In the next version, the weight of the model to be downloaded from the huggingface is added to the Docker image.

**Is your feature request related to a problem? Please describe.** I can't download the model weights from Huggingface due to network problems, but I can download the docker image. When the Docker container of omniparse is started, the image needs to be downloaded from the Internet. If the download fails, the container cannot be started. **Describe the solution you'd like** In the next version, the weight of the model to be downloaded from the huggingface is added to the Docker image. **Additional context** Since the docker image of omniparse has 18 GB, why not put the model weight of about 2 GB into the docker image?