Add VISE LoRA adapter (Qwen3-VL-2B) + 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

README.md

@@ -0,0 +1,211 @@
+---
+base_model: Qwen/Qwen3-VL-2B-Instruct
+base_model_relation: adapter
+library_name: peft
+pipeline_tag: image-text-to-text
+license: apache-2.0
+language:
+- en
+tags:
+- lora
+- peft
+- vise
+- self-evolving
+- multimodal
+- vision-language
+- lmm
+- visual-grounding
+- image-captioning
+- qwen3-vl
+- unsupervised
+---
+
+# VISE — Visual Invariance Self-Evolution (Qwen3-VL-2B)
+
+> **Paying More Attention to Visual Tokens in Self-Evolving Large Multimodal Models** · ECCV 2026
+
+This repository hosts the **VISE** LoRA adapter trained on top of
+[`Qwen/Qwen3-VL-2B-Instruct`](https://huggingface.co/Qwen/Qwen3-VL-2B-Instruct).
+VISE is a **purely unsupervised, single-model self-evolving framework** that
+strengthens a large multimodal model's *visual conditioning* — how much the
+decoder actually attends to the image while generating — instead of optimizing
+answer agreement. It is trained on **raw, unlabeled images** with **no captions,
+bounding boxes, category labels, external reward models, or specialist roles**.
+
+- 📄 **Paper:** *(arXiv link coming soon)*
+- 💻 **Code:** *(GitHub repository coming soon)*
+- 🧩 **Base model:** `Qwen/Qwen3-VL-2B-Instruct`
+- 🔧 **Adapter type:** LoRA (PEFT) · `r=16`, `α=32`, `dropout=0.05`
+
+---
+
+## What problem does VISE solve?
+
+Existing self-evolving LMMs (proposer–solver / questioner–reasoner self-play)
+optimize *answer agreement*. A decoder can reach high self-consistency purely
+from **language priors** without ever attending to the image — a failure mode
+the paper terms **visual under-conditioning**, which causes hallucination,
+modality bypass, and unstable grounding even when the vision encoder is accurate.
+
+VISE replaces answer-agreement rewards with two complementary
+**invariance rewards** computed entirely from the model's *own* predictions:
+
+| Reward | Idea | Signal |
+| --- | --- | --- |
+| **Geometric Invariance** `R_geo` | A correctly-conditioned model should localize the same object consistently under a known spatial transform | `GIoU(B_proj, B_new)` between the box on the transformed view and the analytically projected original box |
+| **Semantic Invariance** `R_sem` | If you blur ("ghost") the predicted region, the evidence for the object disappears — the model should notice | Reward `1` only if the object is judged *visible* on the original image and *not visible* after ghosting |
+
+The composite reward `R = 0.5·R_geo + 0.5·R_sem` is optimized with
+**KL-regularized REINFORCE** against a frozen reference policy. Mechanistically,
+VISE **increases generation-time attention to visual tokens** across decoder
+layers (+2.84% on 2B, +2.56% on 4B), shifting decoding from language-prior-driven
+to image-conditioned.
+
+---
+
+## Results (Qwen3-VL-2B base)
+
+Δ = absolute change vs. the base model.
+
+**Image captioning (CIDEr):**
+
+| Benchmark | Base | VISE | Δ |
+| --- | --- | --- | --- |
+| COCO      | 21.54 | **38.39** | **+16.85** |
+| NoCaps    | 19.52 | **34.25** | **+14.73** |
+| Flickr30k | 26.09 | **42.64** | **+16.55** |
+| TextCaps  | 22.20 | **41.86** | **+19.66** |
+
+**Hallucination & reasoning (selected):**
+
+| Metric | Base | VISE | Δ |
+| --- | --- | --- | --- |
+| CHAIR-I ↓ | — | — | **−5.00** |
+| CHAIR-S ↓ | — | — | **−5.45** |
+| POPE ↑    | — | — | **+1.02** |
+| ScienceQA | — | — | **+4.19** |
+| InfoVQA   | — | — | **+2.41** |
+| MMMU      | — | — | **+1.75** |
+
+VISE improves **all 18 benchmarks** (captioning, VQA, reasoning, hallucination)
+with **no task tradeoffs**, and generalizes across four scales (2B/4B/8B/32B) and
+four backbone families (Qwen3-VL, InternVL3, Gemma-3, Llama-3.2-Vision). See the
+paper for the full tables.
+
+---
+
+## How to use
+
+This is a **LoRA adapter**, so you load the base model first and then attach the
+adapter with PEFT.
+
+```bash
+pip install "transformers>=4.57" peft accelerate pillow torch
+```
+
+```python
+import torch
+from PIL import Image
+from transformers import AutoModelForVision2Seq, AutoProcessor
+from peft import PeftModel
+
+BASE = "Qwen/Qwen3-VL-2B-Instruct"
+ADAPTER = "shravvvv/VISE"
+
+# 1) Load the base model + processor
+model = AutoModelForVision2Seq.from_pretrained(
+    BASE, torch_dtype=torch.bfloat16, device_map="auto"
+)
+processor = AutoProcessor.from_pretrained(ADAPTER)  # adapter ships the processor too
+
+# 2) Attach the VISE LoRA adapter
+model = PeftModel.from_pretrained(model, ADAPTER)
+model.eval()
+
+# 3) Run inference (e.g. captioning / grounding)
+image = Image.open("example.jpg").convert("RGB")
+messages = [{
+    "role": "user",
+    "content": [
+        {"type": "image", "image": image},
+        {"type": "text", "text": "Describe this image in detail."},
+    ],
+}]
+
+text = processor.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
+inputs = processor(text=[text], images=[image], return_tensors="pt").to(model.device)
+
+with torch.inference_mode():
+    out = model.generate(**inputs, max_new_tokens=128)
+gen = out[:, inputs["input_ids"].shape[1]:]
+print(processor.batch_decode(gen, skip_special_tokens=True)[0])
+```
+
+> **Tip — merge for faster inference:** call
+> `model = model.merge_and_unload()` after loading the adapter to fold the LoRA
+> weights into the base model.
+
+### Visual grounding prompt
+
+VISE is self-evolved through a grounding objective and also produces normalized
+`[0, 1000]` boxes. To localize an object, prompt:
+
+```text
+Locate the object described in the query and output the bounding box in
+normalized [0, 1000] coordinates as: <box>x1,y1,x2,y2</box>
+Query: the red car in the center
+```
+
+---
+
+## Training details
+
+| Setting | Value |
+| --- | --- |
+| Base model | `Qwen/Qwen3-VL-2B-Instruct` (vision encoder frozen) |
+| Adapter | LoRA · `r=16`, `α=32`, `dropout=0.05` |
+| LoRA targets | `q_proj, k_proj, v_proj, o_proj, gate_proj, up_proj, down_proj, mm_projector` |
+| Optimizer | AdamW · `lr=1e-6`, `weight_decay=0.01`, grad clip `1.0` |
+| RL objective | KL-regularized REINFORCE (adaptive β, target KL `0.020`, rate `0.10`) |
+| Reward weights | `λ_geo = λ_sem = 0.5` |
+| Transforms | affine (rot ±10°, scale 0.9–1.1, translate ±50px), crop (0.8–1.0), flip |
+| Ghosting | Gaussian blur, `σ = 25` |
+| Precision | bfloat16 |
+| Training data | 4,000 **raw, unlabeled** COCO images (no captions/boxes/labels) |
+| Hardware | 8× AMD MI250X |
+
+No question–answer pairs, annotations, metadata, or external reward models are
+used at any point.
+
+---
+
+## Intended use & limitations
+
+- **Intended use:** research on self-evolving / unsupervised post-training of
+  LMMs, visual grounding, image captioning, VQA, and hallucination reduction.
+- **Limitations:** gains are largest at smaller scales and attenuate as the base
+  model grows (stronger pretrained conditioning leaves less headroom). The
+  adapter inherits the biases and failure modes of the `Qwen3-VL-2B-Instruct`
+  base model. As with any LMM, outputs may be incorrect and should not be relied
+  upon for safety-critical decisions.
+
+## License
+
+Released under the **Apache 2.0** license, consistent with the
+`Qwen3-VL-2B-Instruct` base model. You must also comply with the base model's
+license terms.
+
+## Citation
+
+```bibtex
+@inproceedings{venkatraman2026vise,
+  title     = {Paying More Attention to Visual Tokens in Self-Evolving Large Multimodal Models},
+  author    = {Venkatraman, Shravan and Thawkar, Ritesh and Thawakar, Omkar and
+               Anwer, Rao Muhammad and Cholakkal, Hisham and Khan, Salman and Khan, Fahad Shahbaz},
+  booktitle = {European Conference on Computer Vision (ECCV)},
+  year      = {2026}
+}
+```
+
+### Framework versions
+- PEFT 0.18.1

adapter_config.json

@@ -0,0 +1,47 @@
+{
+  "alora_invocation_tokens": null,
+  "alpha_pattern": {},
+  "arrow_config": null,
+  "auto_mapping": null,
+  "base_model_name_or_path": "Qwen/Qwen3-VL-2B-Instruct",
+  "bias": "none",
+  "corda_config": null,
+  "ensure_weight_tying": false,
+  "eva_config": null,
+  "exclude_modules": null,
+  "fan_in_fan_out": false,
+  "inference_mode": true,
+  "init_lora_weights": true,
+  "layer_replication": null,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "loftq_config": {},
+  "lora_alpha": 32,
+  "lora_bias": false,
+  "lora_dropout": 0.05,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "peft_type": "LORA",
+  "peft_version": "0.18.1",
+  "qalora_group_size": 16,
+  "r": 16,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "down_proj",
+    "up_proj",
+    "gate_proj",
+    "v_proj",
+    "mm_projector",
+    "k_proj",
+    "o_proj",
+    "q_proj"
+  ],
+  "target_parameters": null,
+  "task_type": "CAUSAL_LM",
+  "trainable_token_indices": null,
+  "use_dora": false,
+  "use_qalora": false,
+  "use_rslora": false
+}

adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:23dbee1b3cdb98cbe471a46bae69fbf680cc8c5be679c6a1e567fb7c8032e920
+size 69788264

added_tokens.json

@@ -0,0 +1,28 @@
+{
+  "</think>": 151668,
+  "</tool_call>": 151658,
+  "</tool_response>": 151666,
+  "<think>": 151667,
+  "<tool_call>": 151657,
+  "<tool_response>": 151665,
+  "<|box_end|>": 151649,
+  "<|box_start|>": 151648,
+  "<|endoftext|>": 151643,
+  "<|file_sep|>": 151664,
+  "<|fim_middle|>": 151660,
+  "<|fim_pad|>": 151662,
+  "<|fim_prefix|>": 151659,
+  "<|fim_suffix|>": 151661,
+  "<|im_end|>": 151645,
+  "<|im_start|>": 151644,
+  "<|image_pad|>": 151655,
+  "<|object_ref_end|>": 151647,
+  "<|object_ref_start|>": 151646,
+  "<|quad_end|>": 151651,
+  "<|quad_start|>": 151650,
+  "<|repo_name|>": 151663,
+  "<|video_pad|>": 151656,
+  "<|vision_end|>": 151653,
+  "<|vision_pad|>": 151654,
+  "<|vision_start|>": 151652
+}

chat_template.jinja

@@ -0,0 +1,120 @@
+{%- if tools %}
+    {{- '<|im_start|>system\n' }}
+    {%- if messages[0].role == 'system' %}
+        {%- if messages[0].content is string %}
+            {{- messages[0].content }}
+        {%- else %}
+            {%- for content in messages[0].content %}
+                {%- if 'text' in content %}
+                    {{- content.text }}
+                {%- endif %}
+            {%- endfor %}
+        {%- endif %}
+        {{- '\n\n' }}
+    {%- endif %}
+    {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>" }}
+    {%- for tool in tools %}
+        {{- "\n" }}
+        {{- tool | tojson }}
+    {%- endfor %}
+    {{- "\n</tools>\n\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call><|im_end|>\n" }}
+{%- else %}
+    {%- if messages[0].role == 'system' %}
+        {{- '<|im_start|>system\n' }}
+        {%- if messages[0].content is string %}
+            {{- messages[0].content }}
+        {%- else %}
+            {%- for content in messages[0].content %}
+                {%- if 'text' in content %}
+                    {{- content.text }}
+                {%- endif %}
+            {%- endfor %}
+        {%- endif %}
+        {{- '<|im_end|>\n' }}
+    {%- endif %}
+{%- endif %}
+{%- set image_count = namespace(value=0) %}
+{%- set video_count = namespace(value=0) %}
+{%- for message in messages %}
+    {%- if message.role == "user" %}
+        {{- '<|im_start|>' + message.role + '\n' }}
+        {%- if message.content is string %}
+            {{- message.content }}
+        {%- else %}
+            {%- for content in message.content %}
+                {%- if content.type == 'image' or 'image' in content or 'image_url' in content %}
+                    {%- set image_count.value = image_count.value + 1 %}
+                    {%- if add_vision_id %}Picture {{ image_count.value }}: {% endif -%}
+                    <|vision_start|><|image_pad|><|vision_end|>
+                {%- elif content.type == 'video' or 'video' in content %}
+                    {%- set video_count.value = video_count.value + 1 %}
+                    {%- if add_vision_id %}Video {{ video_count.value }}: {% endif -%}
+                    <|vision_start|><|video_pad|><|vision_end|>
+                {%- elif 'text' in content %}
+                    {{- content.text }}
+                {%- endif %}
+            {%- endfor %}
+        {%- endif %}
+        {{- '<|im_end|>\n' }}
+    {%- elif message.role == "assistant" %}
+        {{- '<|im_start|>' + message.role + '\n' }}
+        {%- if message.content is string %}
+            {{- message.content }}
+        {%- else %}
+            {%- for content_item in message.content %}
+                {%- if 'text' in content_item %}
+                    {{- content_item.text }}
+                {%- endif %}
+            {%- endfor %}
+        {%- endif %}
+        {%- if message.tool_calls %}
+            {%- for tool_call in message.tool_calls %}
+                {%- if (loop.first and message.content) or (not loop.first) %}
+                    {{- '\n' }}
+                {%- endif %}
+                {%- if tool_call.function %}
+                    {%- set tool_call = tool_call.function %}
+                {%- endif %}
+                {{- '<tool_call>\n{"name": "' }}
+                {{- tool_call.name }}
+                {{- '", "arguments": ' }}
+                {%- if tool_call.arguments is string %}
+                    {{- tool_call.arguments }}
+                {%- else %}
+                    {{- tool_call.arguments | tojson }}
+                {%- endif %}
+                {{- '}\n</tool_call>' }}
+            {%- endfor %}
+        {%- endif %}
+        {{- '<|im_end|>\n' }}
+    {%- elif message.role == "tool" %}
+        {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+            {{- '<|im_start|>user' }}
+        {%- endif %}
+        {{- '\n<tool_response>\n' }}
+        {%- if message.content is string %}
+            {{- message.content }}
+        {%- else %}
+            {%- for content in message.content %}
+                {%- if content.type == 'image' or 'image' in content or 'image_url' in content %}
+                    {%- set image_count.value = image_count.value + 1 %}
+                    {%- if add_vision_id %}Picture {{ image_count.value }}: {% endif -%}
+                    <|vision_start|><|image_pad|><|vision_end|>
+                {%- elif content.type == 'video' or 'video' in content %}
+                    {%- set video_count.value = video_count.value + 1 %}
+                    {%- if add_vision_id %}Video {{ video_count.value }}: {% endif -%}
+                    <|vision_start|><|video_pad|><|vision_end|>
+                {%- elif 'text' in content %}
+                    {{- content.text }}
+                {%- endif %}
+            {%- endfor %}
+        {%- endif %}
+        {{- '\n</tool_response>' }}
+        {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+            {{- '<|im_end|>\n' }}
+        {%- endif %}
+    {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+    {{- '<|im_start|>assistant\n' }}
+{%- endif %}

preprocessor_config.json

@@ -0,0 +1,39 @@
+{
+  "crop_size": null,
+  "data_format": "channels_first",
+  "default_to_square": true,
+  "device": null,
+  "disable_grouping": null,
+  "do_center_crop": null,
+  "do_convert_rgb": true,
+  "do_normalize": true,
+  "do_pad": null,
+  "do_rescale": true,
+  "do_resize": true,
+  "image_mean": [
+    0.5,
+    0.5,
+    0.5
+  ],
+  "image_processor_type": "Qwen2VLImageProcessorFast",
+  "image_std": [
+    0.5,
+    0.5,
+    0.5
+  ],
+  "input_data_format": null,
+  "max_pixels": null,
+  "merge_size": 2,
+  "min_pixels": null,
+  "pad_size": null,
+  "patch_size": 16,
+  "processor_class": "Qwen3VLProcessor",
+  "resample": 3,
+  "rescale_factor": 0.00392156862745098,
+  "return_tensors": null,
+  "size": {
+    "longest_edge": 16777216,
+    "shortest_edge": 65536
+  },
+  "temporal_patch_size": 2
+}

special_tokens_map.json

@@ -0,0 +1,31 @@
+{
+  "additional_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>",
+    "<|object_ref_start|>",
+    "<|object_ref_end|>",
+    "<|box_start|>",
+    "<|box_end|>",
+    "<|quad_start|>",
+    "<|quad_end|>",
+    "<|vision_start|>",
+    "<|vision_end|>",
+    "<|vision_pad|>",
+    "<|image_pad|>",
+    "<|video_pad|>"
+  ],
+  "eos_token": {
+    "content": "<|im_end|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": {
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

tokenizer.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:aeb13307a71acd8fe81861d94ad54ab689df773318809eed3cbe794b4492dae4
+size 11422654

tokenizer_config.json

@@ -0,0 +1,240 @@
+{
+  "add_bos_token": false,
+  "add_prefix_space": false,
+  "added_tokens_decoder": {
+    "151643": {
+      "content": "<|endoftext|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151644": {
+      "content": "<|im_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151645": {
+      "content": "<|im_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151646": {
+      "content": "<|object_ref_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151647": {
+      "content": "<|object_ref_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151648": {
+      "content": "<|box_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151649": {
+      "content": "<|box_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151650": {
+      "content": "<|quad_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151651": {
+      "content": "<|quad_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151652": {
+      "content": "<|vision_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151653": {
+      "content": "<|vision_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151654": {
+      "content": "<|vision_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151655": {
+      "content": "<|image_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151656": {
+      "content": "<|video_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151657": {
+      "content": "<tool_call>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151658": {
+      "content": "</tool_call>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151659": {
+      "content": "<|fim_prefix|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151660": {
+      "content": "<|fim_middle|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151661": {
+      "content": "<|fim_suffix|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151662": {
+      "content": "<|fim_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151663": {
+      "content": "<|repo_name|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151664": {
+      "content": "<|file_sep|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151665": {
+      "content": "<tool_response>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151666": {
+      "content": "</tool_response>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151667": {
+      "content": "<think>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151668": {
+      "content": "</think>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    }
+  },
+  "additional_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>",
+    "<|object_ref_start|>",
+    "<|object_ref_end|>",
+    "<|box_start|>",
+    "<|box_end|>",
+    "<|quad_start|>",
+    "<|quad_end|>",
+    "<|vision_start|>",
+    "<|vision_end|>",
+    "<|vision_pad|>",
+    "<|image_pad|>",
+    "<|video_pad|>"
+  ],
+  "bos_token": null,
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|im_end|>",
+  "errors": "replace",
+  "extra_special_tokens": {},
+  "model_max_length": 262144,
+  "pad_token": "<|endoftext|>",
+  "processor_class": "Qwen3VLProcessor",
+  "split_special_tokens": false,
+  "tokenizer_class": "Qwen2Tokenizer",
+  "unk_token": null
+}

video_preprocessor_config.json

@@ -0,0 +1,41 @@
+{
+  "crop_size": null,
+  "data_format": "channels_first",
+  "default_to_square": true,
+  "device": null,
+  "do_center_crop": null,
+  "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
+  ],
+  "input_data_format": null,
+  "max_frames": 768,
+  "merge_size": 2,
+  "min_frames": 4,
+  "num_frames": null,
+  "pad_size": null,
+  "patch_size": 16,
+  "processor_class": "Qwen3VLProcessor",
+  "resample": 3,
+  "rescale_factor": 0.00392156862745098,
+  "return_metadata": false,
+  "size": {
+    "longest_edge": 25165824,
+    "shortest_edge": 4096
+  },
+  "temporal_patch_size": 2,
+  "video_metadata": null,
+  "video_processor_type": "Qwen3VLVideoProcessor"
+}