Add fine-tuned Qwen3.5-2B distill-structure model with Gradio demo

.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 +1,95 @@
+---
+language:
+- en
+license: apache-2.0
+base_model: Qwen/Qwen3.5-2B
+tags:
+- html
+- structure-analysis
+- information-extraction
+- web-scraping
+- lora
+- fine-tuned
+pipeline_tag: text-generation
+---
+
+# distill-structure
+
+A fine-tuned [Qwen3.5-2B](https://huggingface.co/Qwen/Qwen3.5-2B) model for **HTML structure analysis** — given a compact DOM representation of a web page, it identifies the logical sections and outputs structured JSON.
+
+## What it does
+
+Takes a cleaned, heading-stripped HTML page and returns a JSON array describing its sections:
+
+```json
+[
+  {
+    "title": "Main News Feed Content",
+    "start_text": "1. Canada's bill C-22 mandates...",
+    "content_type": "article",
+    "assets": [{"type": "link", "value": "Canada's bill C-22..."}]
+  },
+  {
+    "title": "Site Footer Navigation",
+    "start_text": "Guidelines | FAQ | Lists",
+    "content_type": "footer",
+    "assets": []
+  }
+]
+```
+
+## Use case
+
+This model powers the `StructureAgent` inside the [distill](https://github.com/nahidstaq/distill) pipeline — it handles pages with **no heading tags** where rule-based sectioning fails. The model is trained to recover section structure that headings would normally provide.
+
+## Training
+
+- **Base model**: `Qwen/Qwen3.5-2B`
+- **Method**: LoRA fine-tuning (r=32, α=64) via TRL SFTTrainer
+- **Dataset**: ~3,455 training / 384 eval examples generated from heading-rich web pages (headings stripped and used as labels)
+- **Epochs**: 3 — Train loss: 1.009 — Token accuracy: 80.5%
+
+## Quick start
+
+```python
+from transformers import AutoModelForCausalLM, AutoTokenizer
+import torch, json
+
+model_id = "nahidstaq/distill-structure"
+tokenizer = AutoTokenizer.from_pretrained(model_id)
+model = AutoModelForCausalLM.from_pretrained(model_id, dtype=torch.bfloat16, device_map="auto")
+
+SYSTEM = (
+    "You are an HTML structure analyzer. Given a compact DOM representation "
+    "of a web page (with headings removed), identify the logical sections. "
+    "Output a JSON array of sections, each with title, start_text, content_type, and assets fields."
+)
+
+def analyze(page_title: str, compact_dom: str) -> list[dict]:
+    messages = [
+        {"role": "system", "content": SYSTEM},
+        {"role": "user", "content": f"Page: {page_title}\n\n{compact_dom}"},
+    ]
+    prompt = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
+    inputs = tokenizer(prompt, return_tensors="pt").to(model.device)
+    with torch.no_grad():
+        ids = model.generate(**inputs, max_new_tokens=512, do_sample=False,
+                             pad_token_id=tokenizer.eos_token_id)
+    raw = tokenizer.decode(ids[0][inputs["input_ids"].shape[1]:], skip_special_tokens=True)
+    return json.loads(raw)
+```
+
+## Output fields
+
+| Field | Description |
+|---|---|
+| `title` | Short descriptive section title |
+| `start_text` | First ~50 chars of the section's text (for anchoring) |
+| `content_type` | One of: `article`, `list`, `hero`, `navigation`, `footer`, `table`, `faq`, `other` |
+| `assets` | Extracted links, images, or list items relevant to the section |
+
+## Limitations
+
+- Works best on English pages
+- Table-heavy layouts (e.g. nested `<td>`) may collapse into fewer sections
+- `content_type` classification skews toward `other` for ambiguous sections

app.py

@@ -0,0 +1,266 @@
+"""Gradio demo for distill-structure model."""
+
+import json
+import re
+
+import gradio as gr
+import torch
+from transformers import AutoModelForCausalLM, AutoTokenizer
+
+# ---------------------------------------------------------------------------
+# Model
+# ---------------------------------------------------------------------------
+
+MODEL_ID = "nahidstaq/distill-structure"
+
+SYSTEM = (
+    "You are an HTML structure analyzer. Given a compact DOM representation "
+    "of a web page (with headings removed), identify the logical sections. "
+    "Output a JSON array of sections, each with title, start_text, content_type, and assets fields."
+)
+
+_model = None
+_tokenizer = None
+
+
+def _load():
+    global _model, _tokenizer
+    if _model is None:
+        device = "cuda" if torch.cuda.is_available() else "cpu"
+        dtype = torch.bfloat16 if device == "cuda" else torch.float32
+        _tokenizer = AutoTokenizer.from_pretrained(MODEL_ID)
+        _model = AutoModelForCausalLM.from_pretrained(
+            MODEL_ID, dtype=dtype, device_map="auto"
+        )
+        _model.eval()
+    return _model, _tokenizer
+
+
+# ---------------------------------------------------------------------------
+# Helpers
+# ---------------------------------------------------------------------------
+
+def _compact_dom(html: str) -> str:
+    from lxml import html as lxml_html
+
+    try:
+        doc = lxml_html.fromstring(html)
+    except Exception:
+        return html[:3000]
+
+    for tag in ("h1", "h2", "h3", "h4", "h5", "h6", "script", "style", "head"):
+        for el in doc.findall(f".//{tag}"):
+            p = el.getparent()
+            if p is not None:
+                p.remove(el)
+
+    def _walk(el, depth=0):
+        if not hasattr(el, "tag") or not isinstance(el.tag, str):
+            return ""
+        tag = el.tag
+        indent = "  " * depth
+
+        if tag == "img":
+            alt = el.get("alt", "")
+            return f'{indent}<img alt="{alt}">' if alt else f'{indent}<img>'
+
+        if tag == "a":
+            text = (el.text_content() or "").strip()[:40]
+            href = (el.get("href") or "")[:60]
+            return f'{indent}<a href="{href}"> {text}'
+
+        if tag in ("td", "th"):
+            # Recurse into td if it has block children, otherwise truncate
+            children = [c for c in el if hasattr(c, "tag") and isinstance(c.tag, str)]
+            if children and depth < 8:
+                lines = [f"{indent}<{tag}>"]
+                for child in children:
+                    r = _walk(child, depth + 1)
+                    if r:
+                        lines.append(r)
+                return "\n".join(lines)
+            text = (el.text_content() or "").strip()[:60]
+            return f"{indent}<{tag}> {text}" if text else ""
+
+        if depth > 7:
+            text = (el.text_content() or "").strip()[:80]
+            return f"{indent}[... {text}...]" if text else ""
+
+        text = (el.text or "").strip()[:50]
+        attrs = ""
+        for a in ("id", "class", "role"):
+            v = el.get(a)
+            if v:
+                attrs += f' {a}="{v[:30]}"'
+
+        line = f"{indent}<{tag}{attrs}>"
+        if text:
+            line += f" {text}"
+        lines = [line]
+        for child in el:
+            r = _walk(child, depth + 1)
+            if r:
+                lines.append(r)
+        return "\n".join(lines)
+
+    body = doc.find(".//body") or doc
+    result = _walk(body)
+    # Truncate to 4096 chars
+    if len(result) > 4096:
+        result = result[:4096] + "\n... (truncated)"
+    return result
+
+
+def _extract_title(html: str) -> str:
+    m = re.search(r"<title>(.*?)</title>", html, re.I | re.S)
+    return m.group(1).strip() if m else "Untitled"
+
+
+def _parse(raw: str) -> list[dict]:
+    try:
+        data = json.loads(raw)
+        if isinstance(data, list):
+            return data
+    except json.JSONDecodeError:
+        pass
+    m = re.search(r"\[.*?\]", raw, re.S)
+    if m:
+        try:
+            return json.loads(m.group())
+        except json.JSONDecodeError:
+            pass
+    return []
+
+
+# ---------------------------------------------------------------------------
+# Inference
+# ---------------------------------------------------------------------------
+
+def analyze_html(html: str, page_title: str) -> tuple[str, str]:
+    if not html.strip():
+        return "Please paste some HTML.", ""
+
+    model, tokenizer = _load()
+
+    compact = _compact_dom(html)
+    title = page_title.strip() or _extract_title(html)
+
+    messages = [
+        {"role": "system", "content": SYSTEM},
+        {"role": "user", "content": f"Page: {title}\n\n{compact}"},
+    ]
+    prompt = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
+    inputs = tokenizer(prompt, return_tensors="pt").to(model.device)
+
+    with torch.no_grad():
+        ids = model.generate(
+            **inputs,
+            max_new_tokens=512,
+            do_sample=False,
+            temperature=None,
+            top_p=None,
+            pad_token_id=tokenizer.eos_token_id,
+        )
+
+    raw = tokenizer.decode(ids[0][inputs["input_ids"].shape[1]:], skip_special_tokens=True).strip()
+    sections = _parse(raw)
+
+    # Pretty format sections as markdown table
+    if sections:
+        md = "| # | Type | Title | Start text |\n|---|---|---|---|\n"
+        for i, s in enumerate(sections, 1):
+            title_s = s.get("title", "")
+            ctype = s.get("content_type", "?")
+            start = (s.get("start_text") or "")[:50]
+            md += f"| {i} | `{ctype}` | {title_s} | {start} |\n"
+    else:
+        md = "_Could not parse sections from model output._"
+
+    return md, raw
+
+
+def analyze_url(url: str) -> tuple[str, str, str]:
+    if not url.strip():
+        return "", "Please enter a URL.", ""
+    try:
+        import httpx
+        r = httpx.get(url, follow_redirects=True, timeout=10,
+                      headers={"User-Agent": "Mozilla/5.0"})
+        html = r.text
+        title = _extract_title(html)
+        md, raw = analyze_html(html, title)
+        return html[:5000] + ("..." if len(html) > 5000 else ""), md, raw
+    except Exception as e:
+        return "", f"Error fetching URL: {e}", ""
+
+
+# ---------------------------------------------------------------------------
+# UI
+# ---------------------------------------------------------------------------
+
+EXAMPLE_HTML = """<!DOCTYPE html>
+<html>
+<head><title>Product Page</title></head>
+<body>
+  <h1>Our Amazing Product</h1>
+  <p>Welcome to the best product you've ever seen.</p>
+  <h2>Features</h2>
+  <ul>
+    <li>Lightning fast</li>
+    <li>Easy to use</li>
+    <li>Affordable pricing</li>
+  </ul>
+  <h2>Pricing</h2>
+  <table>
+    <tr><th>Plan</th><th>Price</th></tr>
+    <tr><td>Starter</td><td>$9/mo</td></tr>
+    <tr><td>Pro</td><td>$29/mo</td></tr>
+  </table>
+  <h2>FAQ</h2>
+  <h3>Is there a free trial?</h3>
+  <p>Yes! 14 days free, no credit card required.</p>
+</body>
+</html>"""
+
+with gr.Blocks(title="distill-structure", theme=gr.themes.Soft()) as demo:
+    gr.Markdown("# distill-structure\nHTML section analyzer — fine-tuned Qwen3.5-2B")
+
+    with gr.Tabs():
+        with gr.Tab("Paste HTML"):
+            with gr.Row():
+                with gr.Column():
+                    html_input = gr.Textbox(
+                        label="HTML",
+                        placeholder="Paste HTML here...",
+                        lines=15,
+                        value=EXAMPLE_HTML,
+                    )
+                    title_input = gr.Textbox(label="Page title (optional)", placeholder="Auto-detected from <title>")
+                    btn_html = gr.Button("Analyze", variant="primary")
+                with gr.Column():
+                    sections_out = gr.Markdown(label="Sections")
+                    raw_out = gr.Textbox(label="Raw JSON output", lines=10)
+
+            btn_html.click(analyze_html, inputs=[html_input, title_input], outputs=[sections_out, raw_out])
+
+        with gr.Tab("From URL"):
+            with gr.Row():
+                with gr.Column():
+                    url_input = gr.Textbox(label="URL", placeholder="https://news.ycombinator.com")
+                    btn_url = gr.Button("Fetch & Analyze", variant="primary")
+                    html_preview = gr.Textbox(label="Fetched HTML (preview)", lines=8)
+                with gr.Column():
+                    sections_out2 = gr.Markdown(label="Sections")
+                    raw_out2 = gr.Textbox(label="Raw JSON output", lines=10)
+
+            btn_url.click(analyze_url, inputs=[url_input], outputs=[html_preview, sections_out2, raw_out2])
+
+    gr.Markdown("""
+---
+**Model**: [nahidstaq/distill-structure](https://huggingface.co/nahidstaq/distill-structure) ·
+**Base**: Qwen3.5-2B ·
+**Task**: HTML structure analysis
+""")
+
+if __name__ == "__main__":
+    demo.launch()

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 +1,75 @@
+{
+  "architectures": [
+    "Qwen3_5ForCausalLM"
+  ],
+  "attention_bias": false,
+  "attention_dropout": 0.0,
+  "attn_output_gate": true,
+  "bos_token_id": null,
+  "dtype": "bfloat16",
+  "eos_token_id": 248046,
+  "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,
+  "transformers_version": "5.3.0",
+  "use_cache": false,
+  "vocab_size": 248320
+}

generation_config.json

@@ -0,0 +1,10 @@
+{
+  "_from_model_config": true,
+  "eos_token_id": [
+    248046,
+    248044
+  ],
+  "pad_token_id": 248044,
+  "transformers_version": "5.3.0",
+  "use_cache": true
+}

model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0ac8fe6e5c61f34e58a1247604e167a91dd5cacd984ff7798d8cebedc2fe9fed
+size 3763692048

tokenizer.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:87a7830d63fcf43bf241c3c5242e96e62dd3fdc29224ca26fed8ea333db72de4
+size 19989343

tokenizer_config.json

@@ -0,0 +1,31 @@
+{
+  "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": false,
+  "model_max_length": 4096,
+  "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+",
+  "split_special_tokens": false,
+  "tokenizer_class": "TokenizersBackend",
+  "unk_token": null,
+  "video_token": "<|video_pad|>",
+  "vision_bos_token": "<|vision_start|>",
+  "vision_eos_token": "<|vision_end|>"
+}