Upload model files (v1)

Unmodified copies of NemoStation/Marlin-2B at fd111fca4fc7897876fb0d7e9df22ca5ac8ab965, plus a model card.

.gitattributes

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

LICENSE

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

NOTICE

@@ -0,0 +1,15 @@
+Marlin 2B
+Copyright (c) 2026 NemoStation
+
+This product includes weights derived from Qwen3.5-2B
+(https://huggingface.co/Qwen/Qwen3.5-2B), Copyright (c) 2025 Alibaba Cloud,
+used under the Apache License, Version 2.0.
+
+Modifications by NemoStation include: integration of a video-capable
+visual tower, custom training data curation (~400K clip-level annotations
+with Gemini-3-Flash teacher distillation), two-stage SFT + SimPO
+post-training, and custom modeling code (modeling_marlin.py) exposing
+the .caption() and .find() inference modes.
+
+Marlin 2B is distributed under the Apache License, Version 2.0
+(see LICENSE).

README.md

@@ -0,0 +1,20 @@
+---
+license: apache-2.0
+tags:
+  - video
+---
+
+# Marlin 2B
+
+Marlin 2B is a 2B-parameter video-language model that describes what happens in a video as timestamped events. It is derived from Qwen3.5-2B with a video-capable visual tower. All files in this repository are unmodified copies of the originals published at `NemoStation/Marlin-2B`, revision `fd111fca4fc7897876fb0d7e9df22ca5ac8ab965`. Only this model card is new; the upstream model card is not included.
+
+## Author
+
+**[NemoStation](https://huggingface.co/NemoStation/Marlin-2B)**
+
+## License
+
+Apache License 2.0. The complete license is distributed with this repository in [`LICENSE`](LICENSE), and the upstream attribution notice in [`NOTICE`](NOTICE). Marlin 2B is derived from [Qwen3.5-2B](https://huggingface.co/Qwen/Qwen3.5-2B) by Alibaba Cloud, also released under the Apache License 2.0.
+
+---
+*Hosted by [deAPI-ai](https://huggingface.co/deAPI-ai) for use with [GamerHash AI](https://gamerhash.com)*

chat_template.jinja

@@ -0,0 +1,154 @@
+{%- set image_count = namespace(value=0) %}
+{%- set video_count = namespace(value=0) %}
+{%- macro render_content(content, do_vision_count, is_system_content=false) %}
+    {%- if content is string %}
+        {{- content }}
+    {%- elif content is iterable and content is not mapping %}
+        {%- for item in content %}
+            {%- if 'image' in item or 'image_url' in item or item.type == 'image' %}
+                {%- if is_system_content %}
+                    {{- raise_exception('System message cannot contain images.') }}
+                {%- endif %}
+                {%- if do_vision_count %}
+                    {%- set image_count.value = image_count.value + 1 %}
+                {%- endif %}
+                {%- if add_vision_id %}
+                    {{- 'Picture ' ~ image_count.value ~ ': ' }}
+                {%- endif %}
+                {{- '<|vision_start|><|image_pad|><|vision_end|>' }}
+            {%- elif 'video' in item or item.type == 'video' %}
+                {%- if is_system_content %}
+                    {{- raise_exception('System message cannot contain videos.') }}
+                {%- endif %}
+                {%- if do_vision_count %}
+                    {%- set video_count.value = video_count.value + 1 %}
+                {%- endif %}
+                {%- if add_vision_id %}
+                    {{- 'Video ' ~ video_count.value ~ ': ' }}
+                {%- endif %}
+                {{- '<|vision_start|><|video_pad|><|vision_end|>' }}
+            {%- elif 'text' in item %}
+                {{- item.text }}
+            {%- else %}
+                {{- raise_exception('Unexpected item type in content.') }}
+            {%- endif %}
+        {%- endfor %}
+    {%- elif content is none or content is undefined %}
+        {{- '' }}
+    {%- else %}
+        {{- raise_exception('Unexpected content type.') }}
+    {%- endif %}
+{%- endmacro %}
+{%- if not messages %}
+    {{- raise_exception('No messages provided.') }}
+{%- endif %}
+{%- if tools and tools is iterable and tools is not mapping %}
+    {{- '<|im_start|>system\n' }}
+    {{- "# Tools\n\nYou have access to the following functions:\n\n<tools>" }}
+    {%- for tool in tools %}
+        {{- "\n" }}
+        {{- tool | tojson }}
+    {%- endfor %}
+    {{- "\n</tools>" }}
+    {{- '\n\nIf you choose to call a function ONLY reply in the following format with NO suffix:\n\n<tool_call>\n<function=example_function_name>\n<parameter=example_parameter_1>\nvalue_1\n</parameter>\n<parameter=example_parameter_2>\nThis is the value for the second parameter\nthat can span\nmultiple lines\n</parameter>\n</function>\n</tool_call>\n\n<IMPORTANT>\nReminder:\n- Function calls MUST follow the specified format: an inner <function=...></function> block must be nested within <tool_call></tool_call> XML tags\n- Required parameters MUST be specified\n- You may provide optional reasoning for your function call in natural language BEFORE the function call, but NOT after\n- If there is no function call available, answer the question like normal with your current knowledge and do not tell the user about function calls\n</IMPORTANT>' }}
+    {%- if messages[0].role == 'system' %}
+        {%- set content = render_content(messages[0].content, false, true)|trim %}
+        {%- if content %}
+            {{- '\n\n' + content }}
+        {%- endif %}
+    {%- endif %}
+    {{- '<|im_end|>\n' }}
+{%- else %}
+    {%- if messages[0].role == 'system' %}
+        {%- set content = render_content(messages[0].content, false, true)|trim %}
+        {{- '<|im_start|>system\n' + content + '<|im_end|>\n' }}
+    {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+    {%- set index = (messages|length - 1) - loop.index0 %}
+    {%- if ns.multi_step_tool and message.role == "user" %}
+        {%- set content = render_content(message.content, false)|trim %}
+        {%- if not(content.startswith('<tool_response>') and content.endswith('</tool_response>')) %}
+            {%- set ns.multi_step_tool = false %}
+            {%- set ns.last_query_index = index %}
+        {%- endif %}
+    {%- endif %}
+{%- endfor %}
+{%- if ns.multi_step_tool %}
+    {{- raise_exception('No user query found in messages.') }}
+{%- endif %}
+{%- for message in messages %}
+    {%- set content = render_content(message.content, true)|trim %}
+    {%- if message.role == "system" %}
+        {%- if not loop.first %}
+            {{- raise_exception('System message must be at the beginning.') }}
+        {%- endif %}
+    {%- elif message.role == "user" %}
+        {{- '<|im_start|>' + message.role + '\n' + content + '<|im_end|>' + '\n' }}
+    {%- elif message.role == "assistant" %}
+        {%- set reasoning_content = '' %}
+        {%- if message.reasoning_content is string %}
+            {%- set reasoning_content = message.reasoning_content %}
+        {%- else %}
+            {%- if '</think>' in content %}
+                {%- set reasoning_content = content.split('</think>')[0].rstrip('\n').split('<think>')[-1].lstrip('\n') %}
+                {%- set content = content.split('</think>')[-1].lstrip('\n') %}
+            {%- endif %}
+        {%- endif %}
+        {%- set reasoning_content = reasoning_content|trim %}
+        {%- if loop.index0 > ns.last_query_index %}
+            {{- '<|im_start|>' + message.role + '\n<think>\n' + reasoning_content + '\n</think>\n\n' + content }}
+        {%- else %}
+            {{- '<|im_start|>' + message.role + '\n' + content }}
+        {%- endif %}
+        {%- if message.tool_calls and message.tool_calls is iterable and message.tool_calls is not mapping %}
+            {%- for tool_call in message.tool_calls %}
+                {%- if tool_call.function is defined %}
+                    {%- set tool_call = tool_call.function %}
+                {%- endif %}
+                {%- if loop.first %}
+                    {%- if content|trim %}
+                        {{- '\n\n<tool_call>\n<function=' + tool_call.name + '>\n' }}
+                    {%- else %}
+                        {{- '<tool_call>\n<function=' + tool_call.name + '>\n' }}
+                    {%- endif %}
+                {%- else %}
+                    {{- '\n<tool_call>\n<function=' + tool_call.name + '>\n' }}
+                {%- endif %}
+                {%- if tool_call.arguments is defined %}
+                    {%- for args_name, args_value in tool_call.arguments|items %}
+                        {{- '<parameter=' + args_name + '>\n' }}
+                        {%- set args_value = args_value | tojson | safe if args_value is mapping or (args_value is sequence and args_value is not string) else args_value | string %}
+                        {{- args_value }}
+                        {{- '\n</parameter>\n' }}
+                    {%- endfor %}
+                {%- endif %}
+                {{- '</function>\n</tool_call>' }}
+            {%- endfor %}
+        {%- endif %}
+        {{- '<|im_end|>\n' }}
+    {%- elif message.role == "tool" %}
+        {%- if loop.previtem and loop.previtem.role != "tool" %}
+            {{- '<|im_start|>user' }}
+        {%- endif %}
+        {{- '\n<tool_response>\n' }}
+        {{- content }}
+        {{- '\n</tool_response>' }}
+        {%- if not loop.last and loop.nextitem.role != "tool" %}
+            {{- '<|im_end|>\n' }}
+        {%- elif loop.last %}
+            {{- '<|im_end|>\n' }}
+        {%- endif %}
+    {%- else %}
+        {{- raise_exception('Unexpected message role.') }}
+    {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+    {{- '<|im_start|>assistant\n' }}
+    {%- if enable_thinking is defined and enable_thinking is true %}
+        {{- '<think>\n' }}
+    {%- else %}
+        {{- '<think>\n\n</think>\n\n' }}
+    {%- endif %}
+{%- endif %}

config.json

@@ -0,0 +1,108 @@
+{
+  "architectures": [
+    "MarlinForConditionalGeneration"
+  ],
+  "dtype": "bfloat16",
+  "eos_token_id": 248046,
+  "hidden_size": 2048,
+  "image_token_id": 248056,
+  "model_type": "qwen3_5",
+  "pad_token_id": 248044,
+  "text_config": {
+    "attention_bias": false,
+    "attention_dropout": 0.0,
+    "attn_output_gate": true,
+    "bos_token_id": null,
+    "dtype": "bfloat16",
+    "eos_token_id": 248044,
+    "full_attention_interval": 4,
+    "head_dim": 256,
+    "hidden_act": "silu",
+    "hidden_size": 2048,
+    "initializer_range": 0.02,
+    "intermediate_size": 6144,
+    "layer_types": [
+      "linear_attention",
+      "linear_attention",
+      "linear_attention",
+      "full_attention",
+      "linear_attention",
+      "linear_attention",
+      "linear_attention",
+      "full_attention",
+      "linear_attention",
+      "linear_attention",
+      "linear_attention",
+      "full_attention",
+      "linear_attention",
+      "linear_attention",
+      "linear_attention",
+      "full_attention",
+      "linear_attention",
+      "linear_attention",
+      "linear_attention",
+      "full_attention",
+      "linear_attention",
+      "linear_attention",
+      "linear_attention",
+      "full_attention"
+    ],
+    "linear_conv_kernel_dim": 4,
+    "linear_key_head_dim": 128,
+    "linear_num_key_heads": 16,
+    "linear_num_value_heads": 16,
+    "linear_value_head_dim": 128,
+    "mamba_ssm_dtype": "float32",
+    "max_position_embeddings": 262144,
+    "mlp_only_layers": [],
+    "model_type": "qwen3_5_text",
+    "mtp_num_hidden_layers": 1,
+    "mtp_use_dedicated_embeddings": false,
+    "num_attention_heads": 8,
+    "num_hidden_layers": 24,
+    "num_key_value_heads": 2,
+    "pad_token_id": 248044,
+    "partial_rotary_factor": 0.25,
+    "rms_norm_eps": 1e-06,
+    "rope_parameters": {
+      "mrope_interleaved": true,
+      "mrope_section": [
+        11,
+        11,
+        10
+      ],
+      "partial_rotary_factor": 0.25,
+      "rope_theta": 10000000,
+      "rope_type": "default"
+    },
+    "tie_word_embeddings": true,
+    "use_cache": false,
+    "vocab_size": 248320
+  },
+  "tie_word_embeddings": true,
+  "transformers_version": "5.7.0",
+  "use_cache": false,
+  "video_token_id": 248057,
+  "vision_config": {
+    "deepstack_visual_indexes": [],
+    "depth": 24,
+    "dtype": "bfloat16",
+    "hidden_act": "gelu_pytorch_tanh",
+    "hidden_size": 1024,
+    "in_channels": 3,
+    "initializer_range": 0.02,
+    "intermediate_size": 4096,
+    "model_type": "qwen3_5_vision",
+    "num_heads": 16,
+    "num_position_embeddings": 2304,
+    "out_hidden_size": 2048,
+    "patch_size": 16,
+    "spatial_merge_size": 2,
+    "temporal_patch_size": 2
+  },
+  "vision_end_token_id": 248054,
+  "vision_start_token_id": 248053,
+  "auto_map": {
+    "AutoModelForCausalLM": "modeling_marlin.MarlinForConditionalGeneration"
+  }
+}

generation_config.json

@@ -0,0 +1,9 @@
+{
+  "_from_model_config": true,
+  "eos_token_id": [
+    248044,
+    248046
+  ],
+  "transformers_version": "5.7.0",
+  "use_cache": true
+}

model-00001-of-00002.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5d78fa4dbd856dc89c01b99ffa92072fe31b8a1e6b31e87893734c80304983b7
+size 4999157736

model-00002-of-00002.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:01d40ec9ccf4c2ad8e755604468dd6ee4a5c6551553e5739a03beb4c0673d0db
+size 444519488

model.safetensors.index.json

@@ -0,0 +1,626 @@
+{
+  "metadata": {
+    "total_parameters": 2213241664,
+    "total_size": 5443602048
+  },
+  "weight_map": {
+    "lm_head.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.embed_tokens.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.0.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.0.linear_attn.A_log": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.0.linear_attn.conv1d.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.0.linear_attn.dt_bias": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.0.linear_attn.in_proj_a.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.0.linear_attn.in_proj_b.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.0.linear_attn.in_proj_qkv.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.0.linear_attn.in_proj_z.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.0.linear_attn.norm.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.0.linear_attn.out_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.0.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.0.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.0.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.0.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.1.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.1.linear_attn.A_log": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.1.linear_attn.conv1d.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.1.linear_attn.dt_bias": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.1.linear_attn.in_proj_a.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.1.linear_attn.in_proj_b.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.1.linear_attn.in_proj_qkv.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.1.linear_attn.in_proj_z.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.1.linear_attn.norm.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.1.linear_attn.out_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.1.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.1.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.1.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.1.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.10.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.10.linear_attn.A_log": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.10.linear_attn.conv1d.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.10.linear_attn.dt_bias": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.10.linear_attn.in_proj_a.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.10.linear_attn.in_proj_b.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.10.linear_attn.in_proj_qkv.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.10.linear_attn.in_proj_z.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.10.linear_attn.norm.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.10.linear_attn.out_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.10.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.10.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.10.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.10.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.11.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.11.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.11.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.11.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.11.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.11.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.11.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.11.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.11.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.11.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.11.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.12.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.12.linear_attn.A_log": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.12.linear_attn.conv1d.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.12.linear_attn.dt_bias": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.12.linear_attn.in_proj_a.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.12.linear_attn.in_proj_b.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.12.linear_attn.in_proj_qkv.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.12.linear_attn.in_proj_z.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.12.linear_attn.norm.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.12.linear_attn.out_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.12.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.12.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.12.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.12.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.13.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.13.linear_attn.A_log": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.13.linear_attn.conv1d.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.13.linear_attn.dt_bias": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.13.linear_attn.in_proj_a.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.13.linear_attn.in_proj_b.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.13.linear_attn.in_proj_qkv.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.13.linear_attn.in_proj_z.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.13.linear_attn.norm.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.13.linear_attn.out_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.13.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.13.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.13.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.13.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.14.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.14.linear_attn.A_log": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.14.linear_attn.conv1d.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.14.linear_attn.dt_bias": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.14.linear_attn.in_proj_a.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.14.linear_attn.in_proj_b.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.14.linear_attn.in_proj_qkv.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.14.linear_attn.in_proj_z.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.14.linear_attn.norm.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.14.linear_attn.out_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.14.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.14.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.14.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.14.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.15.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.15.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.15.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.15.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.15.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.15.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.15.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.15.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.15.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.15.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.15.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.16.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.16.linear_attn.A_log": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.16.linear_attn.conv1d.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.16.linear_attn.dt_bias": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.16.linear_attn.in_proj_a.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.16.linear_attn.in_proj_b.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.16.linear_attn.in_proj_qkv.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.16.linear_attn.in_proj_z.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.16.linear_attn.norm.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.16.linear_attn.out_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.16.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.16.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.16.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.16.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.17.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.17.linear_attn.A_log": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.17.linear_attn.conv1d.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.17.linear_attn.dt_bias": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.17.linear_attn.in_proj_a.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.17.linear_attn.in_proj_b.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.17.linear_attn.in_proj_qkv.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.17.linear_attn.in_proj_z.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.17.linear_attn.norm.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.17.linear_attn.out_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.17.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.17.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.17.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.17.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.18.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.18.linear_attn.A_log": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.18.linear_attn.conv1d.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.18.linear_attn.dt_bias": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.18.linear_attn.in_proj_a.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.18.linear_attn.in_proj_b.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.18.linear_attn.in_proj_qkv.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.18.linear_attn.in_proj_z.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.18.linear_attn.norm.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.18.linear_attn.out_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.18.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.18.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.18.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.18.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.19.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.19.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.19.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.19.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.19.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.19.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.19.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.19.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.19.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.19.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.19.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.2.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.2.linear_attn.A_log": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.2.linear_attn.conv1d.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.2.linear_attn.dt_bias": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.2.linear_attn.in_proj_a.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.2.linear_attn.in_proj_b.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.2.linear_attn.in_proj_qkv.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.2.linear_attn.in_proj_z.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.2.linear_attn.norm.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.2.linear_attn.out_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.2.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.2.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.2.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.2.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.20.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.20.linear_attn.A_log": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.20.linear_attn.conv1d.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.20.linear_attn.dt_bias": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.20.linear_attn.in_proj_a.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.20.linear_attn.in_proj_b.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.20.linear_attn.in_proj_qkv.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.20.linear_attn.in_proj_z.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.20.linear_attn.norm.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.20.linear_attn.out_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.20.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.20.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.20.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.20.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.21.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.21.linear_attn.A_log": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.21.linear_attn.conv1d.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.21.linear_attn.dt_bias": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.21.linear_attn.in_proj_a.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.21.linear_attn.in_proj_b.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.21.linear_attn.in_proj_qkv.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.21.linear_attn.in_proj_z.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.21.linear_attn.norm.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.21.linear_attn.out_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.21.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.21.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.21.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.21.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.22.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.22.linear_attn.A_log": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.22.linear_attn.conv1d.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.22.linear_attn.dt_bias": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.22.linear_attn.in_proj_a.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.22.linear_attn.in_proj_b.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.22.linear_attn.in_proj_qkv.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.22.linear_attn.in_proj_z.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.22.linear_attn.norm.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.22.linear_attn.out_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.22.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.22.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.22.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.22.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.23.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.23.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.23.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.23.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.23.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.23.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.23.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.23.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.23.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.23.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.23.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.3.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.3.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.3.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.3.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.3.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.3.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.3.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.3.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.3.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.3.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.3.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.4.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.4.linear_attn.A_log": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.4.linear_attn.conv1d.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.4.linear_attn.dt_bias": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.4.linear_attn.in_proj_a.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.4.linear_attn.in_proj_b.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.4.linear_attn.in_proj_qkv.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.4.linear_attn.in_proj_z.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.4.linear_attn.norm.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.4.linear_attn.out_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.4.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.4.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.4.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.4.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.5.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.5.linear_attn.A_log": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.5.linear_attn.conv1d.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.5.linear_attn.dt_bias": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.5.linear_attn.in_proj_a.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.5.linear_attn.in_proj_b.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.5.linear_attn.in_proj_qkv.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.5.linear_attn.in_proj_z.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.5.linear_attn.norm.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.5.linear_attn.out_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.5.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.5.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.5.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.5.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.6.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.6.linear_attn.A_log": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.6.linear_attn.conv1d.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.6.linear_attn.dt_bias": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.6.linear_attn.in_proj_a.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.6.linear_attn.in_proj_b.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.6.linear_attn.in_proj_qkv.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.6.linear_attn.in_proj_z.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.6.linear_attn.norm.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.6.linear_attn.out_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.6.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.6.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.6.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.6.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.7.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.7.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.7.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.7.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.7.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.7.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.7.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.7.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.7.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.7.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.7.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.8.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.8.linear_attn.A_log": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.8.linear_attn.conv1d.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.8.linear_attn.dt_bias": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.8.linear_attn.in_proj_a.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.8.linear_attn.in_proj_b.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.8.linear_attn.in_proj_qkv.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.8.linear_attn.in_proj_z.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.8.linear_attn.norm.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.8.linear_attn.out_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.8.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.8.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.8.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.8.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.9.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.9.linear_attn.A_log": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.9.linear_attn.conv1d.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.9.linear_attn.dt_bias": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.9.linear_attn.in_proj_a.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.9.linear_attn.in_proj_b.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.9.linear_attn.in_proj_qkv.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.9.linear_attn.in_proj_z.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.9.linear_attn.norm.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.9.linear_attn.out_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.9.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.9.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.9.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.layers.9.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.language_model.norm.weight": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.0.attn.proj.bias": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.0.attn.proj.weight": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.0.attn.qkv.bias": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.0.attn.qkv.weight": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.0.mlp.linear_fc1.bias": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.0.mlp.linear_fc1.weight": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.0.mlp.linear_fc2.bias": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.0.mlp.linear_fc2.weight": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.0.norm1.bias": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.0.norm1.weight": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.0.norm2.bias": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.0.norm2.weight": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.1.attn.proj.bias": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.1.attn.proj.weight": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.1.attn.qkv.bias": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.1.attn.qkv.weight": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.1.mlp.linear_fc1.bias": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.1.mlp.linear_fc1.weight": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.1.mlp.linear_fc2.bias": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.1.mlp.linear_fc2.weight": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.1.norm1.bias": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.1.norm1.weight": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.1.norm2.bias": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.1.norm2.weight": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.10.attn.proj.bias": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.10.attn.proj.weight": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.10.attn.qkv.bias": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.10.attn.qkv.weight": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.10.mlp.linear_fc1.bias": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.10.mlp.linear_fc1.weight": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.10.mlp.linear_fc2.bias": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.10.mlp.linear_fc2.weight": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.10.norm1.bias": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.10.norm1.weight": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.10.norm2.bias": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.10.norm2.weight": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.11.attn.proj.bias": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.11.attn.proj.weight": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.11.attn.qkv.bias": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.11.attn.qkv.weight": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.11.mlp.linear_fc1.bias": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.11.mlp.linear_fc1.weight": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.11.mlp.linear_fc2.bias": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.11.mlp.linear_fc2.weight": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.11.norm1.bias": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.11.norm1.weight": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.11.norm2.bias": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.11.norm2.weight": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.12.attn.proj.bias": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.12.attn.proj.weight": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.12.attn.qkv.bias": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.12.attn.qkv.weight": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.12.mlp.linear_fc1.bias": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.12.mlp.linear_fc1.weight": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.12.mlp.linear_fc2.bias": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.12.mlp.linear_fc2.weight": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.12.norm1.bias": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.12.norm1.weight": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.12.norm2.bias": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.12.norm2.weight": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.13.attn.proj.bias": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.13.attn.proj.weight": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.13.attn.qkv.bias": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.13.attn.qkv.weight": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.13.mlp.linear_fc1.bias": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.13.mlp.linear_fc1.weight": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.13.mlp.linear_fc2.bias": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.13.mlp.linear_fc2.weight": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.13.norm1.bias": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.13.norm1.weight": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.13.norm2.bias": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.13.norm2.weight": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.14.attn.proj.bias": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.14.attn.proj.weight": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.14.attn.qkv.bias": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.14.attn.qkv.weight": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.14.mlp.linear_fc1.bias": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.14.mlp.linear_fc1.weight": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.14.mlp.linear_fc2.bias": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.14.mlp.linear_fc2.weight": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.14.norm1.bias": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.14.norm1.weight": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.14.norm2.bias": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.14.norm2.weight": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.15.attn.proj.bias": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.15.attn.proj.weight": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.15.attn.qkv.bias": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.15.attn.qkv.weight": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.15.mlp.linear_fc1.bias": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.15.mlp.linear_fc1.weight": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.15.mlp.linear_fc2.bias": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.15.mlp.linear_fc2.weight": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.15.norm1.bias": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.15.norm1.weight": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.15.norm2.bias": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.15.norm2.weight": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.16.attn.proj.bias": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.16.attn.proj.weight": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.16.attn.qkv.bias": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.16.attn.qkv.weight": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.16.mlp.linear_fc1.bias": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.16.mlp.linear_fc1.weight": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.16.mlp.linear_fc2.bias": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.16.mlp.linear_fc2.weight": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.16.norm1.bias": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.16.norm1.weight": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.16.norm2.bias": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.16.norm2.weight": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.17.attn.proj.bias": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.17.attn.proj.weight": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.17.attn.qkv.bias": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.17.attn.qkv.weight": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.17.mlp.linear_fc1.bias": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.17.mlp.linear_fc1.weight": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.17.mlp.linear_fc2.bias": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.17.mlp.linear_fc2.weight": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.17.norm1.bias": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.17.norm1.weight": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.17.norm2.bias": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.17.norm2.weight": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.18.attn.proj.bias": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.18.attn.proj.weight": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.18.attn.qkv.bias": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.18.attn.qkv.weight": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.18.mlp.linear_fc1.bias": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.18.mlp.linear_fc1.weight": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.18.mlp.linear_fc2.bias": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.18.mlp.linear_fc2.weight": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.18.norm1.bias": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.18.norm1.weight": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.18.norm2.bias": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.18.norm2.weight": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.19.attn.proj.bias": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.19.attn.proj.weight": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.19.attn.qkv.bias": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.19.attn.qkv.weight": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.19.mlp.linear_fc1.bias": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.19.mlp.linear_fc1.weight": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.19.mlp.linear_fc2.bias": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.19.mlp.linear_fc2.weight": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.19.norm1.bias": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.19.norm1.weight": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.19.norm2.bias": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.19.norm2.weight": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.2.attn.proj.bias": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.2.attn.proj.weight": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.2.attn.qkv.bias": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.2.attn.qkv.weight": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.2.mlp.linear_fc1.bias": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.2.mlp.linear_fc1.weight": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.2.mlp.linear_fc2.bias": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.2.mlp.linear_fc2.weight": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.2.norm1.bias": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.2.norm1.weight": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.2.norm2.bias": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.2.norm2.weight": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.20.attn.proj.bias": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.20.attn.proj.weight": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.20.attn.qkv.bias": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.20.attn.qkv.weight": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.20.mlp.linear_fc1.bias": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.20.mlp.linear_fc1.weight": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.20.mlp.linear_fc2.bias": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.20.mlp.linear_fc2.weight": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.20.norm1.bias": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.20.norm1.weight": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.20.norm2.bias": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.20.norm2.weight": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.21.attn.proj.bias": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.21.attn.proj.weight": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.21.attn.qkv.bias": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.21.attn.qkv.weight": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.21.mlp.linear_fc1.bias": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.21.mlp.linear_fc1.weight": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.21.mlp.linear_fc2.bias": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.21.mlp.linear_fc2.weight": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.21.norm1.bias": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.21.norm1.weight": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.21.norm2.bias": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.21.norm2.weight": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.22.attn.proj.bias": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.22.attn.proj.weight": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.22.attn.qkv.bias": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.22.attn.qkv.weight": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.22.mlp.linear_fc1.bias": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.22.mlp.linear_fc1.weight": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.22.mlp.linear_fc2.bias": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.22.mlp.linear_fc2.weight": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.22.norm1.bias": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.22.norm1.weight": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.22.norm2.bias": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.22.norm2.weight": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.23.attn.proj.bias": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.23.attn.proj.weight": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.23.attn.qkv.bias": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.23.attn.qkv.weight": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.23.mlp.linear_fc1.bias": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.23.mlp.linear_fc1.weight": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.23.mlp.linear_fc2.bias": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.23.mlp.linear_fc2.weight": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.23.norm1.bias": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.23.norm1.weight": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.23.norm2.bias": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.23.norm2.weight": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.3.attn.proj.bias": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.3.attn.proj.weight": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.3.attn.qkv.bias": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.3.attn.qkv.weight": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.3.mlp.linear_fc1.bias": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.3.mlp.linear_fc1.weight": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.3.mlp.linear_fc2.bias": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.3.mlp.linear_fc2.weight": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.3.norm1.bias": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.3.norm1.weight": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.3.norm2.bias": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.3.norm2.weight": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.4.attn.proj.bias": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.4.attn.proj.weight": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.4.attn.qkv.bias": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.4.attn.qkv.weight": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.4.mlp.linear_fc1.bias": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.4.mlp.linear_fc1.weight": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.4.mlp.linear_fc2.bias": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.4.mlp.linear_fc2.weight": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.4.norm1.bias": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.4.norm1.weight": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.4.norm2.bias": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.4.norm2.weight": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.5.attn.proj.bias": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.5.attn.proj.weight": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.5.attn.qkv.bias": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.5.attn.qkv.weight": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.5.mlp.linear_fc1.bias": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.5.mlp.linear_fc1.weight": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.5.mlp.linear_fc2.bias": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.5.mlp.linear_fc2.weight": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.5.norm1.bias": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.5.norm1.weight": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.5.norm2.bias": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.5.norm2.weight": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.6.attn.proj.bias": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.6.attn.proj.weight": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.6.attn.qkv.bias": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.6.attn.qkv.weight": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.6.mlp.linear_fc1.bias": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.6.mlp.linear_fc1.weight": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.6.mlp.linear_fc2.bias": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.6.mlp.linear_fc2.weight": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.6.norm1.bias": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.6.norm1.weight": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.6.norm2.bias": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.6.norm2.weight": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.7.attn.proj.bias": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.7.attn.proj.weight": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.7.attn.qkv.bias": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.7.attn.qkv.weight": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.7.mlp.linear_fc1.bias": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.7.mlp.linear_fc1.weight": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.7.mlp.linear_fc2.bias": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.7.mlp.linear_fc2.weight": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.7.norm1.bias": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.7.norm1.weight": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.7.norm2.bias": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.7.norm2.weight": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.8.attn.proj.bias": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.8.attn.proj.weight": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.8.attn.qkv.bias": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.8.attn.qkv.weight": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.8.mlp.linear_fc1.bias": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.8.mlp.linear_fc1.weight": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.8.mlp.linear_fc2.bias": "model-00001-of-00002.safetensors",
+    "model.visual.blocks.8.mlp.linear_fc2.weight": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.8.norm1.bias": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.8.norm1.weight": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.8.norm2.bias": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.8.norm2.weight": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.9.attn.proj.bias": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.9.attn.proj.weight": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.9.attn.qkv.bias": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.9.attn.qkv.weight": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.9.mlp.linear_fc1.bias": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.9.mlp.linear_fc1.weight": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.9.mlp.linear_fc2.bias": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.9.mlp.linear_fc2.weight": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.9.norm1.bias": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.9.norm1.weight": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.9.norm2.bias": "model-00002-of-00002.safetensors",
+    "model.visual.blocks.9.norm2.weight": "model-00002-of-00002.safetensors",
+    "model.visual.merger.linear_fc1.bias": "model-00002-of-00002.safetensors",
+    "model.visual.merger.linear_fc1.weight": "model-00002-of-00002.safetensors",
+    "model.visual.merger.linear_fc2.bias": "model-00002-of-00002.safetensors",
+    "model.visual.merger.linear_fc2.weight": "model-00002-of-00002.safetensors",
+    "model.visual.merger.norm.bias": "model-00002-of-00002.safetensors",
+    "model.visual.merger.norm.weight": "model-00002-of-00002.safetensors",
+    "model.visual.patch_embed.proj.bias": "model-00002-of-00002.safetensors",
+    "model.visual.patch_embed.proj.weight": "model-00002-of-00002.safetensors",
+    "model.visual.pos_embed.weight": "model-00002-of-00002.safetensors"
+  }
+}

modeling_marlin.py

@@ -0,0 +1,690 @@
+"""Custom HuggingFace modeling code for Marlin.
+
+This module subclasses the upstream ``Qwen3_5ForConditionalGeneration``
+(native in ``transformers >= 5.7.0``) and adds two convenience methods —
+:meth:`MarlinForConditionalGeneration.caption` and
+:meth:`MarlinForConditionalGeneration.find` — that mirror moondream's
+image-SDK ergonomics for video captioning and temporal grounding.
+
+The forward pass is **not** modified: we only add chat-template + generate +
+post-processing wrappers. Loading the model through
+``AutoModelForCausalLM.from_pretrained(..., trust_remote_code=True)`` returns
+this subclass thanks to the ``auto_map`` entry in ``config.json``.
+
+Required environment for video inference (set before importing transformers)::
+
+    FORCE_QWENVL_VIDEO_READER=torchcodec
+    VIDEO_MAX_PIXELS=200704
+    FPS=2.0
+    FPS_MAX_FRAMES=240
+    FPS_MIN_FRAMES=4
+
+System requirements:
+
+* transformers >= 5.7.0
+* torch >= 2.11.0
+* torchcodec
+* qwen-vl-utils >= 0.0.14
+* av, pillow
+"""
+
+from __future__ import annotations
+
+import os
+import re
+import subprocess
+import tempfile
+from typing import Any, Dict, List, Optional, Tuple, TypedDict, Union
+
+import torch
+
+# ``Qwen3_5ForConditionalGeneration`` is the native HF class for Marlin's
+# backbone (Qwen3.5-2B with vision tower). It ships in transformers >= 5.7.0.
+# We import it lazily-friendly at module top so AutoModelForCausalLM resolution
+# works correctly when this file is loaded via ``trust_remote_code=True``.
+from transformers import Qwen3_5ForConditionalGeneration
+
+__all__ = [
+    "CAPTION_PROMPT",
+    "GROUNDING_PROMPT_TEMPLATE",
+    "CaptionResult",
+    "FindResult",
+    "MultiFindResult",
+    "Event",
+    "MarlinForConditionalGeneration",
+    "strip_thinking",
+    "parse_caption",
+    "parse_span",
+]
+
+
+# ---------------------------------------------------------------------------
+# Canonical training-time prompts — DO NOT EDIT
+# ---------------------------------------------------------------------------
+#
+# These strings must match exactly what the model was fine-tuned on. Diverging
+# from them silently degrades quality.
+
+CAPTION_PROMPT: str = (
+    "Provide a spatial description of this clip followed by time-ranged events.\n"
+    "For each event, give the time range as <start - end> and a short description."
+)
+
+GROUNDING_PROMPT_TEMPLATE: str = (
+    'Identify the timestamps during which "{event}" takes place. '
+    'Output the time range as "From <start> to <end>." (numbers in seconds).'
+)
+
+
+# ---------------------------------------------------------------------------
+# Thinking-tag stripping
+# ---------------------------------------------------------------------------
+#
+# ms-swift's Marlin training template uses ``add_non_thinking_prefix=True``,
+# which prefixes every response with a bare ``<think>\n`` (no close tag). The
+# model occasionally also emits a complete ``<think>...</think>`` block. Strip
+# both robustly.
+
+_THINK_BLOCK = re.compile(r"<think>.*?</think>\s*", re.DOTALL)
+_THINK_PREFIX = re.compile(r"^\s*<think>\s*\n*", re.IGNORECASE)
+_THINK_CLOSE = re.compile(r"</think>\s*", re.IGNORECASE)
+
+
+def strip_thinking(text: str) -> str:
+    """Remove ``<think>...</think>`` blocks and bare ``<think>`` prefixes.
+
+    Parameters
+    ----------
+    text:
+        Raw model output.
+
+    Returns
+    -------
+    str
+        The text with any thinking artifacts removed and outer whitespace
+        stripped.
+    """
+    out = _THINK_BLOCK.sub("", text)
+    out = _THINK_PREFIX.sub("", out)
+    out = _THINK_CLOSE.sub("", out)
+    return out.strip()
+
+
+# ---------------------------------------------------------------------------
+# Mode 1 — dense caption parser
+# ---------------------------------------------------------------------------
+
+
+class Event(TypedDict):
+    """A single time-ranged event extracted from a dense caption."""
+
+    start: float
+    end: float
+    description: str
+
+
+# Tolerates ``<1.2 - 3.4>`` / ``1.2 - 3.4`` / ``1.2-3.4`` with optional units.
+# Unit alternation is ordered longest-first so e.g. ``"1.8 seconds"`` consumes
+# the full word instead of leaving ``"econds"`` in the description.
+_EVENT_LINE = re.compile(
+    r"^\s*<?\s*(\d+\.?\d*)\s*(?:seconds?|secs?|s)?\s*-\s*"
+    r"(\d+\.?\d*)\s*(?:seconds?|secs?|s)?\s*>?\s*[:\-]?\s*(.+?)\s*$"
+)
+
+
+def _parse_events(events_block: str) -> List[Event]:
+    """Parse a multi-line events block into a list of :class:`Event` dicts."""
+    out: List[Event] = []
+    for raw_line in events_block.splitlines():
+        line = raw_line.strip()
+        if not line:
+            continue
+        m = _EVENT_LINE.match(line)
+        if not m:
+            continue
+        start = float(m.group(1))
+        end = float(m.group(2))
+        desc = m.group(3).strip().lstrip("-").strip()
+        if end <= start or not desc:
+            continue
+        out.append(Event(start=start, end=end, description=desc))
+    return out
+
+
+def parse_caption(text: str) -> Tuple[str, str, List[Event]]:
+    """Parse a Mode 1 caption into ``(caption, scene, events)``.
+
+    The model is trained to produce::
+
+        Scene: <one-paragraph spatial description>
+
+        Events:
+        <start - end> <description>
+        <start - end> <description>
+
+    The parser is tolerant: if explicit ``Scene:`` / ``Events:`` headers are
+    missing, ``scene`` falls back to everything before the first event line and
+    ``events`` is whatever event-shaped lines were detected.
+
+    Parameters
+    ----------
+    text:
+        Raw model output. Thinking artifacts will be stripped.
+
+    Returns
+    -------
+    tuple
+        ``(caption, scene, events)`` — the post-thinking full text, the parsed
+        scene paragraph, and a list of :class:`Event` dicts in emission order.
+    """
+    cleaned = strip_thinking(text)
+
+    scene_match = re.search(
+        r"(?:^|\n)\s*Scene\s*:\s*(.*?)(?=\n\s*Events\s*:|\Z)",
+        cleaned,
+        re.IGNORECASE | re.DOTALL,
+    )
+    events_match = re.search(
+        r"(?:^|\n)\s*Events\s*:\s*(.*)\Z",
+        cleaned,
+        re.IGNORECASE | re.DOTALL,
+    )
+
+    if scene_match:
+        scene = scene_match.group(1).strip()
+    else:
+        # Fallback: scene = everything before the first event-shaped line.
+        scene_lines: List[str] = []
+        for line in cleaned.splitlines():
+            if _EVENT_LINE.match(line.strip()):
+                break
+            scene_lines.append(line)
+        scene = "\n".join(scene_lines).strip()
+
+    events_block = events_match.group(1) if events_match else cleaned
+    events = _parse_events(events_block)
+
+    return cleaned, scene, events
+
+
+# ---------------------------------------------------------------------------
+# Mode 2 — temporal grounding parser
+# ---------------------------------------------------------------------------
+
+# Tolerates ``From 1.2 to 3.4.``, ``From 1.2s to 3.4 sec``; trailing period
+# optional.
+_SPAN_RE = re.compile(
+    r"From\s+(\d+\.?\d*)\s*(?:s|sec)?\s+to\s+(\d+\.?\d*)\s*(?:s|sec)?\.?",
+    re.IGNORECASE,
+)
+
+
+def parse_span(text: str) -> Tuple[str, Optional[Tuple[float, float]]]:
+    """Parse a Mode 2 grounding output into ``(text, span)``.
+
+    Parameters
+    ----------
+    text:
+        Raw model output. Thinking artifacts will be stripped.
+
+    Returns
+    -------
+    tuple
+        ``(cleaned, span)`` — the post-thinking text and ``(start, end)`` in
+        seconds, or ``None`` if no valid ``"From X to Y"`` substring was found
+        or the span was non-positive.
+    """
+    cleaned = strip_thinking(text)
+    m = _SPAN_RE.search(cleaned)
+    if not m:
+        return cleaned, None
+    start = float(m.group(1))
+    end = float(m.group(2))
+    if end <= start:
+        return cleaned, None
+    return cleaned, (start, end)
+
+
+# ---------------------------------------------------------------------------
+# Result dicts
+# ---------------------------------------------------------------------------
+
+
+class CaptionResult(TypedDict):
+    """Return type for :meth:`MarlinForConditionalGeneration.caption`.
+
+    Keys
+    ----
+    caption : str
+        Post-thinking model output (e.g. ``"Scene: ...\\n\\nEvents:\\n..."``).
+    scene : str
+        Parsed ``Scene:`` paragraph.
+    events : list of :class:`Event`
+        Parsed ``{start, end, description}`` dicts in emission order.
+    raw : str
+        Raw model output *before* thinking-prefix stripping (for debugging).
+    """
+
+    caption: str
+    scene: str
+    events: List[Event]
+    raw: str
+
+
+class FindResult(TypedDict):
+    """Return type for :meth:`MarlinForConditionalGeneration.find`.
+
+    Keys
+    ----
+    raw : str
+        Raw post-thinking model output (e.g. ``"From 1.2 to 3.4."``).
+    span : tuple of (float, float) or None
+        ``(start, end)`` in seconds, or ``None`` if parsing failed.
+    format_ok : bool
+        ``True`` iff the output matched the trained ``"From X to Y."`` format.
+    """
+
+    raw: str
+    span: Optional[Tuple[float, float]]
+    format_ok: bool
+
+
+class MultiFindResult(TypedDict):
+    """Return type for :meth:`MarlinForConditionalGeneration.multi_find`.
+
+    Keys
+    ----
+    event : str
+        The natural-language event that was searched for.
+    spans : list of (float, float)
+        Up to ``max_events`` ``(start, end)`` ranges in seconds, ordered
+        earliest-first across the full video timeline.
+    stopped : str
+        Why the search ended: ``"max_events"`` (the cap was reached),
+        ``"end_of_video"`` (the cursor ran off the timeline), or
+        ``"no_span"`` (``find`` stopped producing a usable span).
+    """
+
+    event: str
+    spans: List[Tuple[float, float]]
+    stopped: str
+
+
+# ---------------------------------------------------------------------------
+# Default video-preprocessing env vars
+# ---------------------------------------------------------------------------
+#
+# qwen-vl-utils reads these from the environment when ``apply_chat_template``
+# decodes a video. We populate them here as a safety net for users who forget
+# to set them before importing transformers. Existing env values are NEVER
+# overwritten — explicit user settings always win.
+
+_DEFAULT_VIDEO_ENV: Dict[str, str] = {
+    "FORCE_QWENVL_VIDEO_READER": "torchcodec",
+    "VIDEO_MAX_PIXELS": "200704",
+    "FPS": "2.0",
+    "FPS_MAX_FRAMES": "240",
+    "FPS_MIN_FRAMES": "4",
+}
+
+for _k, _v in _DEFAULT_VIDEO_ENV.items():
+    os.environ.setdefault(_k, _v)
+
+
+# ---------------------------------------------------------------------------
+# The actual model class
+# ---------------------------------------------------------------------------
+
+
+class MarlinForConditionalGeneration(Qwen3_5ForConditionalGeneration):
+    """Marlin with ``.caption()`` and ``.find()`` convenience methods.
+
+    Inherits the full forward / generate / from_pretrained machinery from
+    :class:`transformers.Qwen3_5ForConditionalGeneration`; only adds two
+    helpers that wrap chat-template construction, generation, and the trained
+    output parsers.
+
+    Use it via the standard auto class::
+
+        from transformers import AutoModelForCausalLM
+        model = AutoModelForCausalLM.from_pretrained(
+            "NemoStation/Marlin-2B",
+            trust_remote_code=True,
+            dtype=torch.bfloat16,
+            device_map={"": "cuda"},
+        )
+        result = model.caption("video.mp4")
+        span = model.find("video.mp4", event="a person enters the room")
+    """
+
+    # ------------------------------------------------------------------ utils
+
+    @property
+    def processor(self):  # type: ignore[override]
+        """Lazily-loaded :class:`~transformers.AutoProcessor` for this checkpoint.
+
+        Cached on the instance to avoid the expensive HF Hub lookup on every
+        call.
+        """
+        cached = getattr(self, "_processor", None)
+        if cached is None:
+            from transformers import AutoProcessor
+
+            cached = AutoProcessor.from_pretrained(
+                self.config._name_or_path,
+                trust_remote_code=True,
+            )
+            self._processor = cached
+        return cached
+
+    def compile(self, *args: Any, **kwargs: Any) -> "MarlinForConditionalGeneration":
+        """Optional ``torch.compile`` wrapper around the model.
+
+        Returns ``self`` so it chains naturally after ``from_pretrained``::
+
+            model = AutoModelForCausalLM.from_pretrained(...).compile()
+
+        All positional / keyword args are forwarded to ``torch.compile``.
+        """
+        # ``torch.compile`` replaces the module's forward with a compiled
+        # version in-place; we still return self for fluent chaining.
+        torch.compile(self, *args, **kwargs)
+        return self
+
+    # ----------------------------------------------------------- core generate
+
+    def _generate_video(
+        self,
+        video_path: Union[str, os.PathLike],
+        prompt: str,
+        max_tokens: int,
+        *,
+        do_sample: bool = False,
+        temperature: float = 1.0,
+        top_p: float = 1.0,
+    ) -> str:
+        """Build a chat message with one video + one text turn and decode.
+
+        Returns the raw decoded string (with any ``<think>`` artifacts still
+        attached — callers are expected to run :func:`strip_thinking`).
+        """
+        messages = [
+            {
+                "role": "user",
+                "content": [
+                    {"type": "video", "video": str(video_path)},
+                    {"type": "text", "text": prompt},
+                ],
+            }
+        ]
+
+        inputs = self.processor.apply_chat_template(
+            messages,
+            tokenize=True,
+            add_generation_prompt=True,
+            return_tensors="pt",
+            return_dict=True,
+        ).to(self.device)
+
+        with torch.inference_mode():
+            out = self.generate(
+                **inputs,
+                max_new_tokens=max_tokens,
+                do_sample=do_sample,
+                temperature=temperature if do_sample else 1.0,
+                top_p=top_p if do_sample else 1.0,
+            )
+
+        # Strip the prompt prefix so we only return the model's continuation.
+        prompt_len = inputs["input_ids"].shape[1]
+        out = out[:, prompt_len:]
+        return self.processor.batch_decode(out, skip_special_tokens=True)[0]
+
+    # ----------------------------------------------------- video sub-clipping
+
+    @staticmethod
+    def _video_duration(video_path: Union[str, os.PathLike]) -> float:
+        """Return the video duration in seconds via the torchcodec decoder."""
+        from torchcodec.decoders import VideoDecoder
+
+        return float(VideoDecoder(str(video_path)).metadata.duration_seconds)
+
+    @staticmethod
+    def _clip_segment(
+        video_path: Union[str, os.PathLike], start: float, end: float
+    ) -> str:
+        """Trim ``video_path`` to ``[start, end]`` (seconds) into a temp mp4.
+
+        Re-encodes with libx264 for frame-accurate boundaries (stream-copy would
+        snap to keyframes and skew the timestamps we remap against). The caller
+        owns the returned path and must ``os.unlink`` it.
+        """
+        fd, out_path = tempfile.mkstemp(suffix=".mp4")
+        os.close(fd)
+        subprocess.run(
+            [
+                "ffmpeg", "-y",
+                "-ss", f"{start:.3f}",
+                "-to", f"{end:.3f}",
+                "-i", str(video_path),
+                "-c:v", "libx264",
+                "-an",
+                "-loglevel", "error",
+                out_path,
+            ],
+            check=True,
+        )
+        return out_path
+
+    # ---------------------------------------------------------------- caption
+
+    def caption(
+        self,
+        video_path: Union[str, os.PathLike],
+        *,
+        prompt: Optional[str] = None,
+        do_sample: bool = False,
+        temperature: float = 1.0,
+        top_p: float = 1.0,
+        max_new_tokens: int = 2048,
+    ) -> CaptionResult:
+        """Generate a dense caption for a video.
+
+        Parameters
+        ----------
+        video_path:
+            Local path to a video file (mp4, webm, etc.).
+        prompt:
+            Override the canonical training prompt. Almost always leave at
+            ``None``; diverging from training silently degrades quality.
+        do_sample:
+            If ``True``, switch to nucleus sampling. Defaults to greedy.
+        temperature, top_p:
+            Sampling params, only used when ``do_sample=True``.
+        max_new_tokens:
+            Generation cap. Default 2048 is enough for any dense caption the
+            model produces in practice.
+
+        Returns
+        -------
+        CaptionResult
+            Dict with keys ``caption``, ``scene``, ``events``, ``raw``.
+        """
+        prompt_text = prompt if prompt is not None else CAPTION_PROMPT
+        raw = self._generate_video(
+            video_path,
+            prompt_text,
+            max_tokens=max_new_tokens,
+            do_sample=do_sample,
+            temperature=temperature,
+            top_p=top_p,
+        )
+        cleaned, scene, events = parse_caption(raw)
+        return CaptionResult(
+            caption=cleaned,
+            scene=scene,
+            events=events,
+            raw=raw,
+        )
+
+    # ------------------------------------------------------------------- find
+
+    def find(
+        self,
+        video_path: Union[str, os.PathLike],
+        event: str,
+        *,
+        prompt_template: Optional[str] = None,
+        do_sample: bool = False,
+        temperature: float = 1.0,
+        top_p: float = 1.0,
+        max_new_tokens: int = 64,
+    ) -> FindResult:
+        """Locate when a natural-language event occurs in a video.
+
+        Parameters
+        ----------
+        video_path:
+            Local path to a video file.
+        event:
+            Free-form description of the event to locate, e.g.
+            ``"a person enters the room"``. Inserted into the trained prompt
+            via the ``{event}`` placeholder.
+        prompt_template:
+            Override the canonical training prompt template. Must include a
+            ``{event}`` placeholder. Almost always leave at ``None``.
+        do_sample:
+            If ``True``, switch to nucleus sampling. Defaults to greedy.
+        temperature, top_p:
+            Sampling params, only used when ``do_sample=True``.
+        max_new_tokens:
+            Output budget. 64 is plenty for the one-line trained format.
+
+        Returns
+        -------
+        FindResult
+            Dict with keys ``raw``, ``span`` and ``format_ok``.
+
+        Raises
+        ------
+        ValueError
+            If ``event`` is empty or whitespace-only.
+        """
+        event_str = (event or "").strip()
+        if not event_str:
+            raise ValueError("`event` must be a non-empty string")
+
+        template = prompt_template if prompt_template is not None else GROUNDING_PROMPT_TEMPLATE
+        prompt_text = template.format(event=event_str)
+
+        raw = self._generate_video(
+            video_path,
+            prompt_text,
+            max_tokens=max_new_tokens,
+            do_sample=do_sample,
+            temperature=temperature,
+            top_p=top_p,
+        )
+        cleaned, span = parse_span(raw)
+        return FindResult(
+            raw=cleaned,
+            span=span,
+            format_ok=span is not None,
+        )
+
+    # ------------------------------------------------------------- multi_find
+
+    def multi_find(
+        self,
+        video_path: Union[str, os.PathLike],
+        event: str,
+        max_events: int,
+        *,
+        min_gap: float = 0.3,
+        min_segment: float = 0.5,
+        **find_kwargs: Any,
+    ) -> MultiFindResult:
+        """Locate up to ``max_events`` occurrences of ``event``, left to right.
+
+        :meth:`find` returns only the *earliest* occurrence in whatever clip it
+        is given. ``multi_find`` walks the timeline by repeatedly trimming the
+        video to the not-yet-searched tail ``[cursor, end]``, running
+        :meth:`find` on that sub-clip, remapping the returned (clip-local) span
+        back to global time, then advancing the cursor past it.
+
+        The search stops on exactly two conditions: ``max_events`` spans have
+        been collected, or the cursor runs off the end of the video. Because
+        :meth:`find` always emits *some* span, there is no built-in
+        "event absent" signal — ``max_events`` is the caller-supplied upper
+        bound that bounds the walk. If the event actually occurs fewer than
+        ``max_events`` times, trailing spans may be spurious; set
+        ``max_events`` to a tight expected maximum, not an arbitrarily large
+        number, and treat later spans as lower confidence.
+
+        Parameters
+        ----------
+        video_path:
+            Local path to a video file.
+        event:
+            Free-form description of the event to locate (see :meth:`find`).
+        max_events:
+            Maximum number of occurrences to return (must be >= 1).
+        min_gap:
+            Seconds to advance the cursor past each hit. Also guarantees the
+            cursor always moves forward, so the loop can never stall.
+        min_segment:
+            Stop once the remaining clip is shorter than this, and drop any
+            individual span shorter than this.
+        **find_kwargs:
+            Forwarded verbatim to :meth:`find` (e.g. ``do_sample``,
+            ``temperature``, ``top_p``, ``max_new_tokens``).
+
+        Returns
+        -------
+        MultiFindResult
+            Dict with keys ``event``, ``spans`` and ``stopped``.
+
+        Raises
+        ------
+        ValueError
+            If ``max_events < 1`` or ``event`` is empty / whitespace-only.
+        """
+        if max_events < 1:
+            raise ValueError("`max_events` must be >= 1")
+        event_str = (event or "").strip()
+        if not event_str:
+            raise ValueError("`event` must be a non-empty string")
+
+        duration = self._video_duration(video_path)
+        spans: List[Tuple[float, float]] = []
+        cursor = 0.0
+        stopped = "end_of_video"
+
+        while cursor < duration - min_segment:
+            if len(spans) >= max_events:
+                stopped = "max_events"
+                break
+
+            clip = self._clip_segment(video_path, cursor, duration)
+            try:
+                result = self.find(clip, event_str, **find_kwargs)
+            finally:
+                os.unlink(clip)
+
+            if not result["format_ok"]:
+                stopped = "no_span"
+                break
+
+            local_start, local_end = result["span"]
+            # Remap clip-local timestamps back onto the full-video timeline.
+            global_start = cursor + local_start
+            global_end = min(cursor + local_end, duration)
+            if global_end - global_start < min_segment:
+                stopped = "no_span"
+                break
+
+            spans.append((round(global_start, 1), round(global_end, 1)))
+            # Always move strictly forward, even if the span maps near `cursor`.
+            cursor = max(global_end, cursor + min_gap) + min_gap
+
+        return MultiFindResult(event=event_str, spans=spans, stopped=stopped)

preprocessor_config.json

@@ -0,0 +1,21 @@
+{
+    "size": {
+        "longest_edge": 16777216,
+        "shortest_edge": 65536
+    },
+    "patch_size": 16,
+    "temporal_patch_size": 2,
+    "merge_size": 2,
+    "image_mean": [
+        0.5,
+        0.5,
+        0.5
+    ],
+    "image_std": [
+        0.5,
+        0.5,
+        0.5
+    ],
+    "processor_class": "Qwen3VLProcessor",
+    "image_processor_type": "Qwen2VLImageProcessorFast"
+}

processor_config.json

@@ -0,0 +1,60 @@
+{
+  "image_processor": {
+    "do_convert_rgb": true,
+    "do_normalize": true,
+    "do_rescale": true,
+    "do_resize": true,
+    "image_mean": [
+      0.5,
+      0.5,
+      0.5
+    ],
+    "image_processor_type": "Qwen2VLImageProcessor",
+    "image_std": [
+      0.5,
+      0.5,
+      0.5
+    ],
+    "merge_size": 2,
+    "patch_size": 16,
+    "resample": 3,
+    "rescale_factor": 0.00392156862745098,
+    "size": {
+      "longest_edge": 16777216,
+      "shortest_edge": 65536
+    },
+    "temporal_patch_size": 2
+  },
+  "processor_class": "Qwen3VLProcessor",
+  "video_processor": {
+    "do_convert_rgb": true,
+    "do_normalize": true,
+    "do_rescale": true,
+    "do_resize": true,
+    "do_sample_frames": true,
+    "fps": 2,
+    "image_mean": [
+      0.5,
+      0.5,
+      0.5
+    ],
+    "image_std": [
+      0.5,
+      0.5,
+      0.5
+    ],
+    "max_frames": 768,
+    "merge_size": 2,
+    "min_frames": 4,
+    "patch_size": 16,
+    "resample": 3,
+    "rescale_factor": 0.00392156862745098,
+    "return_metadata": false,
+    "size": {
+      "longest_edge": 25165824,
+      "shortest_edge": 4096
+    },
+    "temporal_patch_size": 2,
+    "video_processor_type": "Qwen3VLVideoProcessor"
+  }
+}

tokenizer.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:06b9509352d2af50381ab2247e083b80d32d5c0aba91c272ca9ff729b6a0e523
+size 19989325

tokenizer_config.json

@@ -0,0 +1,33 @@
+{
+  "add_prefix_space": false,
+  "audio_bos_token": "<|audio_start|>",
+  "audio_eos_token": "<|audio_end|>",
+  "audio_token": "<|audio_pad|>",
+  "backend": "tokenizers",
+  "bos_token": null,
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|im_end|>",
+  "errors": "replace",
+  "image_token": "<|image_pad|>",
+  "is_local": true,
+  "local_files_only": false,
+  "model_max_length": 262144,
+  "model_specific_special_tokens": {
+    "audio_bos_token": "<|audio_start|>",
+    "audio_eos_token": "<|audio_end|>",
+    "audio_token": "<|audio_pad|>",
+    "image_token": "<|image_pad|>",
+    "video_token": "<|video_pad|>",
+    "vision_bos_token": "<|vision_start|>",
+    "vision_eos_token": "<|vision_end|>"
+  },
+  "pad_token": "<|endoftext|>",
+  "pretokenize_regex": "(?i:'s|'t|'re|'ve|'m|'ll|'d)|[^\\r\\n\\p{L}\\p{N}]?[\\p{L}\\p{M}]+|\\p{N}| ?[^\\s\\p{L}\\p{M}\\p{N}]+[\\r\\n]*|\\s*[\\r\\n]+|\\s+(?!\\S)|\\s+",
+  "processor_class": "Qwen3VLProcessor",
+  "split_special_tokens": false,
+  "tokenizer_class": "Qwen2Tokenizer",
+  "unk_token": null,
+  "video_token": "<|video_pad|>",
+  "vision_bos_token": "<|vision_start|>",
+  "vision_eos_token": "<|vision_end|>"
+}