Instructions to use DIYIN/Youtu-LLM-A7B-Base with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use DIYIN/Youtu-LLM-A7B-Base with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("image-text-to-text", model="DIYIN/Youtu-LLM-A7B-Base", trust_remote_code=True)
messages = [
    {
        "role": "user",
        "content": [
            {"type": "image", "url": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/p-blog/candy.JPG"},
            {"type": "text", "text": "What animal is on the candy?"}
        ]
    },
]
pipe(text=messages)

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoModelForCausalLM
model = AutoModelForCausalLM.from_pretrained("DIYIN/Youtu-LLM-A7B-Base", trust_remote_code=True, device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use DIYIN/Youtu-LLM-A7B-Base with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "DIYIN/Youtu-LLM-A7B-Base"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "DIYIN/Youtu-LLM-A7B-Base",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

Use Docker

docker model run hf.co/DIYIN/Youtu-LLM-A7B-Base

SGLang

How to use DIYIN/Youtu-LLM-A7B-Base with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "DIYIN/Youtu-LLM-A7B-Base" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "DIYIN/Youtu-LLM-A7B-Base",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "DIYIN/Youtu-LLM-A7B-Base" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "DIYIN/Youtu-LLM-A7B-Base",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

Docker Model Runner
How to use DIYIN/Youtu-LLM-A7B-Base with Docker Model Runner:
```
docker model run hf.co/DIYIN/Youtu-LLM-A7B-Base
```

Youtu-LLM-A7B-Base / processing_youtu_vl_moe.py

DIYIN

Duplicate from tencent/Youtu-LLM-A7B-Base

7162ea8 4 days ago

Raw History Blame Contribute Delete

1.79 kB

# coding=utf-8
"""
UTU-VL MoE 处理器（适配 transformers 5.3）
直接复用 Qwen3VLProcessor，UTU-VL MoE 使用与 V4 相同的图像处理流程和 tokenizer 特殊 token 设置。
- image_token_id = 128017 (<|image_pad|>)
- video_token_id = 128040 (<｜place▁holder▁no▁40｜>, 与 Stage3 VIDEO_TOKEN_ID 一致；128032 在 union tokenizer 中是字面量 ｜DSML｜)
"""

from transformers.models.qwen3_vl import Qwen3VLProcessor


class YoutuVLMoEProcessor(Qwen3VLProcessor):
    """
    UTU-VL MoE processor.
    It is intentionally a thin Qwen3VLProcessor subclass with UTU-VL token
    aliases populated on the tokenizer.
    """

    image_token = "<|image_pad|>"
    video_token = "<｜place▁holder▁no▁40｜>"
    vision_start_token = "<|vision_start|>"
    vision_end_token = "<|vision_end|>"

    tokenizer_class = ("PreTrainedTokenizerFast",)

    def __init__(self, image_processor=None, tokenizer=None, video_processor=None, chat_template=None, **kwargs):
        if tokenizer is not None:
            tokenizer.image_token = self.image_token
            tokenizer.video_token = self.video_token
            tokenizer.vision_start_token = self.vision_start_token
            tokenizer.vision_end_token = self.vision_end_token
            tokenizer.image_token_id = tokenizer.convert_tokens_to_ids(self.image_token)
            tokenizer.video_token_id = tokenizer.convert_tokens_to_ids(self.video_token)
            tokenizer.vision_start_token_id = tokenizer.convert_tokens_to_ids(self.vision_start_token)
            tokenizer.vision_end_token_id = tokenizer.convert_tokens_to_ids(self.vision_end_token)
        super().__init__(image_processor, tokenizer, video_processor, chat_template=chat_template, **kwargs)


__all__ = ["YoutuVLMoEProcessor"]

	# coding=utf-8
	"""
	UTU-VL MoE 处理器（适配 transformers 5.3）

	直接复用 Qwen3VLProcessor，UTU-VL MoE 使用与 V4 相同的图像处理流程和 tokenizer 特殊 token 设置。
	- image_token_id = 128017 (<\|image_pad\|>)
	- video_token_id = 128040 (<｜place▁holder▁no▁40｜>, 与 Stage3 VIDEO_TOKEN_ID 一致；128032 在 union tokenizer 中是字面量｜DSML｜)
	"""

	from transformers.models.qwen3_vl import Qwen3VLProcessor


	class YoutuVLMoEProcessor(Qwen3VLProcessor):
	"""
	UTU-VL MoE processor.

	It is intentionally a thin Qwen3VLProcessor subclass with UTU-VL token
	aliases populated on the tokenizer.
	"""

	image_token = "<\|image_pad\|>"
	video_token = "<｜place▁holder▁no▁40｜>"
	vision_start_token = "<\|vision_start\|>"
	vision_end_token = "<\|vision_end\|>"

	tokenizer_class = ("PreTrainedTokenizerFast",)

	def __init__(self, image_processor=None, tokenizer=None, video_processor=None, chat_template=None, **kwargs):
	if tokenizer is not None:
	tokenizer.image_token = self.image_token
	tokenizer.video_token = self.video_token
	tokenizer.vision_start_token = self.vision_start_token
	tokenizer.vision_end_token = self.vision_end_token
	tokenizer.image_token_id = tokenizer.convert_tokens_to_ids(self.image_token)
	tokenizer.video_token_id = tokenizer.convert_tokens_to_ids(self.video_token)
	tokenizer.vision_start_token_id = tokenizer.convert_tokens_to_ids(self.vision_start_token)
	tokenizer.vision_end_token_id = tokenizer.convert_tokens_to_ids(self.vision_end_token)
	super().__init__(image_processor, tokenizer, video_processor, chat_template=chat_template, **kwargs)


	__all__ = ["YoutuVLMoEProcessor"]