the first uploading

.gitattributes

@@ -33,3 +33,5 @@ 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
+unimo-context/tokenizer.json filter=lfs diff=lfs merge=lfs -text
+image/*.png filter=lfs diff=lfs merge=lfs -text

.gitignore

@@ -0,0 +1 @@
+unitale_env/

README.md

@@ -1,3 +1,45 @@
 ---
 license: agpl-3.0
+base_model: Qwen/Qwen3-0.6B
+pipeline_tag: text-to-speech
+tags:
+- emotion-
+- emotion-recognition
+- tts
+- unimo
 ---
+
+
+# Unimo IndexTTS2情绪描述文本解析优化模型
+
+Unimo-context 是一个情绪上下文模型，能够根据情绪的上下文（如上一句话的情绪向量）以及当前的情绪描述，综合情绪的上下文信息后生成更加自然的情绪，使情绪的变化更加丝滑流畅
+
+基于 Qwen3-0.6B 微调的轻量级情感特征工程模型，专门用于解决多轮对话中TTS语音合成情感表现过于突变、缺乏物理惯性的问题。
+
+目的是为解决在传统的情感 TTS 流程中，情感随文本瞬间切换（例如从愤怒直接跳到冷静）过渡不自然的问题。
+
+Unimo-context 充当了情感变化中的“阻尼器”或“润滑剂”：它接收上一句的情绪历史状态和当前句的情绪意图，输出较符合人类习惯的情绪惯性残留 的混合情绪向量，8维情绪特征向量可直接输入IndexTTS2使用合成音频。
+
+本项目可配合另一个项目:https://huggingface.co/sdsds222moyu/Unimo-indexTTS2-emotext 使用，这个项目可以更精确地解析自然语言的情绪指令文本，转化为IndexTTS2可用的情绪向量。
+
+
+## 部署方法：
+安装依赖：
+```
+python -m venv venv
+source venv/bin/activate  # Linux
+.\venv\Scripts\activate   # Windows
+pip install -r requirements.txt
+```
+
+启动情绪测试对比界面（可以生成上一句的情绪音频，当前的情绪音频和混合后的情绪音频，以供对比参考）：
+```
+python Unimo_test_qwen2.5.py
+```
+```
+python Unimo_test_qwen3.py
+```
+
+![image](image/1.png)
+![image](image/2.png)
+![image](image/2.png)

Unimo_context_qwen3.py

@@ -0,0 +1,210 @@
+import gradio as gr
+import torch
+import json
+import re
+import os
+import requests
+from transformers import AutoModelForCausalLM, AutoTokenizer
+
+# --- 0. 全局配置 ---
+modelname = 'unimo-context'  # 你的“完全体”模型文件夹名
+
+# --- 简洁 CSS：仅保留核心换行逻辑 ---
+custom_css = """
+/* 向量框：强制换行，确保长 JSON 自动折行 */
+.wrap-json-code .cm-content, .wrap-json-code pre {
+    white-space: pre-wrap !important;
+    word-break: break-all !important;
+}
+.wrap-json-code .cm-editor {
+    border: 1px solid #ddd !important;
+}
+
+/* 容器样式：加深视觉区隔 */
+.test-row {
+    border: 1px solid #e0e0e0;
+    border-radius: 12px;
+    padding: 24px !important;
+    margin-bottom: 35px !important;
+    background-color: transparent !important;
+}
+
+/* 标题样式：更具层次感 */
+.test-row h4 {
+    border-left: 4px solid #4A90E2;
+    padding-left: 10px;
+    margin-bottom: 15px !important;
+}
+"""
+
+# --- 1. 情感引擎 (逻辑不变) ---
+class EmotionEngine:
+    def __init__(self, model_folder_name):
+        current_dir = os.path.dirname(os.path.abspath(__file__))
+        model_path = os.path.join(current_dir, model_folder_name)
+        print(f"--- 引擎初始化 ---")
+        self.tokenizer = AutoTokenizer.from_pretrained(model_path, trust_remote_code=True)
+        self.model = AutoModelForCausalLM.from_pretrained(
+            model_path, dtype=torch.float16, device_map="cuda", trust_remote_code=True
+        )
+        self.model.eval()
+        print(f"Unimo 引擎就绪")
+
+    def predict_context(self, prev_v, intent_v):
+        instruction_str = "根据上句情感状态和此句意图，计算并输出符合心理阻尼逻辑的实操结果向量。"
+        inner_input_data = {"上句状态": prev_v, "此句意图": intent_v}
+        inner_input_str = json.dumps(inner_input_data, ensure_ascii=False)
+        prompt_obj = {"instruction": instruction_str, "input": inner_input_str}
+        full_text = json.dumps(prompt_obj, ensure_ascii=False, indent=0) 
+        
+        print(f"\n[模型调用] 正在请求阻尼计算...")
+        messages = [{"role": "user", "content": full_text}]
+        prompt = self.tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True, enable_thinking=False)
+        
+        inputs = self.tokenizer(prompt, return_tensors="pt").to("cuda")
+        with torch.no_grad():
+            outputs = self.model.generate(**inputs, max_new_tokens=256, do_sample=False)
+        
+        response = self.tokenizer.decode(outputs[0][len(inputs["input_ids"][0]):], skip_special_tokens=True).strip()
+        print(f"[模型输出] 原始响应: {response}")
+        
+        try:
+            json_match = re.search(r'\{.*\}', response, re.DOTALL)
+            vector_res = json.loads(json_match.group())
+            return vector_res.get("实操结果", vector_res)
+        except Exception as e:
+            print(f"[解析失败] 响应格式不匹配: {e}")
+            return intent_v
+
+# --- 2. 接口调用 (增加防御性检查) ---
+def call_indextts(text, emotion_config, ref_audio, base_url):
+    print(f"\n[TTS 链路] 开始处理任务...")
+    if not base_url: return None, "URL缺失"
+    if not ref_audio: 
+        print("⚠️ 失败: 请先在上方上传全局参考音频")
+        return None, "参考音频缺失"
+    
+    base_url = base_url.rstrip('/')
+    ref_filename = os.path.basename(ref_audio)
+    
+    try:
+        check_res = requests.get(f"{base_url}/v1/check/audio", params={"file_name": ref_filename}, timeout=5)
+        if check_res.status_code == 404:
+            print(f"[上传中] 发送参考音频: {ref_filename}")
+            with open(ref_audio, "rb") as f:
+                requests.post(f"{base_url}/v1/upload_audio", files={"audio": (ref_filename, f)}, data={"full_path": ref_filename}, timeout=30)
+        
+        payload = {"text": text, "audio_path": ref_filename}
+        if isinstance(emotion_config, dict):
+            keys = ["高兴", "愤怒", "悲伤", "恐惧", "反感", "低落", "惊讶", "自然"]
+            payload["emo_vector"] = [float(emotion_config.get(k, 0.0)) for k in keys]
+        else: payload["emo_text"] = emotion_config
+
+        s_res = requests.post(f"{base_url}/v2/synthesize", json=payload, timeout=60)
+        if s_res.status_code == 200:
+            out_path = f"output_{hash(text + str(emotion_config))}.wav"
+            with open(out_path, "wb") as f: f.write(s_res.content)
+            print(f"[成功] 音频生成完毕")
+            return out_path, "成功"
+        print(f"[失败] 状态码: {s_res.status_code}")
+        return None, "失败"
+    except Exception as e:
+        print(f"[报错] {str(e)}")
+        return None, "报错"
+
+# 初始化引擎
+engine = EmotionEngine(model_folder_name=modelname)
+
+# --- 3. Gradio 界面 ---
+DEFAULT_V = {"高兴": 0.0, "愤怒": 0.0, "悲伤": 0.0, "恐惧": 0.0, "反感": 0.0, "低落": 0.0, "惊讶": 0.0, "自然": 1.0}
+
+def get_default_row():
+    return {
+        "txt_p": "（前序台词）",
+        "txt_i": "刚才的事，我还没找你算账呢！",
+        "txt_m": "刚才的事，我还没找你算账呢！",
+        "v_prev": DEFAULT_V,
+        "v_intent": DEFAULT_V,
+        "v_mixed": None
+    }
+
+with gr.Blocks(title="Unimo Context", css=custom_css) as demo:
+    gr.Markdown("# 🎭 Unitale 情感上下文测试台")
+
+    with gr.Row(variant="panel"):
+        api_url = gr.Textbox(label="IndexTTS API URL", value="http://127.0.0.1:8300", interactive=True, scale=4)
+        global_ref = gr.Audio(label="全局参考音色", type="filepath", scale=4)
+
+    rows_data = gr.State(value=[get_default_row()])
+
+    def update_state(data, idx, key, val):
+        data[idx][key] = val
+        return data
+
+    @gr.render(inputs=rows_data)
+    def show_rows(data_list):
+        for i, row in enumerate(data_list):
+            with gr.Column(elem_classes="test-row"):
+                gr.Markdown(f"#### 测试案例 #{i+1} - 1. 上句残留状态")
+                with gr.Row():
+                    txt_p = gr.Textbox(label="对白内容", value=row["txt_p"], interactive=True, scale=4)
+                    btn_p = gr.Button("📢 试听残留", variant="secondary", scale=1)
+                out_p = gr.Audio(label="音频输出", show_label=True)
+                v_prev = gr.Code(label="历史向量 (JSON)", value=json.dumps(row["v_prev"], ensure_ascii=False), language="json", interactive=True, elem_classes="wrap-json-code")
+                
+                gr.Markdown("#### 2. 此句原始意图")
+                with gr.Row():
+                    txt_i = gr.Textbox(label="对白内容", value=row["txt_i"], interactive=True, scale=4)
+                    btn_i = gr.Button("📢 试听意图", variant="secondary", scale=1)
+                out_i = gr.Audio(label="音频输出", show_label=True)
+                v_intent = gr.Code(label="意图向量 (JSON)", value=json.dumps(row["v_intent"], ensure_ascii=False), language="json", interactive=True, elem_classes="wrap-json-code")
+
+                gr.Markdown("#### 3. Unimo 混合实操")
+                with gr.Row():
+                    txt_m = gr.Textbox(label="对白内容", value=row["txt_m"], interactive=True, scale=4)
+                    btn_m = gr.Button("🪄 开始混合合成", variant="primary", scale=1)
+                out_m = gr.Audio(label="音频输出", show_label=True)
+                mixed_val = json.dumps(row["v_mixed"], ensure_ascii=False, indent=4) if row["v_mixed"] else ""
+                v_mixed = gr.Code(label="预测向量结果 (JSON)", value=mixed_val, language="json", interactive=False, elem_classes="wrap-json-code")
+
+                # --- 核心修复：这里必须直接指向全局组件 global_ref 和 api_url ---
+                btn_p.click(
+                    lambda t, v, r, u: call_indextts(t, json.loads(v), r, u)[0], 
+                    [txt_p, v_prev, global_ref, api_url], # <-- 这里修好了
+                    out_p
+                )
+                btn_i.click(
+                    lambda t, v, r, u: call_indextts(t, json.loads(v), r, u)[0], 
+                    [txt_i, v_intent, global_ref, api_url], # <-- 这里修好了
+                    out_i
+                )
+                
+                def calculate_and_gen(tp, ti, tm, vp, vi, r, u, d, idx):
+                    vm = engine.predict_context(json.loads(vp), json.loads(vi))
+                    audio_path, _ = call_indextts(tm, vm, r, u)
+                    d[idx]["v_mixed"] = vm
+                    return json.dumps(vm, ensure_ascii=False, indent=4), audio_path, d
+
+                btn_m.click(
+                    calculate_and_gen, 
+                    [txt_p, txt_i, txt_m, v_prev, v_intent, global_ref, api_url, rows_data, gr.State(i)], # <-- 这里修好了
+                    [v_mixed, out_m, rows_data]
+                )
+
+                # 同步文本框状态，防止添加行时丢失已改台词
+                txt_p.blur(update_state, [rows_data, gr.State(i), gr.State("txt_p"), txt_p], [rows_data])
+                txt_i.blur(update_state, [rows_data, gr.State(i), gr.State("txt_i"), txt_i], [rows_data])
+                txt_m.blur(update_state, [rows_data, gr.State(i), gr.State("txt_m"), txt_m], [rows_data])
+                v_prev.blur(lambda d, idx, val: update_state(d, idx, "v_prev", json.loads(val)), [rows_data, gr.State(i), v_prev], [rows_data])
+                v_intent.blur(lambda d, idx, val: update_state(d, idx, "v_intent", json.loads(val)), [rows_data, gr.State(i), v_intent], [rows_data])
+
+    def add_new_case(data):
+        data.append(get_default_row())
+        return data
+
+    with gr.Row():
+        add_btn = gr.Button("➕ 添加测试案例", variant="outline")
+        add_btn.click(add_new_case, [rows_data], [rows_data])
+
+if __name__ == "__main__":
+    demo.launch(server_port=7878, inbrowser=True)

unimo-context/Modelfile

@@ -0,0 +1,14 @@
+# ollama modelfile auto-generated by llamafactory
+
+FROM .
+
+TEMPLATE """{{ if .System }}<|im_start|>system
+{{ .System }}<|im_end|>
+{{ end }}{{ range .Messages }}{{ if eq .Role "user" }}<|im_start|>user
+{{ .Content }}<|im_end|>
+<|im_start|>assistant
+{{ else if eq .Role "assistant" }}{{ .Content }}<|im_end|>
+{{ end }}{{ end }}"""
+
+PARAMETER stop "<|im_end|>"
+PARAMETER num_ctx 4096

unimo-context/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
+}

unimo-context/chat_template.jinja

@@ -0,0 +1,89 @@
+{%- if tools %}
+    {{- '<|im_start|>system\n' }}
+    {%- if messages[0].role == 'system' %}
+        {{- messages[0].content + '\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' + messages[0].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" and message.content is string and not(message.content.startswith('<tool_response>') and message.content.endswith('</tool_response>')) %}
+        {%- set ns.multi_step_tool = false %}
+        {%- set ns.last_query_index = index %}
+    {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+    {%- if message.content is string %}
+        {%- set content = message.content %}
+    {%- else %}
+        {%- set content = '' %}
+    {%- endif %}
+    {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+        {{- '<|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 %}
+        {%- if loop.index0 > ns.last_query_index %}
+            {%- if loop.last or (not loop.last and reasoning_content) %}
+                {{- '<|im_start|>' + message.role + '\n<think>\n' + reasoning_content.strip('\n') + '\n</think>\n\n' + content.lstrip('\n') }}
+            {%- else %}
+                {{- '<|im_start|>' + message.role + '\n' + content }}
+            {%- endif %}
+        {%- else %}
+            {{- '<|im_start|>' + message.role + '\n' + content }}
+        {%- endif %}
+        {%- if message.tool_calls %}
+            {%- for tool_call in message.tool_calls %}
+                {%- if (loop.first and 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' }}
+        {{- content }}
+        {{- '\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' }}
+    {%- if enable_thinking is defined and enable_thinking is false %}
+        {{- '<think>\n\n</think>\n\n' }}
+    {%- endif %}
+{%- endif %}

unimo-context/config.json

@@ -0,0 +1,60 @@
+{
+  "architectures": [
+    "Qwen3ForCausalLM"
+  ],
+  "attention_bias": false,
+  "attention_dropout": 0.0,
+  "bos_token_id": 151643,
+  "dtype": "bfloat16",
+  "eos_token_id": 151645,
+  "head_dim": 128,
+  "hidden_act": "silu",
+  "hidden_size": 1024,
+  "initializer_range": 0.02,
+  "intermediate_size": 3072,
+  "layer_types": [
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention"
+  ],
+  "max_position_embeddings": 40960,
+  "max_window_layers": 28,
+  "model_type": "qwen3",
+  "num_attention_heads": 16,
+  "num_hidden_layers": 28,
+  "num_key_value_heads": 8,
+  "rms_norm_eps": 1e-06,
+  "rope_scaling": null,
+  "rope_theta": 1000000,
+  "sliding_window": null,
+  "tie_word_embeddings": true,
+  "transformers_version": "4.57.1",
+  "use_cache": true,
+  "use_sliding_window": false,
+  "vocab_size": 151936
+}

unimo-context/generation_config.json

@@ -0,0 +1,13 @@
+{
+  "bos_token_id": 151643,
+  "do_sample": true,
+  "eos_token_id": [
+    151645,
+    151643
+  ],
+  "pad_token_id": 151643,
+  "temperature": 0.6,
+  "top_k": 20,
+  "top_p": 0.95,
+  "transformers_version": "4.57.1"
+}

unimo-context/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1bfeb9985549944bea8333013ba6bd07dcf04fce9bd9091fffd65841dd605699
+size 1192135096

unimo-context/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
+  }
+}

unimo-context/tokenizer.json

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

unimo-context/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": 131072,
+  "pad_token": "<|endoftext|>",
+  "padding_side": "left",
+  "split_special_tokens": false,
+  "tokenizer_class": "Qwen2Tokenizer",
+  "unk_token": null
+}