Downloads · 30 days
12
2% of all-time downloads
OpenSparX/Qwen2.5-VL-3B-Cabin-CyberSpark
Qwen2.5-VL-3B-Cabin-CyberSpark is a machine learning model from OpenSparX. Use it for the machine learning task on the model card, and read the license before you ship it in a product.
Downloads · 30 days
12
2% of all-time downloads
All-time downloads
552
Public
Parameters
3.8B
8.1 GB on disk
Likes
0
Public
Click a slice to open those files.
.safetensors8.1 GB · 100%
From the Hugging Face model README

这是一个基于 Qwen2.5-VL-3B-Instruct 的多模态座舱助手微调模型,支持图像理解与中文语音助手场景的指令响应。模型面向车内多种感知与交互任务:情绪安抚与功能推荐、驾驶分心行为提醒、儿童安全带佩戴检测等。
采用 ChatML 风格消息结构,结合 <image> 标记与操作串:
<|SPEECH_END|> playOperation|PAUSE| ... 或 <|TAG|> phone_left_hand。{
"messages": [
{"role": "system", "content": "You are Qwen, a virtual human developed by the Qwen Team, Alibaba Group, capable of perceiving auditory and visual inputs, as well as generating text and speech."},
{"role": "user", "content": "<image>你现在是一个优秀的汽车座舱智能语音助手,你的名字是大众,可以通过输入图片信息做出对应的回答和功能推荐。如果识别到了用户有开心、悲伤、愤怒、平静等情绪,可以安抚并给出功能推荐。"}
],
"images": ["/data/LLM-SFT/datasets/face_emotion_dataset_0429/images/3675.jpg"]
}
期望输出示例:
我帮您保持柔和的中性色灯光,播放了一首安静的背景乐 <|SPEECH_END|> enableAmbientLight|true| setAmbientLightMode|3| playOperation|PLAY|tag|neutral|
{
"messages": [
{"role": "system", "content": "You are Qwen, a virtual human developed by the Qwen Team, Alibaba Group, capable of perceiving auditory and visual inputs, as well as generating text and speech."},
{"role": "user", "content": "<image>你现在是一个优秀的汽车座舱智能语音助手,你的名字是大众,可以通过输入图片信息做出对应的回答和功能推荐。\n如果识别到了用户在吸烟、打瞌睡、接打电话等行为,可以做出提醒并给出功能推荐。"}
],
"images": ["/data/DMS_CV/datasets/state-farm-distracted-driver-detection/imgs/val/c4/img_22127.jpg"]
}
期望输出示例:
您正在打电话,我帮您把车内音乐调低了,请专心驾驶 <|SPEECH_END|> playOperation|PAUSE| setMediaVolume|6| setMute|true| <|TAG|> phone_left_hand
{
"messages": [
{"role": "system", "content": "You are Qwen, a virtual human developed by the Qwen Team, Alibaba Group, capable of perceiving auditory and visual inputs, as well as generating text and speech."},
{"role": "user", "content": "<image>你现在是一个优秀的汽车座舱智能语音助手,你的名字是大众,可以通过输入图片信息做出对应的回答和功能推荐。看下是儿童是否正确佩戴安全带,若没有,请提醒用户。"}
],
"images": ["/data/LLM-SFT/datasets/safetyBelt/correct-wear/0701_00027_ (137).png"]
}
期望输出示例:
检测到儿童已正确佩戴安全带 <|SPEECH_END|> tag|correct-wear|
使用 Transformers 与内置 chat_template:
from transformers import AutoTokenizer, AutoModelForCausalLM
from PIL import Image
repo_id = "PiloBi/Qwen2.5-VL-3B-Cabin"
tokenizer = AutoTokenizer.from_pretrained(repo_id, trust_remote_code=True)
model = AutoModelForCausalLM.from_pretrained(
repo_id,
trust_remote_code=True,
torch_dtype="auto",
device_map="auto",
)
messages = [
{"role": "system", "content": "You are Qwen, a virtual human developed by the Qwen Team, Alibaba Group, capable of perceiving auditory and visual inputs, as well as generating text and speech."},
{"role": "user", "content": "<image>你现在是一个优秀的汽车座舱智能语音助手,你的名字是大众,可以通过输入图片信息做出对应的回答和功能推荐。"},
]
query = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
image = Image.open("/path/to/your/image.jpg").convert("RGB")
inputs = tokenizer([query], images=[image], return_tensors="pt").to(model.device)
outputs = model.generate(**inputs, max_new_tokens=128)
print(tokenizer.decode(outputs[0], skip_special_tokens=False))
model-00001/00002.safetensors 与 model.safetensors.index.json:权重与索引。config.json、generation_config.json:模型与推理配置。tokenizer.json/tokenizer_config.json/merges.txt/vocab.json:分词器相关。chat_template.jinja:对话模板,建议通过 apply_chat_template 使用。preprocessor_config.json、video_preprocessor_config.json:多模态预处理配置。license: mit datasets: