Update Ollama and Open WebUI image tags to support new models

[foxminchan]

Summary

The current Ollama and Open WebUI container image tags are outdated and do not support several newly released models (e.g., embeddinggemma, DeepSeek-OCR). These models require newer runtime capabilities that are only available in the latest stable versions of the images.

Requested Change

Update the following image tags in OllamaContainerImageTags:

• ollama/ollama → 0.13.0
• open-webui/open-webui → 0.6.38

Reason

• Ensures compatibility with recently published models.
• Avoids runtime errors due to unsupported model formats.
• Keeps the integration aligned with upstream releases.

Impact

• No breaking changes.
• Only updates version constants.