorm.weight": "model-00003-of-00013.safetensors",
+    "language_model.model.layers.15.pre_feedforward_layernorm.weight": "model-00003-of-00013.safetensors",
+    "language_model.model.layers.15.self_attn.k_norm.weight": "model-00003-of-00013.safetensors",
+    "language_model.model.layers.15.self_attn.k_proj.weight": "model-00003-of-00013.safetensors",
+    "language_model.model.layers.15.self_attn.k_proj.scales": "model-00003-of-00013.safetensors",
+    "language_model.model.layers.15.self_attn.o_proj.weight": "model-00003-of-00013.safetensors",
+    "language_model.model.layers.15.self_attn.o_proj.scales": "model-00003-of-00013.safetensors",
+    "language_model.model.layers.15.self_attn.q_norm.weight": "model-00003-of-00013.safetensors",
+    "language_model.model.layers.15.self_attn.q_proj.weight": "model-00003-of-00013.safetensors",
+    "language_model.model.layers.15.self_attn.q_proj.scales": "model-00003-of-00013.safetensors",
+    "language_model.model.layers.15.self_attn.v_proj.weight": "model-00003-of-00013.safetensors",
+    "language_model.model.layers.15.self_attn.v_proj.scales": "model-00003-of-00013.safetensors",
+    "language_model.model.layers.16.input_layernorm.weight": "model-00003-of-00013.safetensors",
+    "language_model.model.layers.16.layer_scalar": "model-00003-of-00013.safetensors",
+    "language_model.model.layers.16.mlp.down_proj.weight": "model-00003-of-00013.safetensors",
+    "language_model.model.layers.16.mlp.down_proj.scales": "model-00003-of-00013.safetensors",
+    "language_model.model.layers.16.mlp.gate_proj.weight": "model-00003-of-00013.safetensors",
+    "language_model.model.layers.16.mlp.gate_proj.scales": "model-00003-of-00013.safetensors",
+    "language_model.model.layers.16.mlp.up_proj.weight": "model-00003-of-00013.safetensors",
+    "language_model.model.layers.16.mlp.up_proj.scales": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.16.post_attention_layernorm.weight": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.16.post_feedforward_layernorm.weight": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.16.pre_feedforward_layernorm.weight": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.16.self_attn.k_norm.weight": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.16.self_attn.k_proj.weight": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.16.self_attn.k_proj.scales": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.16.self_attn.o_proj.weight": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.16.self_attn.o_proj.scales": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.16.self_attn.q_norm.weight": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.16.self_attn.q_proj.weight": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.16.self_attn.q_proj.scales": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.16.self_attn.v_proj.weight": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.16.self_attn.v_proj.scales": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.17.input_layernorm.weight": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.17.layer_scalar": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.17.mlp.down_proj.weight": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.17.mlp.down_proj.scales": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.17.mlp.gate_proj.weight": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.17.mlp.gate_proj.scales": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.17.mlp.up_proj.weight": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.17.mlp.up_proj.scales": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.17.post_attention_layernorm.weight": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.17.post_feedforward_layernorm.weight": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.17.pre_feedforward_layernorm.weight": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.17.self_attn.k_norm.weight": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.17.self_attn.k_proj.weight": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.17.self_attn.k_proj.scales": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.17.self_attn.o_proj.weight": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.17.self_attn.o_proj.scales": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.17.self_attn.q_norm.weight": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.17.self_attn.q_proj.weight": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.17.self_attn.q_proj.scales": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.18.input_layernorm.weight": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.18.layer_scalar": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.18.mlp.down_proj.weight": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.18.mlp.down_proj.scales": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.18.mlp.gate_proj.weight": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.18.mlp.gate_proj.scales": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.18.mlp.up_proj.weight": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.18.mlp.up_proj.scales": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.18.post_attention_layernorm.weight": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.18.post_feedforward_layernorm.weight": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.18.pre_feedforward_layernorm.weight": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.18.self_attn.k_norm.weight": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.18.self_attn.k_proj.weight": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.18.self_attn.k_proj.scales": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.18.self_attn.o_proj.weight": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.18.self_attn.o_proj.scales": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.18.self_attn.q_norm.weight": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.18.self_attn.q_proj.weight": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.18.self_attn.q_proj.scales": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.18.self_attn.v_proj.weight": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.18.self_attn.v_proj.scales": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.19.input_layernorm.weight": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.19.layer_scalar": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.19.mlp.down_proj.weight": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.19.mlp.down_proj.scales": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.19.mlp.gate_proj.weight": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.19.mlp.gate_proj.scales": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.19.mlp.up_proj.weight": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.19.mlp.up_proj.scales": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.19.post_attention_layernorm.weight": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.19.post_feedforward_layernorm.weight": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.19.pre_feedforward_layernorm.weight": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.19.self_attn.k_norm.weight": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.19.self_attn.k_proj.weight": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.19.self_attn.k_proj.scales": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.19.self_attn.o_proj.weight": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.19.self_attn.o_proj.scales": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.19.self_attn.q_norm.weight": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.19.self_attn.q_proj.weight": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.19.self_attn.q_proj.scales": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.19.self_attn.v_proj.weight": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.19.self_attn.v_proj.scales": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.2.input_layernorm.weight": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.2.layer_scalar": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.2.mlp.down_proj.weight": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.2.mlp.down_proj.scales": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.2.mlp.gate_proj.weight": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.2.mlp.gate_proj.scales": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.2.mlp.up_proj.weight": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.2.mlp.up_proj.scales": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.2.post_attention_layernorm.weight": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.2.post_feedforward_layernorm.weight": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.2.pre_feedforward_layernorm.weight": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.2.self_attn.k_norm.weight": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.2.self_attn.k_proj.weight": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.2.self_attn.k_proj.scales": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.2.self_attn.o_proj.weight": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.2.self_attn.o_proj.scales": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.2.self_attn.q_norm.weight": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.2.self_attn.q_proj.weight": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.2.self_attn.q_proj.scales": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.2.self_attn.v_proj.weight": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.2.self_attn.v_proj.scales": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.20.input_layernorm.weight": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.20.layer_scalar": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.20.mlp.down_proj.weight": "model-00004-of-00013.safetensors",
+    "language_model.model.layers.20.mlp.down_proj.scales": "model-00005-of-00013.safetensors",
+    "language_model.model.layers.20.mlp.gate_proj.weight": "model-00005-of-00013.safetensors",
+    "language_model.model.layers.20.mlp.gate_proj.scales": "model-00005-of-00013.safetensors",
+    "language_model.model.layers.20.mlp.up_proj.weight": "model-00005-of-00013.safetensors",
+    "language_model.model.layers.20.mlp.up_proj.scales": "model-00005-of-00013.safetensors",
+    "language_model.model.layers.20.post_attention_layernorm.weight": "model-00005-of-00013.safetensors",
+    "language_model.model.layers.20.post_feedforward_layernorm.weight": "model-00005-of-00013.safetensors",
+    "language_model.model.layers.20.pre_feedforward_layernorm.weight": "model-00005-of-00013.safetensors",
+    "language_model.model.layers.20.self_attn.k_norm.weight": "model-00005-of-00013.safetensors",
+    "language_model.model.layers.20.self_attn.k_proj.weight": "model-00005-of-00013.safetensors",
+    "language_model.model.layers.20.self_attn.k_proj.scales": "model-00005-of-00013.safetensors",
+    "language_model.model.layers.20.self_attn.o_proj.weight": "model-00005-of-00013.safetensors",
+    "language_model.model.layers.20.self_attn.o_proj.scales": "model-00005-of-00013.safetensors",
+    "language_model.model.layers.20.self_attn.q_norm.weight": "model-00005-of-00013.safetensors",
+    "language_model.model.layers.20.self_attn.q_proj.weight": "model-00005-of-00013.safetensors",
+    "language_model.model.layers.20.self_attn.q_proj.scales": "model-00005-of-00013.safetensors",
+    "language_model.model.layers.20.self_attn.v_proj.weight": "model-00005-of-00013.safetensors",
+    "language_model.model.layers.20.self_attn.v_proj.scales": "model-00005-of-00013.safetensors",
+    "language_model.model.layers.21.input_layernorm.weight": "model-00005-of-00013.safetensors",
+    "language_model.model.layers.21.layer_scalar": "model-00005-of-00013.safetensors",
+    "language_model.model.layers.21.mlp.down_proj.weight": "model-00005-of-00013.safetensors",
+    "language_model.model.layers.21.mlp.down_proj.scales": "model-00005-of-00013.safetensors",
+    "language_model.model.layers.21.mlp.gate_proj.weight": "model-00005-of-00013.safetensors",
+    "language_model.model.layers.21.mlp.gate_proj.scales": "model-00005-of-00013.safetensors",
+    "language_model.model.layers.21.mlp.up_proj.weight": "model-00005-of-00013.safetensors",
+    "language_model.model.layers.21.mlp.up_proj.scales": "model-00005-of-00013.safetensors",
+    "language_model.model.layers.21.post_attention_layernorm.weight": "model-00005-of-00013.safetensors",
+    "language_model.model.layers.21.post_feedforward_layernorm.weight": "model-00005-of-00013.safetensors",
+    "language_model.model.layers.21.pre_feedforward_layernorm.weight": "model-00005-of-00013.safetensors",
+    "language_model.model.layers.21.self_attn.k_norm.weight": "model-00005-of-00013.safetensors",
+    "language_model.model.layers.21.self_attn.k_proj.weight": "model-00005-of-00013.safetensors",
+    "language_model.model.layers.21.self_attn.k_proj.scales": "model-00005-of-00013.safetensors",
+    "language_model.model.layers.21.self_attn.o_proj.weight": "model-00005-of-00013.safetensors",
+    "language_model.model.layers.21.self_attn.o_proj.scales": "model-00005-of-00013.safetensors",
+    "language_model.model.layers.21.self_attn.q_norm.weight": "model-00005-of-00013.safetensors",
+    "language_model.model.layers.21.self_attn.q_proj.weight": "model-00005-of-00013.safetensors",
+    "language_model.model.layers.21.self_attn.q_proj.scales": "model-00005-of-00013.safetensors",
+    "language_model.model.layers.21.self_attn.v_proj.weight": "model-00005-of-00013.safetensors",
+    "language_model.model.layers.21.self_attn.v_proj.scales": "model-00005-of-00013.safetensors",
+    "language_model.model.layers.22.input_layernorm.weight": "model-00005-of-00013.safetensors",
+    "language_model.model.layers.22.layer_scalar": "model-00005-of-00013.safetensors",
+    "language_model.model.layers.22.mlp.down_proj.weight": "model-00005-of-00013.safetensors",
+    "language_model.model.layers.22.mlp.down_proj.scales": "model-00005-of-00013.safetensors",
+    "language_model.model.layers.22.mlp.gate_proj.weight": "model-00005-of-00013.safetensors",
+    "language_model.model.layers.22.mlp.gate_proj.scales": "model-00005-of-00013.safetensors",
+    "language_model.model.layers.22.mlp.up_proj.weight": "model-00005-of-00013.safetensors",
+    "language_model.model.layers.22.mlp.up_proj.scales": "model-00005-of-00013.safetensors",
+    "language_model.model.layers.22.post_attention_layernorm.weight": "model-00005-of-00013.safetensors",
+    "language_model.model.layers.22.post_feedforward_layernorm.weight": "model-00005-of-00013.safetensors",
+    "language_model.model.layers.22.pre_feedforward_layernorm.weight": "model-00005-of-00013.safetensors",
+    "language_model.model.layers.22.self_attn.k_norm.weight": "model-00005-of-00013.safetensors",
+    "language_model.model.layers.22.self_attn.k_proj.weight": "model-00005-of-00013.safetensors",
+    "language_model.model.layers.22.self_attn.k_proj.scales": "model-00005-of-00013.safetensors",
+    "language_model.model.layers.22.self_attn.o_proj.weight": "model-00005-of-00013.safetensors",
+    "language_model.model.layers.22.self_attn.o_proj.scales": "model-00005-of-00013.safetensors",
+    "language_model.model.layers.22.self_attn.q_norm.weight": "model-00005-of-00013.safetensors",
+    "language_model.model.layers.22.self_attn.q_proj.weight": "model-00005-of-00013.safetensors",
+    "language_model.model.layers.22.self_attn.q_proj.scales": "model-00005-of-00013.safetensors",
+    "language_model.model.layers.22.self_attn.v_proj.weight": "model-00005-of-00013.safetensors",
+    "language_model.model.layers.22.self_attn.v_proj.scales": "model-00005-of-00013.safetensors",
+    "language_model.model.layers.23.input_layernorm.weight": "model-00005-of-00013.safetensors",
+    "language_model.model.layers.23.layer_scalar": "model-00005-of-00013.safetensors",
+    "language_model.model.layers.23.mlp.down_proj.weight": "model-00005-of-00013.safetensors",
+    "language_model.model.layers.23.mlp.down_proj.scales": "model-00005-of-00013.safetensors",
+    "language_model.model.layers.23.mlp.gate_proj.weight": "model-00005-of-00013.safetensors",
+    "language_model.model.layers.23.mlp.gate_proj.scales": "model-00005-of-00013.safetensors",
+    "language_model.model.layers.23.mlp.up_proj.weight": "model-00005-of-00013.safetensors",
+    "language_model.model.layers.23.mlp.up_proj.scales": "model-00005-of-00013.safetensors",
+    "language_model.model.layers.23.post_attention_layernorm.weight": "model-00005-of-00013.safetensors",
+    "language_model.model.layers.23.post_feedforward_layernorm.weight": "model-00005-of-00013.safetensors",
+    "language_model.model.layers.23.pre_feedforward_layernorm.weight": "model-00005-of-00013.safetensors",
+    "language_model.model.layers.23.self_attn.k_norm.weight": "model-00005-of-00013.safetensors",
+    "language_model.model.layers.23.self_attn.k_proj.weight": "model-00005-of-00013.safetensors",
+    "language_model.model.layers.23.self_attn.k_proj.scales": "model-00005-of-00013.safetensors",
+    "language_model.model.layers.23.self_attn.o_proj.weight": "model-00005-of-00013.safetensors",
+    "language_model.model.layers.23.self_attn.o_proj.scales": "model-00005-of-00013.safetensors",
+    "language_model.model.layers.23.self_attn.q_norm.weight": "model-00005-of-00013.safetensors",
+    "language_model.model.layers.23.self_attn.q_proj.weight": "model-00005-of-00013.safetensors",
+    "language_model.model.layers.23.self_attn.q_proj.scales": "model-00005-of-00013.safetensors",
+    "language_model.model.layers.24.input_layernorm.weight": "model-00005-of-00013.safetensors",
+    "language_model.model.layers.24.layer_scalar": "model-00005-of-00013.safetensors",
+    "language_model.model.layers.24.mlp.down_proj.weight": "model-00005-of-00013.safetensors",
+    "language_model.model.layers.24.mlp.down_proj.scales": "model-00005-of-00013.safetensors",
+    "language_model.model.layers.24.mlp.gate_proj.weight": "model-00005-of-00013.safetensors",
+    "language_model.model.layers.24.mlp.gate_proj.scales": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.24.mlp.up_proj.weight": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.24.mlp.up_proj.scales": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.24.post_attention_layernorm.weight": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.24.post_feedforward_layernorm.weight": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.24.pre_feedforward_layernorm.weight": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.24.self_attn.k_norm.weight": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.24.self_attn.k_proj.weight": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.24.self_attn.k_proj.scales": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.24.self_attn.o_proj.weight": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.24.self_attn.o_proj.scales": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.24.self_attn.q_norm.weight": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.24.self_attn.q_proj.weight": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.24.self_attn.q_proj.scales": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.24.self_attn.v_proj.weight": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.24.self_attn.v_proj.scales": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.25.input_layernorm.weight": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.25.layer_scalar": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.25.mlp.down_proj.weight": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.25.mlp.down_proj.scales": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.25.mlp.gate_proj.weight": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.25.mlp.gate_proj.scales": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.25.mlp.up_proj.weight": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.25.mlp.up_proj.scales": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.25.post_attention_layernorm.weight": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.25.post_feedforward_layernorm.weight": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.25.pre_feedforward_layernorm.weight": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.25.self_attn.k_norm.weight": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.25.self_attn.k_proj.weight": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.25.self_attn.k_proj.scales": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.25.self_attn.o_proj.weight": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.25.self_attn.o_proj.scales": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.25.self_attn.q_norm.weight": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.25.self_attn.q_proj.weight": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.25.self_attn.q_proj.scales": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.25.self_attn.v_proj.weight": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.25.self_attn.v_proj.scales": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.26.input_layernorm.weight": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.26.layer_scalar": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.26.mlp.down_proj.weight": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.26.mlp.down_proj.scales": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.26.mlp.gate_proj.weight": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.26.mlp.gate_proj.scales": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.26.mlp.up_proj.weight": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.26.mlp.up_proj.scales": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.26.post_attention_layernorm.weight": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.26.post_feedforward_layernorm.weight": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.26.pre_feedforward_layernorm.weight": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.26.self_attn.k_norm.weight": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.26.self_attn.k_proj.weight": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.26.self_attn.k_proj.scales": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.26.self_attn.o_proj.weight": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.26.self_attn.o_proj.scales": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.26.self_attn.q_norm.weight": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.26.self_attn.q_proj.weight": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.26.self_attn.q_proj.scales": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.26.self_attn.v_proj.weight": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.26.self_attn.v_proj.scales": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.27.input_layernorm.weight": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.27.layer_scalar": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.27.mlp.down_proj.weight": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.27.mlp.down_proj.scales": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.27.mlp.gate_proj.weight": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.27.mlp.gate_proj.scales": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.27.mlp.up_proj.weight": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.27.mlp.up_proj.scales": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.27.post_attention_layernorm.weight": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.27.post_feedforward_layernorm.weight": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.27.pre_feedforward_layernorm.weight": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.27.self_attn.k_norm.weight": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.27.self_attn.k_proj.weight": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.27.self_attn.k_proj.scales": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.27.self_attn.o_proj.weight": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.27.self_attn.o_proj.scales": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.27.self_attn.q_norm.weight": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.27.self_attn.q_proj.weight": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.27.self_attn.q_proj.scales": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.27.self_attn.v_proj.weight": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.27.self_attn.v_proj.scales": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.28.input_layernorm.weight": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.28.layer_scalar": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.28.mlp.down_proj.weight": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.28.mlp.down_proj.scales": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.28.mlp.gate_proj.weight": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.28.mlp.gate_proj.scales": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.28.mlp.up_proj.weight": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.28.mlp.up_proj.scales": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.28.post_attention_layernorm.weight": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.28.post_feedforward_layernorm.weight": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.28.pre_feedforward_layernorm.weight": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.28.self_attn.k_norm.weight": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.28.self_attn.k_proj.weight": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.28.self_attn.k_proj.scales": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.28.self_attn.o_proj.weight": "model-00006-of-00013.safetensors",
+    "language_model.model.layers.28.self_attn.o_proj.scales": "model-00007-of-00013.safetensors",
+    "language_model.model.layers.28.self_attn.q_norm.weight": "model-00007-of-00013.safetensors",
+    "language_model.model.layers.28.self_attn.q_proj.weight": "model-00007-of-00013.safetensors",
+    "language_model.model.layers.28.self_attn.q_proj.scales": "model-00007-of-00013.safetensors",
+    "language_model.model.layers.28.self_attn.v_proj.weight": "model-00007-of-00013.safetensors",
+    "language_model.model.layers.28.self_attn.v_proj.scales": "model-00007-of-00013.safetensors",
+    "language_model.model.layers.29.input_layernorm.weight": "model-00007-of-00013.safetensors",
+    "language_model.model.layers.29.layer_scalar": "model-00007-of-00013.safetensors",
+    "language_model.model.layers.29.mlp.down_proj.weight": "model-00007-of-00013.safetensors",
+    "language_model.model.layers.29.mlp.down_proj.scales": "model-00007-of-00013.safetensors",
+    "language_model.model.layers.29.mlp.gate_proj.weight": "model-00007-of-00013.safetensors",
+    "language_model.model.layers.29.mlp.gate_proj.scales": "model-00007-of-00013.safetensors",
+    "language_model.model.layers.29.mlp.up_proj.weight": "model-00007-of-00013.safetensors",
+    "language_model.model.layers.29.mlp.up_proj.scales": "model-00007-of-00013.safetensors",
+    "language_model.model.layers.29.post_attention_layernorm.weight": "model-00007-of-00013.safetensors",
+    "language_model.model.layers.29.post_feedforward_layernorm.weight": "model-00007-of-00013.safetensors",
+    "language_model.model.layers.29.pre_feedforward_layernorm.weight": "model-00007-of-00013.safetensors",
+    "language_model.model.layers.29.self_attn.k_norm.weight": "model-00007-of-00013.safetensors",
+    "language_model.model.layers.29.self_attn.k_proj.weight": "model-00007-of-00013.safetensors",
+    "language_model.model.layers.29.self_attn.k_proj.scales": "model-00007-of-00013.safetensors",
+    "language_model.model.layers.29.self_attn.o_proj.weight": "model-00007-of-00013.safetensors",
+    "language_model.model.layers.29.self_attn.o_proj.scales": "model-00007-of-00013.safetensors",
+    "language_model.model.layers.29.self_attn.q_norm.weight": "model-00007-of-00013.safetensors",
+    "language_model.model.layers.29.self_attn.q_proj.weight": "model-00007-of-00013.safetensors",
+    "language_model.model.layers.29.self_attn.q_proj.scales": "model-00007-of-00013.safetensors",
+    "language_model.model.layers.3.input_layernorm.weight": "model-00007-of-00013.safetensors",
+    "language_model.model.layers.3.layer_scalar": "model-00007-of-00013.safetensors",
+    "language_model.model.layers.3.mlp.down_proj.weight": "model-00007-of-00013.safetensors",
+    "language_model.model.layers.3.mlp.down_proj.scales": "model-00007-of-00013.safetensors",
+    "language_model.model.layers.3.mlp.gate_proj.weight": "model-00007-of-00013.safetensors",
+    "language_model.model.layers.3.mlp.gate_proj.scales": "model-00007-of-00013.safetensors",
+    "language_model.model.layers.3.mlp.up_proj.weight": "model-00007-of-00013.safetensors",
+    "language_model.model.layers.3.mlp.up_proj.scales": "model-00007-of-00013.safetensors",
+    "language_model.model.layers.3.post_attention_layernorm.weight": "model-00007-of-00013.safetensors",
+    "language_model.model.layers.3.post_feedforward_layernorm.weight": "model-00007-of-00013.safetensors",
+    "language_model.model.layers.3.pre_feedforward_layernorm.weight": "model-00007-of-00013.safetensors",
+    "language_model.model.layers.3.self_attn.k_norm.weight": "model-00007-of-00013.safetensors",
+    "language_model.model.layers.3.self_attn.k_proj.weight": "model-00007-of-00013.safetensors",
+    "language_model.model.layers.3.self_attn.k_proj.scales": "model-00007-of-00013.safetensors",
+    "language_model.model.layers.3.self_attn.o_proj.weight": "model-00007-of-00013.safetensors",
+    "language_model.model.layers.3.self_attn.o_proj.scales": "model-00007-of-00013.safetensors",
+    "language_model.model.layers.3.self_attn.q_norm.weight": "model-00007-of-00013.safetensors",
+    "language_model.model.layers.3.self_attn.q_proj.weight": "model-00007-of-00013.safetensors",
+    "language_model.model.layers.3.self_attn.q_proj.scales": "model-00007-of-00013.safetensors",
+    "language_model.model.layers.3.self_attn.v_proj.weight": "model-00007-of-00013.safetensors",
+    "language_model.model.layers.3.self_attn.v_proj.scales": "model-00007-of-00013.safetensors",
+    "language_model.model.layers.30.input_layernorm.weight": "model-00007-of-00013.safetensors",
+    "language_model.model.layers.30.layer_scalar": "model-00007-of-00013.safetensors",
+    "language_model.model.layers.30.mlp.down_proj.weight": "model-00007-of-00013.safetensors",
+    "language_model.model.layers.30.mlp.down_proj.scales": "model-00007-of-00013.safetensors",
+    "language_model.model.layers.30.mlp.gate_proj.weight": "model-00007-of-00013.safetensors",
+    "language_model.model.layers.30.mlp.gate_proj.scales": "model-00007-of-00013.safetensors",
+    "language_model.model.layers.30.mlp.up_proj.weight": "model-00007-of-00013.safetensors",
+    "language_model.model.layers.30.mlp.up_proj.scales": "model-00007-of-00013.safetensors",
+    "language_model.model.layers.30.post_attention_layernorm.weight": "model-00007-of-00013.safetensors",
+    "language_model.model.layers.30.post_feedforward_layernorm.weight": "model-00007-of-00013.safetensors",
+    "language_model.model.layers.30.pre_feedforward_layernorm.weight": "model-00007-of-00013.safetensors",
+    "language_model.model.layers.30.self_attn.k_norm.weight": "model-00007-of-00013.safetensors",
+    "language_model.model.layers.30.self_attn.k_proj.weight": "model-00007-of-00013.safetensors",
+    "language_model.model.layers.30.self_attn.k_proj.scales": "model-00007-of-00013.safetensors",
+    "language_model.model.layers.30.self_attn.o_proj.weight": "model-00007-of-00013.safetensors",
+    "language_model.model.layers.30.self_attn.o_proj.scales": "model-00007-of-00013.safetensors",
+    "language_model.model.layers.30.self_attn.q_norm.weight": "model-00007-of-00013.safetensors",
+    "language_model.model.layers.30.self_attn.q_proj.weight": "model-00007-of-00013.safetensors",
+    "language_model.model.layers.30.self_attn.q_proj.scales": "model-00007-of-00013.safetensors",
+    "language_model.model.layers.30.self_attn.v_proj.weight": "model-00007-of-00013.safetensors",
+    "language_model.model.layers.30.self_attn.v_proj.scales": "model-00007-of-00013.safetensors",
+    "language_model.model.layers.31.input_layernorm.weight": "model-00007-of-00013.safetensors",
+    "language_model.model.layers.31.layer_scalar": "model-00007-of-00013.safetensors",
+    "language_model.model.layers.31.mlp.down_proj.weight": "model-00007-of-00013.safetensors",
+    "language_model.model.layers.31.mlp.down_proj.scales": "model-00007-of-00013.safetensors",
+    "language_model.model.layers.31.mlp.gate_proj.weight": "model-00007-of-00013.safetensors",
+    "language_model.model.layers.31.mlp.gate_proj.scales": "model-00007-of-00013.safetensors",
+    "language_model.model.layers.31.mlp.up_proj.weight": "model-00007-of-00013.safetensors",
+    "language_model.model.layers.31.mlp.up_proj.scales": "model-00007-of-00013.safetensors",
+    "language_model.model.layers.31.post_attention_layernorm.weight": "model-00007-of-00013.safetensors",
+    "language_model.model.layers.31.post_feedforward_layernorm.weight": "model-00007-of-00013.safetensors",
+    "language_model.model.layers.31.pre_feedforward_layernorm.weight": "model-00007-of-00013.safetensors",
+    "language_model.model.layers.31.self_attn.k_norm.weight": "model-00007-of-00013.safetensors",
+    "language_model.model.layers.31.self_attn.k_proj.weight": "model-00007-of-00013.safetensors",
+    "language_model.model.layers.31.self_attn.k_proj.scales": "model-00007-of-00013.safetensors",
+    "language_model.model.layers.31.self_attn.o_proj.weight": "model-00007-of-00013.safetensors",
+    "language_model.model.layers.31.self_attn.o_proj.scales": "model-00007-of-00013.safetensors",
+    "language_model.model.layers.31.self_attn.q_norm.weight": "model-00007-of-00013.safetensors",
+    "language_model.model.layers.31.self_attn.q_proj.weight": "model-00007-of-00013.safetensors",
+    "language_model.model.layers.31.self_attn.q_proj.scales": "model-00007-of-00013.safetensors",
+    "language_model.model.layers.31.self_attn.v_proj.weight": "model-00007-of-00013.safetensors",
+    "language_model.model.layers.31.self_attn.v_proj.scales": "model-00007-of-00013.safetensors",
+    "language_model.model.layers.32.input_layernorm.weight": "model-00007-of-00013.safetensors",
+    "language_model.model.layers.32.layer_scalar": "model-00007-of-00013.safetensors",
+    "language_model.model.layers.32.mlp.down_proj.weight": "model-00007-of-00013.safetensors",
+    "language_model.model.layers.32.mlp.down_proj.scales": "model-00008-of-00013.safetensors",
+    "language_model.model.layers.32.mlp.gate_proj.weight": "model-00008-of-00013.safetensors",
+    "language_model.model.layers.32.mlp.gate_proj.scales": "model-00008-of-00013.safetensors",
+    "language_model.model.layers.32.mlp.up_proj.weight": "model-00008-of-00013.safetensors",
+    "language_model.model.layers.32.mlp.up_proj.scales": "model-00008-of-00013.safetensors",
+    "language_model.model.layers.32.post_attention_layernorm.weight": "model-00008-of-00013.safetensors",
+    "language_model.model.layers.32.post_feedforward_layernorm.weight": "model-00008-of-00013.safetensors",
+    "language_model.model.layers.32.pre_feedforward_layernorm.weight": "model-00008-of-00013.safetensors",
+    "language_model.model.layers.32.self_attn.k_norm.weight": "model-00008-of-00013.safetensors",
+    "language_model.model.layers.32.self_attn.k_proj.weight": "model-00008-of-00013.safetensors",
+    "language_model.model.layers.32.self_attn.k_proj.scales": "model-00008-of-00013.safetensors",
+    "language_model.model.layers.32.self_attn.o_proj.weight": "model-00008-of-00013.safetensors",
+    "language_model.model.layers.32.self_attn.o_proj.scales": "model-00008-of-00013.safetensors",
+    "language_model.model.layers.32.self_attn.q_norm.weight": "model-00008-of-00013.safetensors",
+    "language_model.model.layers.32.self_attn.q_proj.weight": "model-00008-of-00013.safetensors",
+    "language_model.model.layers.32.self_attn.q_proj.scales": "model-00008-of-00013.safetensors",
+    "language_model.model.layers.32.self_attn.v_proj.weight": "model-00008-of-00013.safetensors",
+    "language_model.model.layers.32.self_attn.v_proj.scales": "model-00008-of-00013.safetensors",
+    "language_model.model.layers.33.input_layernorm.weight": "model-00008-of-00013.safetensors",
+    "language_model.model.layers.33.layer_scalar": "model-00008-of-00013.safetensors",
+    "language_model.model.layers.33.mlp.down_proj.weight": "model-00008-of-00013.safetensors",
+    "language_model.model.layers.33.mlp.down_proj.scales": "model-00008-of-00013.safetensors",
+    "language_model.model.layers.33.mlp.gate_proj.weight": "model-00008-of-00013.safetensors",
+    "language_model.model.layers.33.mlp.gate_proj.scales": "model-00008-of-00013.safetensors",
+    "language_model.model.layers.33.mlp.up_proj.weight": "model-00008-of-00013.safetensors",
+    "language_model.model.layers.33.mlp.up_proj.scales": "model-00008-of-00013.safetensors",
+    "language_model.model.layers.33.post_attention_layernorm.weight": "model-00008-of-00013.safetensors",
+    "language_model.model.layers.33.post_feedforward_layernorm.weight": "model-00008-of-00013.safetensors",
+    "language_model.model.layers.33.pre_feedforward_layernorm.weight": "model-00008-of-00013.safetensors",
+    "language_model.model.layers.33.self_attn.k_norm.weight": "model-00008-of-00013.safetensors",
+    "language_model.model.layers.33.self_attn.k_proj.weight": "model-00008-of-00013.safetensors",
+    "language_model.model.layers.33.self_attn.k_proj.scales": "model-00008-of-00013.safetensors",
+    "language_model.model.layers.33.self_attn.o_proj.weight": "model-00008-of-00013.safetensors",
+    "language_model.model.layers.33.self_attn.o_proj.scales": "model-00008-of-00013.safetensors",
+    "language_model.model.layers.33.self_attn.q_norm.weight": "model-00008-of-00013.safetensors",
+    "language_model.model.layers.33.self_attn.q_proj.weight": "model-00008-of-00013.safetensors",
+    "language_model.model.layers.33.self_attn.q_proj.scales": "model-00008-of-00013.safetensors",
+    "language_model.model.layers.33.self_attn.v_proj.weight": "model-00008-of-00013.safetensors",
+    "language_model.model.layers.33.self_attn.v_proj.scales": "model-00008-of-00013.safetensors",
+    "language_model.model.layers.34.input_layernorm.weight": "model-00008-of-00013.safetensors",
+    "language_model.model.layers.34.layer_scalar": "model-00008-of-00013.safetensors",
+    "language_model.model.layers.34.mlp.down_proj.weight": "model-00008-of-00013.safetensors",
+    "language_model.model.layers.34.mlp.down_proj.scales": "model-00008-of-00013.safetensors",
+    "language_model.model.layers.34.mlp.gate_proj.weight": "model-00008-of-00013.safetensors",
+    "language_model.model.layers.34.mlp.gate_proj.scales": "model-00008-of-00013.safetensors",
+    "language_model.model.layers.34.mlp.up_proj.weight": "model-00008-of-00013.safetensors",
+    "language_model.model.layers.34.mlp.up_proj.scales": "model-00008-of-00013.safetensors",
+    "language_model.model.layers.34.post_attention_layernorm.weight": "model-00008-of-00013.safetensors",
+    "language_model.model.layers.34.post_feedforward_layernorm.weight": "model-00008-of-00013.safetensors",
+    "language_model.model.layers.34.pre_feedforward_layernorm.weight": "model-00008-of-00013.safetensors",
+    "language_model.model.layers.34.self_attn.k_norm.weight": "model-00008-of-00013.safetensors",
+    "language_model.model.layers.34.self_attn.k_proj.weight": "model-00008-of-00013.safetensors",
+    "language_model.model.layers.34.self_attn.k_proj.scales": "model-00008-of-00013.safetensors",
+    "language_model.model.layers.34.self_attn.o_proj.weight": "model-00008-of-00013.safetensors",
+    "language_model.model.layers.34.self_attn.o_proj.scales": "model-00008-of-00013.safetensors",
+    "language_model.model.layers.34.self_attn.q_norm.weight": "model-00008-of-00013.safetensors",
+    "language_model.model.layers.34.self_attn.q_proj.weight": "model-00008-of-00013.safetensors",
+    "language_model.model.layers.34.self_attn.q_proj.scales": "model-00008-of-00013.safetensors",
+    "language_model.model.layers.34.self_attn.v_proj.weight": "model-00008-of-00013.safetensors",
+    "language_model.model.layers.34.self_attn.v_proj.scales": "model-00008-of-00013.safetensors",
+    "language_model.model.layers.35.input_layernorm.weight": "model-00008-of-00013.safetensors",
+    "language_model.model.layers.35.layer_scalar": "model-00008-of-00013.safetensors",
+    "language_model.model.layers.35.mlp.down_proj.weight": "model-00008-of-00013.safetensors",
+    "language_model.model.layers.35.mlp.down_proj.scales": "model-00008-of-00013.safetensors",
+    "language_model.model.layers.35.mlp.gate_proj.weight": "model-00008-of-00013.safetensors",
+    "language_model.model.layers.35.mlp.gate_proj.scales": "model-00008-of-00013.safetensors",
+    "language_model.model.layers.35.mlp.up_proj.weight": "model-00008-of-00013.safetensors",
+    "language_model.model.layers.35.mlp.up_proj.scales": "model-00008-of-00013.safetensors",
+    "language_model.model.layers.35.post_attention_layernorm.weight": "model-00008-of-00013.safetensors",
+    "language_model.model.layers.35.post_feedforward_layernorm.weight": "model-00008-of-00013.safetensors",
+    "language_model.model.layers.35.pre_feedforward_layernorm.weight": "model-00008-of-00013.safetensors",
+    "language_model.model.layers.35.self_attn.k_norm.weight": "model-00008-of-00013.safetensors",
+    "language_model.model.layers.35.self_attn.k_proj.weight": "model-00008-of-00013.safetensors",
+    "language_model.model.layers.35.self_attn.k_proj.scales": "model-00008-of-00013.safetensors",
+    "language_model.model.layers.35.self_attn.o_proj.weight": "model-00008-of-00013.safetensors",
+    "language_model.model.layers.35.self_attn.o_proj.scales": "model-00008-of-00013.safetensors",
+    "language_model.model.layers.35.self_attn.q_norm.weight": "model-00008-of-00013.safetensors",
+    "language_model.model.layers.35.self_attn.q_proj.weight": "model-00008-of-00013.safetensors",
+    "language_model.model.layers.35.self_attn.q_proj.scales": "model-00008-of-00013.safetensors",
+    "language_model.model.layers.36.input_layernorm.weight": "model-00008-of-00013.safetensors",
+    "language_model.model.layers.36.layer_scalar": "model-00008-of-00013.safetensors",
+    "language_model.model.layers.36.mlp.down_proj.weight": "model-00008-of-00013.safetensors",
+    "language_model.model.layers.36.mlp.down_proj.scales": "model-00008-of-00013.safetensors",
+    "language_model.model.layers.36.mlp.gate_proj.weight": "model-00008-of-00013.safetensors",
+    "language_model.model.layers.36.mlp.gate_proj.scales": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.36.mlp.up_proj.weight": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.36.mlp.up_proj.scales": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.36.post_attention_layernorm.weight": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.36.post_feedforward_layernorm.weight": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.36.pre_feedforward_layernorm.weight": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.36.self_attn.k_norm.weight": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.36.self_attn.k_proj.weight": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.36.self_attn.k_proj.scales": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.36.self_attn.o_proj.weight": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.36.self_attn.o_proj.scales": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.36.self_attn.q_norm.weight": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.36.self_attn.q_proj.weight": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.36.self_attn.q_proj.scales": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.36.self_attn.v_proj.weight": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.36.self_attn.v_proj.scales": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.37.input_layernorm.weight": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.37.layer_scalar": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.37.mlp.down_proj.weight": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.37.mlp.down_proj.scales": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.37.mlp.gate_proj.weight": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.37.mlp.gate_proj.scales": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.37.mlp.up_proj.weight": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.37.mlp.up_proj.scales": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.37.post_attention_layernorm.weight": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.37.post_feedforward_layernorm.weight": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.37.pre_feedforward_layernorm.weight": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.37.self_attn.k_norm.weight": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.37.self_attn.k_proj.weight": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.37.self_attn.k_proj.scales": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.37.self_attn.o_proj.weight": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.37.self_attn.o_proj.scales": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.37.self_attn.q_norm.weight": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.37.self_attn.q_proj.weight": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.37.self_attn.q_proj.scales": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.37.self_attn.v_proj.weight": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.37.self_attn.v_proj.scales": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.38.input_layernorm.weight": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.38.layer_scalar": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.38.mlp.down_proj.weight": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.38.mlp.down_proj.scales": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.38.mlp.gate_proj.weight": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.38.mlp.gate_proj.scales": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.38.mlp.up_proj.weight": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.38.mlp.up_proj.scales": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.38.post_attention_layernorm.weight": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.38.post_feedforward_layernorm.weight": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.38.pre_feedforward_layernorm.weight": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.38.self_attn.k_norm.weight": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.38.self_attn.k_proj.weight": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.38.self_attn.k_proj.scales": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.38.self_attn.o_proj.weight": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.38.self_attn.o_proj.scales": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.38.self_attn.q_norm.weight": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.38.self_attn.q_proj.weight": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.38.self_attn.q_proj.scales": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.38.self_attn.v_proj.weight": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.38.self_attn.v_proj.scales": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.39.input_layernorm.weight": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.39.layer_scalar": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.39.mlp.down_proj.weight": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.39.mlp.down_proj.scales": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.39.mlp.gate_proj.weight": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.39.mlp.gate_proj.scales": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.39.mlp.up_proj.weight": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.39.mlp.up_proj.scales": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.39.post_attention_layernorm.weight": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.39.post_feedforward_layernorm.weight": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.39.pre_feedforward_layernorm.weight": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.39.self_attn.k_norm.weight": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.39.self_attn.k_proj.weight": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.39.self_attn.k_proj.scales": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.39.self_attn.o_proj.weight": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.39.self_attn.o_proj.scales": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.39.self_attn.q_norm.weight": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.39.self_attn.q_proj.weight": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.39.self_attn.q_proj.scales": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.39.self_attn.v_proj.weight": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.39.self_attn.v_proj.scales": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.4.input_layernorm.weight": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.4.layer_scalar": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.4.mlp.down_proj.weight": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.4.mlp.down_proj.scales": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.4.mlp.gate_proj.weight": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.4.mlp.gate_proj.scales": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.4.mlp.up_proj.weight": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.4.mlp.up_proj.scales": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.4.post_attention_layernorm.weight": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.4.post_feedforward_layernorm.weight": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.4.pre_feedforward_layernorm.weight": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.4.self_attn.k_norm.weight": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.4.self_attn.k_proj.weight": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.4.self_attn.k_proj.scales": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.4.self_attn.o_proj.weight": "model-00009-of-00013.safetensors",
+    "language_model.model.layers.4.self_attn.o_proj.scales": "model-00010-of-00013.safetensors",
+    "language_model.model.layers.4.self_attn.q_norm.weight": "model-00010-of-00013.safetensors",
+    "language_model.model.layers.4.self_attn.q_proj.weight": "model-00010-of-00013.safetensors",
+    "language_model.model.layers.4.self_attn.q_proj.scales": "model-00010-of-00013.safetensors",
+    "language_model.model.layers.4.self_attn.v_proj.weight": "model-00010-of-00013.safetensors",
+    "language_model.model.layers.4.self_attn.v_proj.scales": "model-00010-of-00013.safetensors",
+    "language_model.model.layers.40.input_layernorm.weight": "model-00010-of-00013.safetensors",
+    "language_model.model.layers.40.layer_scalar": "model-00010-of-00013.safetensors",
+    "language_model.model.layers.40.mlp.down_proj.weight": "model-00010-of-00013.safetensors",
+    "language_model.model.layers.40.mlp.down_proj.scales": "model-00010-of-00013.safetensors",
+    "language_model.model.layers.40.mlp.gate_proj.weight": "model-00010-of-00013.safetensors",
+    "language_model.model.layers.40.mlp.gate_proj.scales": "model-00010-of-00013.safetensors",
+    "language_model.model.layers.40.mlp.up_proj.weight": "model-00010-of-00013.safetensors",
+    "language_model.model.layers.40.mlp.up_proj.scales": "model-00010-of-00013.safetensors",
+    "language_model.model.layers.40.post_attention_layernorm.weight": "model-00010-of-00013.safetensors",
+    "language_model.model.layers.40.post_feedforward_layernorm.weight": "model-00010-of-00013.safetensors",
+    "language_model.model.layers.40.pre_feedforward_layernorm.weight": "model-00010-of-00013.safetensors",
+    "language_model.model.layers.40.self_attn.k_norm.weight": "model-00010-of-00013.safetensors",
+    "language_model.model.layers.40.self_attn.k_proj.weight": "model-00010-of-00013.safetensors",
+    "language_model.model.layers.40.self_attn.k_proj.scales": "model-00010-of-00013.safetensors",
+    "language_model.model.layers.40.self_attn.o_proj.weight": "model-00010-of-00013.safetensors",
+    "language_model.model.layers.40.self_attn.o_proj.scales": "model-00010-of-00013.safetensors",
+    "language_model.model.layers.40.self_attn.q_norm.weight": "model-00010-of-00013.safetensors",
+    "language_model.model.layers.40.self_attn.q_proj.weight": "model-00010-of-00013.safetensors",
+    "language_model.model.layers.40.self_attn.q_proj.scales": "model-00010-of-00013.safetensors",
+    "language_model.model.layers.40.self_attn.v_proj.weight": "model-00010-of-00013.safetensors",
+    "language_model.model.layers.40.self_attn.v_proj.scales": "model-00010-of-00013.safetensors",
+    "language_model.model.layers.41.input_layernorm.weight": "model-00010-of-00013.safetensors",
+    "language_model.model.layers.41.layer_scalar": "model-00010-of-00013.safetensors",
+    "language_model.model.layers.41.mlp.down_proj.weight": "model-00010-of-00013.safetensors",
+    "language_model.model.layers.41.mlp.down_proj.scales": "model-00010-of-00013.safetensors",
+    "language_model.model.layers.41.mlp.gate_proj.weight": "model-00010-of-00013.safetensors",
+    "language_model.model.layers.41.mlp.gate_proj.scales": "model-00010-of-00013.safetensors",
+    "language_model.model.layers.41.mlp.up_proj.weight": "model-00010-of-00013.safetensors",
+    "language_model.model.layers.41.mlp.up_proj.scales": "model-00010-of-00013.safetensors",
+    "language_model.model.layers.41.post_attention_layernorm.weight": "model-00010-of-00013.safetensors",
+    "language_model.model.layers.41.post_feedforward_layernorm.weight": "model-00010-of-00013.safetensors",
+    "language_model.model.layers.41.pre_feedforward_layernorm.weight": "model-00010-of-00013.safetensors",
+    "language_model.model.layers.41.self_attn.k_norm.weight": "model-00010-of-00013.safetensors",
+    "language_model.model.layers.41.self_attn.k_proj.weight": "model-00010-of-00013.safetensors",
+    "language_model.model.layers.41.self_attn.k_proj.scales": "model-00010-of-00013.safetensors",
+    "language_model.model.layers.41.self_attn.o_proj.weight": "model-00010-of-00013.safetensors",
+    "language_model.model.layers.41.self_attn.o_proj.scales": "model-00010-of-00013.safetensors",
+    "language_model.model.layers.41.self_attn.q_norm.weight": "model-00010-of-00013.safetensors",
+    "language_model.model.layers.41.self_attn.q_proj.weight": "model-00010-of-00013.safetensors",
+    "language_model.model.layers.41.self_attn.q_proj.scales": "model-00010-of-00013.safetensors",
+    "language_model.model.layers.42.input_layernorm.weight": "model-00010-of-00013.safetensors",
+    "language_model.model.layers.42.layer_scalar": "model-00010-of-00013.safetensors",
+    "language_model.model.layers.42.mlp.down_proj.weight": "model-00010-of-00013.safetensors",
+    "language_model.model.layers.42.mlp.down_proj.scales": "model-00010-of-00013.safetensors",
+    "language_model.model.layers.42.mlp.gate_proj.weight": "model-00010-of-00013.safetensors",
+    "language_model.model.layers.42.mlp.gate_proj.scales": "model-00010-of-00013.safetensors",
+    "language_model.model.layers.42.mlp.up_proj.weight": "model-00010-of-00013.safetensors",
+    "language_model.model.layers.42.mlp.up_proj.scales": "model-00010-of-00013.safetensors",
+    "language_model.model.layers.42.post_attention_layernorm.weight": "model-00010-of-00013.safetensors",
+    "language_model.model.layers.42.post_feedforward_layernorm.weight": "model-00010-of-00013.safetensors",
+    "language_model.model.layers.42.pre_feedforward_layernorm.weight": "model-00010-of-00013.safetensors",
+    "language_model.model.layers.42.self_attn.k_norm.weight": "model-00010-of-00013.safetensors",
+    "language_model.model.layers.42.self_attn.k_proj.weight": "model-00010-of-00013.safetensors",
+    "language_model.model.layers.42.self_attn.k_proj.scales": "model-00010-of-00013.safetensors",
+    "language_model.model.layers.42.self_attn.o_proj.weight": "model-00010-of-00013.safetensors",
+    "language_model.model.layers.42.self_attn.o_proj.scales": "model-00010-of-00013.safetensors",
+    "language_model.model.layers.42.self_attn.q_norm.weight": "model-00010-of-00013.safetensors",
+    "language_model.model.layers.42.self_attn.q_proj.weight": "model-00010-of-00013.safetensors",
+    "language_model.model.layers.42.self_attn.q_proj.scales": "model-00010-of-00013.safetensors",
+    "language_model.model.layers.42.self_attn.v_proj.weight": "model-00010-of-00013.safetensors",
+    "language_model.model.layers.42.self_attn.v_proj.scales": "model-00010-of-00013.safetensors",
+    "language_model.model.layers.43.input_layernorm.weight": "model-00010-of-00013.safetensors",
+    "language_model.model.layers.43.layer_scalar": "model-00010-of-00013.safetensors",
+    "language_model.model.layers.43.mlp.down_proj.weight": "model-00010-of-00013.safetensors",
+    "language_model.model.layers.43.mlp.down_proj.scales": "model-00010-of-00013.safetensors",
+    "language_model.model.layers.43.mlp.gate_proj.weight": "model-00010-of-00013.safetensors",
+    "language_model.model.layers.43.mlp.gate_proj.scales": "model-00010-of-00013.safetensors",
+    "language_model.model.layers.43.mlp.up_proj.weight": "model-00010-of-00013.safetensors",
+    "language_model.model.layers.43.mlp.up_proj.scales": "model-00010-of-00013.safetensors",
+    "language_model.model.layers.43.post_attention_layernorm.weight": "model-00010-of-00013.safetensors",
+    "language_model.model.layers.43.post_feedforward_layernorm.weight": "model-00010-of-00013.safetensors",
+    "language_model.model.layers.43.pre_feedforward_layernorm.weight": "model-00010-of-00013.safetensors",
+    "language_model.model.layers.43.self_attn.k_norm.weight": "model-00010-of-00013.safetensors",
+    "language_model.model.layers.43.self_attn.k_proj.weight": "model-00010-of-00013.safetensors",
+    "language_model.model.layers.43.self_attn.k_proj.scales": "model-00010-of-00013.safetensors",
+    "language_model.model.layers.43.self_attn.o_proj.weight": "model-00010-of-00013.safetensors",
+    "language_model.model.layers.43.self_attn.o_proj.scales": "model-00010-of-00013.safetensors",
+    "language_model.model.layers.43.self_attn.q_norm.weight": "model-00010-of-00013.safetensors",
+    "language_model.model.layers.43.self_attn.q_proj.weight": "model-00010-of-00013.safetensors",
+    "language_model.model.layers.43.self_attn.q_proj.scales": "model-00010-of-00013.safetensors",
+    "language_model.model.layers.43.self_attn.v_proj.weight": "model-00010-of-00013.safetensors",
+    "language_model.model.layers.43.self_attn.v_proj.scales": "model-00010-of-00013.safetensors",
+    "language_model.model.layers.44.input_layernorm.weight": "model-00010-of-00013.safetensors",
+    "language_model.model.layers.44.layer_scalar": "model-00010-of-00013.safetensors",
+    "language_model.model.layers.44.mlp.down_proj.weight": "model-00010-of-00013.safetensors",
+    "language_model.model.layers.44.mlp.down_proj.scales": "model-00011-of-00013.safetensors",
+    "language_model.model.layers.44.mlp.gate_proj.weight": "model-00011-of-00013.safetensors",
+    "language_model.model.layers.44.mlp.gate_proj.scales": "model-00011-of-00013.safetensors",
+    "language_model.model.layers.44.mlp.up_proj.weight": "model-00011-of-00013.safetensors",
+    "language_model.model.layers.44.mlp.up_proj.scales": "model-00011-of-00013.safetensors",
+    "language_model.model.layers.44.post_attention_layernorm.weight": "model-00011-of-00013.safetensors",
+    "language_model.model.layers.44.post_feedforward_layernorm.weight": "model-00011-of-00013.safetensors",
+    "language_model.model.layers.44.pre_feedforward_layernorm.weight": "model-00011-of-00013.safetensors",
+    "language_model.model.layers.44.self_attn.k_norm.weight": "model-00011-of-00013.safetensors",
+    "language_model.model.layers.44.self_attn.k_proj.weight": "model-00011-of-00013.safetensors",
+    "language_model.model.layers.44.self_attn.k_proj.scales": "model-00011-of-00013.safetensors",
+    "language_model.model.layers.44.self_attn.o_proj.weight": "model-00011-of-00013.safetensors",
+    "language_model.model.layers.44.self_attn.o_proj.scales": "model-00011-of-00013.safetensors",
+    "language_model.model.layers.44.self_attn.q_norm.weight": "model-00011-of-00013.safetensors",
+    "language_model.model.layers.44.self_attn.q_proj.weight": "model-00011-of-00013.safetensors",
+    "language_model.model.layers.44.self_attn.q_proj.scales": "model-00011-of-00013.safetensors",
+    "language_model.model.layers.44.self_attn.v_proj.weight": "model-00011-of-00013.safetensors",
+    "language_model.model.layers.44.self_attn.v_proj.scales": "model-00011-of-00013.safetensors",
+    "language_model.model.layers.45.input_layernorm.weight": "model-00011-of-00013.safetensors",
+    "language_model.model.layers.45.layer_scalar": "model-00011-of-00013.safetensors",
+    "language_model.model.layers.45.mlp.down_proj.weight": "model-00011-of-00013.safetensors",
+    "language_model.model.layers.45.mlp.down_proj.scales": "model-00011-of-00013.safetensors",
+    "language_model.model.layers.45.mlp.gate_proj.weight": "model-00011-of-00013.safetensors",
+    "language_model.model.layers.45.mlp.gate_proj.scales": "model-00011-of-00013.safetensors",
+    "language_model.model.layers.45.mlp.up_proj.weight": "model-00011-of-00013.safetensors",
+    "language_model.model.layers.45.mlp.up_proj.scales": "model-00011-of-00013.safetensors",
+    "language_model.model.layers.45.post_attention_layernorm.weight": "model-00011-of-00013.safetensors",
+    "language_model.model.layers.45.post_feedforward_layernorm.weight": "model-00011-of-00013.safetensors",
+    "language_model.model.layers.45.pre_feedforward_layernorm.weight": "model-00011-of-00013.safetensors",
+    "language_model.model.layers.45.self_attn.k_norm.weight": "model-00011-of-00013.safetensors",
+    "language_model.model.layers.45.self_attn.k_proj.weight": "model-00011-of-00013.safetensors",
+    "language_model.model.layers.45.self_attn.k_proj.scales": "model-00011-of-00013.safetensors",
+    "language_model.model.layers.45.self_attn.o_proj.weight": "model-00011-of-00013.safetensors",
+    "language_model.model.layers.45.self_attn.o_proj.scales": "model-00011-of-00013.safetensors",
+    "language_model.model.layers.45.self_attn.q_norm.weight": "model-00011-of-00013.safetensors",
+    "language_model.model.layers.45.self_attn.q_proj.weight": "model-00011-of-00013.safetensors",
+    "language_model.model.layers.45.self_attn.q_proj.scales": "model-00011-of-00013.safetensors",
+    "language_model.model.layers.45.self_attn.v_proj.weight": "model-00011-of-00013.safetensors",
+    "language_model.model.layers.45.self_attn.v_proj.scales": "model-00011-of-00013.safetensors",
+    "language_model.model.layers.46.input_layernorm.weight": "model-00011-of-00013.safetensors",
+    "language_model.model.layers.46.layer_scalar": "model-00011-of-00013.safetensors",
+    "language_model.model.layers.46.mlp.down_proj.weight": "model-00011-of-00013.safetensors",
+    "language_model.model.layers.46.mlp.down_proj.scales": "model-00011-of-00013.safetensors",
+    "language_model.model.layers.46.mlp.gate_proj.weight": "model-00011-of-00013.safetensors",
+    "language_model.model.layers.46.mlp.gate_proj.scales": "model-00011-of-00013.safetensors",
+    "language_model.model.layers.46.mlp.up_proj.weight": "model-00011-of-00013.safetensors",
+    "language_model.model.layers.46.mlp.up_proj.scales": "model-00011-of-00013.safetensors",
+    "language_model.model.layers.46.post_attention_layernorm.weight": "model-00011-of-00013.safetensors",
+    "language_model.model.layers.46.post_feedforward_layernorm.weight": "model-00011-of-00013.safetensors",
+    "language_model.model.layers.46.pre_feedforward_layernorm.weight": "model-00011-of-00013.safetensors",
+    "language_model.model.layers.46.self_attn.k_norm.weight": "model-00011-of-00013.safetensors",
+    "language_model.model.layers.46.self_attn.k_proj.weight": "model-00011-of-00013.safetensors",
+    "language_model.model.layers.46.self_attn.k_proj.scales": "model-00011-of-00013.safetensors",
+    "language_model.model.layers.46.self_attn.o_proj.weight": "model-00011-of-00013.safetensors",
+    "language_model.model.layers.46.self_attn.o_proj.scales": "model-00011-of-00013.safetensors",
+    "language_model.model.layers.46.self_attn.q_norm.weight": "model-00011-of-00013.safetensors",
+    "language_model.model.layers.46.self_attn.q_proj.weight": "model-00011-of-00013.safetensors",
+    "language_model.model.layers.46.self_attn.q_proj.scales": "model-00011-of-00013.safetensors",
+    "language_model.model.layers.46.self_attn.v_proj.weight": "model-00011-of-00013.safetensors",
+    "language_model.model.layers.46.self_attn.v_proj.scales": "model-00011-of-00013.safetensors",
+    "language_model.model.layers.47.input_layernorm.weight": "model-00011-of-00013.safetensors",
+    "language_model.model.layers.47.layer_scalar": "model-00011-of-00013.safetensors",
+    "language_model.model.layers.47.mlp.down_proj.weight": "model-00011-of-00013.safetensors",
+    "language_model.model.layers.47.mlp.down_proj.scales": "model-00011-of-00013.safetensors",
+    "language_model.model.layers.47.mlp.gate_proj.weight": "model-00011-of-00013.safetensors",
+    "language_model.model.layers.47.mlp.gate_proj.scales": "model-00011-of-00013.safetensors",
+    "language_model.model.layers.47.mlp.up_proj.weight": "model-00011-of-00013.safetensors",
+    "language_model.model.layers.47.mlp.up_proj.scales": "model-00011-of-00013.safetensors",
+    "language_model.model.layers.47.post_attention_layernorm.weight": "model-00011-of-00013.safetensors",
+    "language_model.model.layers.47.post_feedforward_layernorm.weight": "model-00011-of-00013.safetensors",
+    "language_model.model.layers.47.pre_feedforward_layernorm.weight": "model-00011-of-00013.safetensors",
+    "language_model.model.layers.47.self_attn.k_norm.weight": "model-00011-of-00013.safetensors",
+    "language_model.model.layers.47.self_attn.k_proj.weight": "model-00011-of-00013.safetensors",
+    "language_model.model.layers.47.self_attn.k_proj.scales": "model-00011-of-00013.safetensors",
+    "language_model.model.layers.47.self_attn.o_proj.weight": "model-00011-of-00013.safetensors",
+    "language_model.model.layers.47.self_attn.o_proj.scales": "model-00011-of-00013.safetensors",
+    "language_model.model.layers.47.self_attn.q_norm.weight": "model-00011-of-00013.safetensors",
+    "language_model.model.layers.47.self_attn.q_proj.weight": "model-00011-of-00013.safetensors",
+    "language_model.model.layers.47.self_attn.q_proj.scales": "model-00011-of-00013.safetensors",
+    "language_model.model.layers.5.input_layernorm.weight": "model-00011-of-00013.safetensors",
+    "language_model.model.layers.5.layer_scalar": "model-00011-of-00013.safetensors",
+    "language_model.model.layers.5.mlp.down_proj.weight": "model-00011-of-00013.safetensors",
+    "language_model.model.layers.5.mlp.down_proj.scales": "model-00011-of-00013.safetensors",
+    "language_model.model.layers.5.mlp.gate_proj.weight": "model-00011-of-00013.safetensors",
+    "language_model.model.layers.5.mlp.gate_proj.scales": "model-00012-of-00013.safetensors",
+    "language_model.model.layers.5.mlp.up_proj.weight": "model-00012-of-00013.safetensors",
+    "language_model.model.layers.5.mlp.up_proj.scales": "model-00012-of-00013.safetensors",
+    "language_model.model.layers.5.post_attention_layernorm.weight": "model-00012-of-00013.safetensors",
+    "language_model.model.layers.5.post_feedforward_layernorm.weight": "model-00012-of-00013.safetensors",
+    "language_model.model.layers.5.pre_feedforward_layernorm.weight": "model-00012-of-00013.safetensors",
+    "language_model.model.layers.5.self_attn.k_norm.weight": "model-00012-of-00013.safetensors",
+    "language_model.model.layers.5.self_attn.k_proj.weight": "model-00012-of-00013.safetensors",
+    "language_model.model.layers.5.self_attn.k_proj.scales": "model-00012-of-00013.safetensors",
+    "language_model.model.layers.5.self_attn.o_proj.weight": "model-00012-of-00013.safetensors",
+    "language_model.model.layers.5.self_attn.o_proj.scales": "model-00012-of-00013.safetensors",
+    "language_model.model.layers.5.self_attn.q_norm.weight": "model-00012-of-00013.safetensors",
+    "language_model.model.layers.5.self_attn.q_proj.weight": "model-00012-of-00013.safetensors",
+    "language_model.model.layers.5.self_attn.q_proj.scales": "model-00012-of-00013.safetensors",
+    "language_model.model.layers.6.input_layernorm.weight": "model-00012-of-00013.safetensors",
+    "language_model.model.layers.6.layer_scalar": "model-00012-of-00013.safetensors",
+    "language_model.model.layers.6.mlp.down_proj.weight": "model-00012-of-00013.safetensors",
+    "language_model.model.layers.6.mlp.down_proj.scales": "model-00012-of-00013.safetensors",
+    "language_model.model.layers.6.mlp.gate_proj.weight": "model-00012-of-00013.safetensors",
+    "language_model.model.layers.6.mlp.gate_proj.scales": "model-00012-of-00013.safetensors",
+    "language_model.model.layers.6.mlp.up_proj.weight": "model-00012-of-00013.safetensors",
+    "language_model.model.layers.6.mlp.up_proj.scales": "model-00012-of-00013.safetensors",
+    "language_model.model.layers.6.post_attention_layernorm.weight": "model-00012-of-00013.safetensors",
+    "language_model.model.layers.6.post_feedforward_layernorm.weight": "model-00012-of-00013.safetensors",
+    "language_model.model.layers.6.pre_feedforward_layernorm.weight": "model-00012-of-00013.safetensors",
+    "language_model.model.layers.6.self_attn.k_norm.weight": "model-00012-of-00013.safetensors",
+    "language_model.model.layers.6.self_attn.k_proj.weight": "model-00012-of-00013.safetensors",
+    "language_model.model.layers.6.self_attn.k_proj.scales": "model-00012-of-00013.safetensors",
+    "language_model.model.layers.6.self_attn.o_proj.weight": "model-00012-of-00013.safetensors",
+    "language_model.model.layers.6.self_attn.o_proj.scales": "model-00012-of-00013.safetensors",
+    "language_model.model.layers.6.self_attn.q_norm.weight": "model-00012-of-00013.safetensors",
+    "language_model.model.layers.6.self_attn.q_proj.weight": "model-00012-of-00013.safetensors",
+    "language_model.model.layers.6.self_attn.q_proj.scales": "model-00012-of-00013.safetensors",
+    "language_model.model.layers.6.self_attn.v_proj.weight": "model-00012-of-00013.safetensors",
+    "language_model.model.layers.6.self_attn.v_proj.scales": "model-00012-of-00013.safetensors",
+    "language_model.model.layers.7.input_layernorm.weight": "model-00012-of-00013.safetensors",
+    "language_model.model.layers.7.layer_scalar": "model-00012-of-00013.safetensors",
+    "language_model.model.layers.7.mlp.down_proj.weight": "model-00012-of-00013.safetensors",
+    "language_model.model.layers.7.mlp.down_proj.scales": "model-00012-of-00013.safetensors",
+    "language_model.model.layers.7.mlp.gate_proj.weight": "model-00012-of-00013.safetensors",
+    "language_model.model.layers.7.mlp.gate_proj.scales": "model-00012-of-00013.safetensors",
+    "language_model.model.layers.7.mlp.up_proj.weight": "model-00012-of-00013.safetensors",
+    "language_model.model.layers.7.mlp.up_proj.scales": "model-00012-of-00013.safetensors",
+    "language_model.model.layers.7.post_attention_layernorm.weight": "model-00012-of-00013.safetensors",
+    "language_model.model.layers.7.post_feedforward_layernorm.weight": "model-00012-of-00013.safetensors",
+    "language_model.model.layers.7.pre_feedforward_layernorm.weight": "model-00012-of-00013.safetensors",
+    "language_model.model.layers.7.self_attn.k_norm.weight": "model-00012-of-00013.safetensors",
+    "language_model.model.layers.7.self_attn.k_proj.weight": "model-00012-of-00013.safetensors",
+    "language_model.model.layers.7.self_attn.k_proj.scales": "model-00012-of-00013.safetensors",
+    "language_model.model.layers.7.self_attn.o_proj.weight": "model-00012-of-00013.safetensors",
+    "language_model.model.layers.7.self_attn.o_proj.scales": "model-00012-of-00013.safetensors",
+    "language_model.model.layers.7.self_attn.q_norm.weight": "model-00012-of-00013.safetensors",
+    "language_model.model.layers.7.self_attn.q_proj.weight": "model-00012-of-00013.safetensors",
+    "language_model.model.layers.7.self_attn.q_proj.scales": "model-00012-of-00013.safetensors",
+    "language_model.model.layers.7.self_attn.v_proj.weight": "model-00012-of-00013.safetensors",
+    "language_model.model.layers.7.self_attn.v_proj.scales": "model-00012-of-00013.safetensors",
+    "language_model.model.layers.8.input_layernorm.weight": "model-00012-of-00013.safetensors",
+    "language_model.model.layers.8.layer_scalar": "model-00012-of-00013.safetensors",
+    "language_model.model.layers.8.mlp.down_proj.weight": "model-00012-of-00013.safetensors",
+    "language_model.model.layers.8.mlp.down_proj.scales": "model-00012-of-00013.safetensors",
+    "language_model.model.layers.8.mlp.gate_proj.weight": "model-00012-of-00013.safetensors",
+    "language_model.model.layers.8.mlp.gate_proj.scales": "model-00012-of-00013.safetensors",
+    "language_model.model.layers.8.mlp.up_proj.weight": "model-00012-of-00013.safetensors",
+    "language_model.model.layers.8.mlp.up_proj.scales": "model-00012-of-00013.safetensors",
+    "language_model.model.layers.8.post_attention_layernorm.weight": "model-00012-of-00013.safetensors",
+    "language_model.model.layers.8.post_feedforward_layernorm.weight": "model-00012-of-00013.safetensors",
+    "language_model.model.layers.8.pre_feedforward_layernorm.weight": "model-00012-of-00013.safetensors",
+    "language_model.model.layers.8.self_attn.k_norm.weight": "model-00012-of-00013.safetensors",
+    "language_model.model.layers.8.self_attn.k_proj.weight": "model-00012-of-00013.safetensors",
+    "language_model.model.layers.8.self_attn.k_proj.scales": "model-00012-of-00013.safetensors",
+    "language_model.model.layers.8.self_attn.o_proj.weight": "model-00012-of-00013.safetensors",
+    "language_model.model.layers.8.self_attn.o_proj.scales": "model-00012-of-00013.safetensors",
+    "language_model.model.layers.8.self_attn.q_norm.weight": "model-00012-of-00013.safetensors",
+    "language_model.model.layers.8.self_attn.q_proj.weight": "model-00012-of-00013.safetensors",
+    "language_model.model.layers.8.self_attn.q_proj.scales": "model-00012-of-00013.safetensors",
+    "language_model.model.layers.8.self_attn.v_proj.weight": "model-00012-of-00013.safetensors",
+    "language_model.model.layers.8.self_attn.v_proj.scales": "model-00012-of-00013.safetensors",
+    "language_model.model.layers.9.input_layernorm.weight": "model-00012-of-00013.safetensors",
+    "language_model.model.layers.9.layer_scalar": "model-00012-of-00013.safetensors",
+    "language_model.model.layers.9.mlp.down_proj.weight": "model-00012-of-00013.safetensors",
+    "language_model.model.layers.9.mlp.down_proj.scales": "model-00012-of-00013.safetensors",
+    "language_model.model.layers.9.mlp.gate_proj.weight": "model-00012-of-00013.safetensors",
+    "language_model.model.layers.9.mlp.gate_proj.scales": "model-00012-of-00013.safetensors",
+    "language_model.model.layers.9.mlp.up_proj.weight": "model-00012-of-00013.safetensors",
+    "language_model.model.layers.9.mlp.up_proj.scales": "model-00013-of-00013.safetensors",
+    "language_model.model.layers.9.post_attention_layernorm.weight": "model-00013-of-00013.safetensors",
+    "language_model.model.layers.9.post_feedforward_layernorm.weight": "model-00013-of-00013.safetensors",
+    "language_model.model.layers.9.pre_feedforward_layernorm.weight": "model-00013-of-00013.safetensors",
+    "language_model.model.layers.9.self_attn.k_norm.weight": "model-00013-of-00013.safetensors",
+    "language_model.model.layers.9.self_attn.k_proj.weight": "model-00013-of-00013.safetensors",
+    "language_model.model.layers.9.self_attn.k_proj.scales": "model-00013-of-00013.safetensors",
+    "language_model.model.layers.9.self_attn.o_proj.weight": "model-00013-of-00013.safetensors",
+    "language_model.model.layers.9.self_attn.o_proj.scales": "model-00013-of-00013.safetensors",
+    "language_model.model.layers.9.self_attn.q_norm.weight": "model-00013-of-00013.safetensors",
+    "language_model.model.layers.9.self_attn.q_proj.weight": "model-00013-of-00013.safetensors",
+    "language_model.model.layers.9.self_attn.q_proj.scales": "model-00013-of-00013.safetensors",
+    "language_model.model.layers.9.self_attn.v_proj.weight": "model-00013-of-00013.safetensors",
+    "language_model.model.layers.9.self_attn.v_proj.scales": "model-00013-of-00013.safetensors",
+    "language_model.model.norm.weight": "model-00013-of-00013.safetensors",
+    "vision_embedder.patch_dense.bias": "model-00013-of-00013.safetensors",
+    "vision_embedder.patch_dense.weight": "model-00013-of-00013.safetensors",
+    "vision_embedder.patch_ln1.bias": "model-00013-of-00013.safetensors",
+    "vision_embedder.patch_ln1.weight": "model-00013-of-00013.safetensors",
+    "vision_embedder.patch_ln2.bias": "model-00013-of-00013.safetensors",
+    "vision_embedder.patch_ln2.weight": "model-00013-of-00013.safetensors",
+    "vision_embedder.pos_embedding": "model-00013-of-00013.safetensors",
+    "vision_embedder.pos_norm.bias": "model-00013-of-00013.safetensors",
+    "vision_embedder.pos_norm.weight": "model-00013-of-00013.safetensors"
+  }
+}

processor_config.json

@@ -0,0 +1,62 @@
+{
+  "audio_ms_per_token": 40,
+  "audio_seq_length": 750,
+  "feature_extractor": {
+    "audio_samples_per_token": 640,
+    "feature_extractor_type": "Gemma4UnifiedAudioFeatureExtractor",
+    "feature_size": 640,
+    "padding_side": "right",
+    "padding_value": 0.0,
+    "return_attention_mask": true,
+    "sampling_rate": 16000
+  },
+  "image_processor": {
+    "do_convert_rgb": true,
+    "do_normalize": false,
+    "do_rescale": true,
+    "do_resize": true,
+    "image_mean": [
+      0.0,
+      0.0,
+      0.0
+    ],
+    "image_processor_type": "Gemma4UnifiedImageProcessor",
+    "image_std": [
+      1.0,
+      1.0,
+      1.0
+    ],
+    "max_soft_tokens": 280,
+    "patch_size": 16,
+    "pooling_kernel_size": 3,
+    "resample": 3,
+    "rescale_factor": 0.00392156862745098
+  },
+  "image_seq_length": 280,
+  "processor_class": "Gemma4UnifiedProcessor",
+  "video_processor": {
+    "do_convert_rgb": true,
+    "do_normalize": true,
+    "do_rescale": true,
+    "do_resize": true,
+    "do_sample_frames": true,
+    "image_mean": [
+      0.0,
+      0.0,
+      0.0
+    ],
+    "image_std": [
+      1.0,
+      1.0,
+      1.0
+    ],
+    "max_soft_tokens": 70,
+    "num_frames": 32,
+    "patch_size": 16,
+    "pooling_kernel_size": 3,
+    "resample": 3,
+    "rescale_factor": 0.00392156862745098,
+    "return_metadata": false,
+    "video_processor_type": "Gemma4UnifiedVideoProcessor"
+  }
+}

tokenizer.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cc8d3a0ce36466ccc1278bf987df5f71db1719b9ca6b4118264f45cb627bfe0f
+size 32169626

tokenizer_config.json

@@ -0,0 +1,75 @@
+{
+  "audio_token": "<|audio|>",
+  "backend": "tokenizers",
+  "boa_token": "<|audio>",
+  "boi_token": "<|image>",
+  "bos_token": "<bos>",
+  "eoa_token": "<audio|>",
+  "eoc_token": "<channel|>",
+  "eoi_token": "<image|>",
+  "eos_token": "<eos>",
+  "eot_token": "<turn|>",
+  "escape_token": "<|\"|>",
+  "etc_token": "<tool_call|>",
+  "etd_token": "<tool|>",
+  "etr_token": "<tool_response|>",
+  "extra_special_tokens": [
+    "<|video|>"
+  ],
+  "image_token": "<|image|>",
+  "mask_token": "<mask>",
+  "model_max_length": 1000000000000000019884624838656,
+  "pad_token": "<pad>",
+  "padding_side": "left",
+  "processor_class": "Gemma4UnifiedProcessor",
+  "response_schema": {
+    "properties": {
+      "content": {
+        "type": "string"
+      },
+      "role": {
+        "const": "assistant"
+      },
+      "thinking": {
+        "type": "string"
+      },
+      "tool_calls": {
+        "items": {
+          "properties": {
+            "function": {
+              "properties": {
+                "arguments": {
+                  "additionalProperties": {},
+                  "type": "object",
+                  "x-parser": "gemma4-tool-call"
+                },
+                "name": {
+                  "type": "string"
+                }
+              },
+              "type": "object",
+              "x-regex": "call\\:(?P<name>\\w+)(?P<arguments>\\{.*\\})"
+            },
+            "type": {
+              "const": "function"
+            }
+          },
+          "type": "object"
+        },
+        "type": "array",
+        "x-regex-iterator": "<\\|tool_call>(.*?)<tool_call\\|>"
+      }
+    },
+    "type": "object",
+    "x-regex": "(\\<\\|channel\\>thought\\n(?P<thinking>.*?)\\<channel\\|\\>)?(?P<tool_calls>\\<\\|tool_call\\>.*\\<tool_call\\|\\>)?(?P<content>(?:(?!\\<turn\\|\\>)(?!\\<\\|tool_response\\>).)+)?(?:\\<turn\\|\\>|\\<\\|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>",
+  "chat_template": "{%- macro format_parameters(properties, required, filter_keys=false) -%}\n    {%- set standard_keys = ['description', 'type', 'properties', 'required', 'nullable'] -%}\n    {%- set ns = namespace(found_first=false) -%}\n    {%- for key, value in properties | dictsort -%}\n        {%- set add_comma = false -%}\n        {%- if not filter_keys or key not in standard_keys -%}\n            {%- if ns.found_first %},{% endif -%}\n            {%- set ns.found_first = true -%}\n            {{ key }}:{\n            {%- if value['description'] -%}\n                description:<|\"|>{{ value['description'] }}<|\"|>\n                {%- set add_comma = true -%}\n            {%- endif -%}\n            {%- if value['type'] | upper == 'STRING' -%}\n                {%- if value['enum'] -%}\n                    {%- if add_comma %},{%- else -%} {%- set add_comma = true -%} {% endif -%}\n                    enum:{{ format_argument(value['enum']) }}\n                {%- endif -%}\n            {%- elif value['type'] | upper == 'ARRAY' -%}\n                {%- if value['items'] is mapping and value['items'] -%}\n                    {%- if add_comma %},{%- else -%} {%- set add_comma = true -%} {% endif -%}\n                    items:{\n                    {%- set ns_items = namespace(found_first=false) -%}\n                    {%- for item_key, item_value in value['items'] | dictsort -%}\n                        {%- if item_value is not none -%}\n                            {%- if ns_items.found_first %},{% endif -%}\n                            {%- set ns_items.found_first = true -%}\n                            {%- if item_key == 'properties' -%}\n                                properties:{\n                                {%- if item_value is mapping -%}\n                                    {{- format_parameters(item_value, value['items']['required'] | default([])) -}}\n                                {%- endif -%}\n                                }\n                            {%- elif item_key == 'required' -%}\n                                required:[\n                                {%- for req_item in item_value -%}\n                                    <|\"|>{{- req_item -}}<|\"|>\n                                    {%- if not loop.last %},{% endif -%}\n                                {%- endfor -%}\n                                ]\n                            {%- elif item_key == 'type' -%}\n                                {%- if item_value is string -%}\n                                    type:{{ format_argument(item_value | upper) }}\n                                {%- else -%}\n                                    type:{{ format_argument(item_value | map('upper') | list) }}\n                                {%- endif -%}\n                            {%- else -%}\n                                {{ item_key }}:{{ format_argument(item_value) }}\n                            {%- endif -%}\n                        {%- endif -%}\n                    {%- endfor -%}\n                    }\n                {%- endif -%}\n            {%- endif -%}\n            {%- if value['nullable'] %}\n                {%- if add_comma %},{%- else -%} {%- set add_comma = true -%} {% endif -%}\n                nullable:true\n            {%- endif -%}\n            {%- if value['type'] | upper == 'OBJECT' -%}\n                {%- if value['properties'] is defined and value['properties'] is mapping -%}\n                    {%- if add_comma %},{%- else -%} {%- set add_comma = true -%} {% endif -%}\n                    properties:{\n                    {{- format_parameters(value['properties'], value['required'] | default([])) -}}\n                    }\n                {%- elif value is mapping -%}\n                    {%- if add_comma %},{%- else -%} {%- set add_comma = true -%} {% endif -%}\n                    properties:{\n                    {{- format_parameters(value, value['required'] | default([]), filter_keys=true) -}}\n                    }\n                {%- endif -%}\n                {%- if value['required'] -%}\n                    {%- if add_comma %},{%- else -%} {%- set add_comma = true -%} {% endif -%}\n                    required:[\n                    {%- for item in value['required'] | default([]) -%}\n                        <|\"|>{{- item -}}<|\"|>\n                        {%- if not loop.last %},{% endif -%}\n                    {%- endfor -%}\n                    ]\n                {%- endif -%}\n            {%- endif -%}\n            {%- if add_comma %},{%- else -%} {%- set add_comma = true -%} {% endif -%}\n            type:<|\"|>{{ value['type'] | upper }}<|\"|>}\n        {%- endif -%}\n    {%- endfor -%}\n{%- endmacro -%}\n{%- macro format_function_declaration(tool_data) -%}\n    declaration:{{- tool_data['function']['name'] -}}{description:<|\"|>{{- tool_data['function']['description'] -}}<|\"|>\n    {%- set params = tool_data['function']['parameters'] -%}\n    {%- if params -%}\n        ,parameters:{\n        {%- if params['properties'] -%}\n            properties:{ {{- format_parameters(params['properties'], params['required']) -}} },\n        {%- endif -%}\n        {%- if params['required'] -%}\n            required:[\n            {%- for item in params['required'] -%}\n                <|\"|>{{- item -}}<|\"|>\n                {{- ',' if not loop.last -}}\n            {%- endfor -%}\n            ],\n        {%- endif -%}\n        {%- if params['type'] -%}\n            type:<|\"|>{{- params['type'] | upper -}}<|\"|>}\n        {%- endif -%}\n    {%- endif -%}\n    {%- if 'response' in tool_data['function'] -%}\n        {%- set response_declaration = tool_data['function']['response'] -%}\n        ,response:{\n        {%- if response_declaration['description'] -%}\n            description:<|\"|>{{- response_declaration['description'] -}}<|\"|>,\n        {%- endif -%}\n        {%- if response_declaration['type'] | upper == 'OBJECT' -%}\n            type:<|\"|>{{- response_declaration['type'] | upper -}}<|\"|>}\n        {%- endif -%}\n    {%- endif -%}\n    }\n{%- endmacro -%}\n{%- macro format_argument(argument, escape_keys=True) -%}\n    {%- if argument is string -%}\n        {{- '<|\"|>' + argument + '<|\"|>' -}}\n    {%- elif argument is boolean -%}\n        {{- 'true' if argument else 'false' -}}\n    {%- elif argument is mapping -%}\n        {{- '{' -}}\n        {%- set ns = namespace(found_first=false) -%}\n        {%- for key, value in argument | dictsort -%}\n            {%- if ns.found_first %},{% endif -%}\n            {%- set ns.found_first = true -%}\n            {%- if escape_keys -%}\n                {{- '<|\"|>' + key + '<|\"|>' -}}\n            {%- else -%}\n                {{- key -}}\n            {%- endif -%}\n            :{{- format_argument(value, escape_keys=escape_keys) -}}\n        {%- endfor -%}\n        {{- '}' -}}\n    {%- elif argument is sequence -%}\n        {{- '[' -}}\n        {%- for item in argument -%}\n            {{- format_argument(item, escape_keys=escape_keys) -}}\n            {%- if not loop.last %},{% endif -%}\n        {%- endfor -%}\n        {{- ']' -}}\n    {%- else -%}\n        {{- argument -}}\n    {%- endif -%}\n{%- endmacro -%}\n{%- macro strip_thinking(text) -%}\n    {%- set ns = namespace(result='') -%}\n    {%- for part in text.split('<channel|>') -%}\n        {%- if '<|channel>' in part -%}\n            {%- set ns.result = ns.result + part.split('<|channel>')[0] -%}\n        {%- else -%}\n            {%- set ns.result = ns.result + part -%}\n        {%- endif -%}\n    {%- endfor -%}\n    {{- ns.result | trim -}}\n{%- endmacro -%}\n\n{%- macro format_tool_response_block(tool_name, response) -%}\n    {{- '<|tool_response>' -}}\n    {%- if response is mapping -%}\n        {{- 'response:' + tool_name + '{' -}}\n        {%- for key, value in response | dictsort -%}\n            {{- key -}}:{{- format_argument(value, escape_keys=False) -}}\n            {%- if not loop.last %},{% endif -%}\n        {%- endfor -%}\n        {{- '}' -}}\n    {%- else -%}\n        {{- 'response:' + tool_name + '{value:' + format_argument(response, escape_keys=False) + '}' -}}\n    {%- endif -%}\n    {{- '<tool_response|>' -}}\n{%- endmacro -%}\n\n{%- set ns = namespace(prev_message_type=None) -%}\n{%- set loop_messages = messages -%}\n{{- bos_token -}}\n{#- Handle System/Tool Definitions Block -#}\n{%- if (enable_thinking is defined and enable_thinking) or tools or messages[0]['role'] in ['system', 'developer'] -%}\n    {{- '<|turn>system\\n' -}}\n    {#- Inject Thinking token at the very top of the FIRST system turn -#}\n    {%- if enable_thinking is defined and enable_thinking -%}\n        {{- '<|think|>\\n' -}}\n        {%- set ns.prev_message_type = 'think' -%}\n    {%- endif -%}\n    {%- if messages[0]['role'] in ['system', 'developer'] -%}\n        {%- if messages[0]['content'] is string -%}\n            {{- messages[0]['content'] | trim -}}\n        {%- elif messages[0]['content'] is sequence -%}\n            {%- for item in messages[0]['content'] -%}\n                {{- item['text'] | trim + ' '-}}\n            {%- endfor -%}\n        {%- endif -%}\n        {%- set loop_messages = messages[1:] -%}\n    {%- endif -%}\n    {%- if tools -%}\n        {%- for tool in tools %}\n            {{- '<|tool>' -}}\n            {{- format_function_declaration(tool) | trim -}}\n            {{- '<tool|>' -}}\n        {%- endfor %}\n        {%- set ns.prev_message_type = 'tool' -%}\n        {%- if tool_choice is defined and ((tool_choice is string and tool_choice == 'required') or (tool_choice is mapping and tool_choice.get('type') == 'required')) -%}\n            {{- '\\nTool use is REQUIRED for this turn: you must call exactly one of the declared tools. Output only the tool call, and copy every argument value verbatim \\u2014 character for character, preserving punctuation and newlines, with no added or removed whitespace.' -}}\n        {%- endif -%}\n    {%- endif -%}\n    {{- '<turn|>\\n' -}}\n{%- endif %}\n\n{#- Pre-scan: find last user message index for reasoning guard -#}\n{%- set ns_turn = namespace(last_user_idx=-1) -%}\n{%- for i in range(loop_messages | length) -%}\n    {%- if loop_messages[i]['role'] == 'user' -%}\n        {%- set ns_turn.last_user_idx = i -%}\n    {%- endif -%}\n{%- endfor -%}\n\n{#- Loop through messages -#}\n{%- for message in loop_messages -%}\n    {%- if message['role'] != 'tool' -%}\n    {%- set ns.prev_message_type = None -%}\n    {%- set role = 'model' if message['role'] == 'assistant' else message['role'] -%}\n    {#- Detect continuation: suppress duplicate <|turn>model when previous non-tool message was also assistant -#}\n    {%- set prev_nt = namespace(role=None, found=false) -%}\n    {%- if loop.index0 > 0 -%}\n        {%- for j in range(loop.index0 - 1, -1, -1) -%}\n            {%- if not prev_nt.found -%}\n                {%- if loop_messages[j]['role'] != 'tool' -%}\n                    {%- set prev_nt.role = loop_messages[j]['role'] -%}\n                    {%- set prev_nt.found = true -%}\n                {%- endif -%}\n            {%- endif -%}\n        {%- endfor -%}\n    {%- endif -%}\n    {%- set continue_same_model_turn = (role == 'model' and prev_nt.role == 'assistant') -%}\n    {%- if not continue_same_model_turn -%}\n        {{- '<|turn>' + role + '\\n' }}\n    {%- endif -%}\n\n    {#- Render reasoning/reasoning_content as thinking channel -#}\n    {%- set thinking_text = message.get('reasoning') or message.get('reasoning_content') -%}\n    {%- if thinking_text and loop.index0 > ns_turn.last_user_idx and message.get('tool_calls') -%}\n        {{- '<|channel>thought\\n' + thinking_text + '\\n<channel|>' -}}\n    {%- endif -%}\n\n            {%- if message['tool_calls'] -%}\n                {%- for tool_call in message['tool_calls'] -%}\n                    {%- set function = tool_call['function'] -%}\n                    {{- '<|tool_call>call:' + function['name'] + '{' -}}\n                    {%- if function['arguments'] is mapping -%}\n                        {%- set ns_args = namespace(found_first=false) -%}\n                        {%- for key, value in function['arguments'] | dictsort -%}\n                            {%- if ns_args.found_first %},{% endif -%}\n                            {%- set ns_args.found_first = true -%}\n                            {{- key -}}:{{- format_argument(value, escape_keys=False) -}}\n                        {%- endfor -%}\n                    {%- elif function['arguments'] is string -%}\n                        {{- function['arguments'] -}}\n                    {%- endif -%}\n                    {{- '}<tool_call|>' -}}\n                {%- endfor -%}\n                {%- set ns.prev_message_type = 'tool_call' -%}\n            {%- endif -%}\n\n            {%- set ns_tr_out = namespace(flag=false) -%}\n            {%- if message.get('tool_responses') -%}\n                {#- Legacy: tool_responses embedded on the assistant message (Google/Gemma native) -#}\n                {%- for tool_response in message['tool_responses'] -%}\n                    {{- format_tool_response_block(tool_response['name'] | default('unknown'), tool_response['response']) -}}\n                    {%- set ns_tr_out.flag = true -%}\n                    {%- set ns.prev_message_type = 'tool_response' -%}\n                {%- endfor -%}\n            {%- elif message.get('tool_calls') -%}\n                {#- OpenAI Chat Completions: forward-scan consecutive role:tool messages -#}\n                {%- set ns_tool_scan = namespace(stopped=false) -%}\n                {%- for k in range(loop.index0 + 1, loop_messages | length) -%}\n                    {%- if ns_tool_scan.stopped -%}\n                    {%- elif loop_messages[k]['role'] != 'tool' -%}\n                        {%- set ns_tool_scan.stopped = true -%}\n                    {%- else -%}\n                        {%- set follow = loop_messages[k] -%}\n                        {#- Resolve tool_call_id to function name -#}\n                        {%- set ns_tname = namespace(name=follow.get('name') | default('unknown')) -%}\n                        {%- for tc in message['tool_calls'] -%}\n                            {%- if tc.get('id') == follow.get('tool_call_id') -%}\n                                {%- set ns_tname.name = tc['function']['name'] -%}\n                            {%- endif -%}\n                        {%- endfor -%}\n                        {#- Handle content as string or content-parts array -#}\n                        {%- set tool_body = follow.get('content') -%}\n                        {%- if tool_body is string -%}\n                            {{- format_tool_response_block(ns_tname.name, tool_body) -}}\n                        {%- elif tool_body is sequence and tool_body is not string -%}\n                            {%- set ns_txt = namespace(s='') -%}\n                            {%- for part in tool_body -%}\n                                {%- if part.get('type') == 'text' -%}\n                                    {%- set ns_txt.s = ns_txt.s + (part.get('text') | default('')) -%}\n                                {%- endif -%}\n                            {%- endfor -%}\n                            {{- format_tool_response_block(ns_tname.name, ns_txt.s) -}}\n                            {%- for part in tool_body -%}\n                                {%- if part.get('type') == 'image' -%}\n                                    {{- '<|image|>' -}}\n                                {%- elif part.get('type') == 'audio' -%}\n                                    {{- '<|audio|>' -}}\n                                {%- elif part.get('type') == 'video' -%}\n                                    {{- '<|video|>' -}}\n                                {%- endif -%}\n                            {%- endfor -%}\n                        {%- else -%}\n                            {{- format_tool_response_block(ns_tname.name, tool_body) -}}\n                        {%- endif -%}\n                        {%- set ns_tr_out.flag = true -%}\n                        {%- set ns.prev_message_type = 'tool_response' -%}\n                    {%- endif -%}\n                {%- endfor -%}\n            {%- endif -%}\n\n            {%- set captured_content -%}\n            {%- if message['content'] is string -%}\n                {%- if role == 'model' -%}\n                    {{- strip_thinking(message['content']) -}}\n                {%- else -%}\n                    {{- message['content'] | trim -}}\n                {%- endif -%}\n            {%- elif message['content'] is sequence -%}\n                {%- for item in message['content'] -%}\n                    {%- if item['type'] == 'text' -%}\n                        {%- if role == 'model' -%}\n                            {{- strip_thinking(item['text']) -}}\n                        {%- else -%}\n                            {{- item['text'] | trim -}}\n                        {%- endif -%}\n                    {%- elif item['type'] == 'image' -%}\n                        {{- '<|image|>' -}}\n                        {%- set ns.prev_message_type = 'image' -%}\n                    {%- elif item['type'] == 'audio' -%}\n                        {{- '<|audio|>' -}}\n                        {%- set ns.prev_message_type = 'audio' -%}\n                    {%- elif item['type'] == 'video' -%}\n                        {{- '<|video|>' -}}\n                        {%- set ns.prev_message_type = 'video' -%}\n                    {%- endif -%}\n                {%- endfor -%}\n            {%- endif -%}\n            {%- endset -%}\n\n            {{- captured_content -}}\n            {%- set has_content = captured_content | trim | length > 0 -%}\n\n        {%- if ns.prev_message_type == 'tool_call' and not ns_tr_out.flag -%}\n            {{- '<|tool_response>' -}}\n        {%- elif not (ns_tr_out.flag and not has_content) -%}\n            {{- '<turn|>\\n' -}}\n        {%- endif -%}\n    {%- endif -%}\n{%- endfor -%}\n\n{%- if add_generation_prompt -%}\n    {%- if ns.prev_message_type != 'tool_response' and ns.prev_message_type != 'tool_call' -%}\n        {{- '<|turn>model\\n' -}}\n    {%- endif -%}\n{%- endif -%}"
+}