Skip to content

process hangs while use multimodel #133

@ray0809

Description

@ray0809

When I use the latest main branch(2026-03-17)
running the service(rkllama_server --port 11434 --models /home/orangepi/Projects/Ray/rkllama/models/ --debug)
do post with image, I will get stuck as the logger show

Image





here is my Modelfile

Image





here is my request demo

from ollama import chat, generate
import base64
from pathlib import Path
import requests

path = '/home/orangepi/Projects/Ray/rkllama/images/06b4ef5a7e3a2e79229139417e1d0b9a.jpg'
img = base64.b64encode(Path(path).read_bytes()).decode()


model_list = [
    'tinyllama-chat:1.1b',
    'qwen3-vl-4b',
    'qwen3-vl-2b',
]

# response, code = generate(
#     model='qwen3-vl-2b',
#     prompt='Help me identify the type, material and color of the clothing in the picture',
#     images=[img],
#     format='json',
#     stream=False,
# )
# print(code)


payload = {
    "model": model_list[2],
    
    "prompt": "Help me identify the type, material and color of the clothing in the picture",
    "images": [img],
    # "prompt": "Hello",
 
    "stream": False,
    'format': 'json',
    'system': 'You are a helpful AI assistant. Keep your answer clear and straightforward.'
}
OLLAMA_URL = 'http://localhost:11434/api/generate'
resp = requests.post(OLLAMA_URL, json=payload, timeout=300)
print(resp.json())




Does anyone know how to debug this problem?

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions