fix(deploy): use DeepSeek for worker vision by EricNGOntos · Pull Request #454 · Ontos-AI/knowhere

EricNGOntos · 2026-10-05T01:50:14Z

Summary

set the shared ECS worker IMAGE_MODEL override to deepseek-flash
update the deployment rendering contract to prevent Qwen from returning
applies to both staging and production because both render the shared worker template

make check
uv run pytest deploy/ecs/test_render_task_definitions.py -q
PYTHONPATH=packages/shared-python .venv/bin/pytest packages/shared-python/shared/tests/test_page_memory_config.py -q

fix(deploy): use deepseek for worker vision

600f0dc

EricNGOntos merged commit 06854e3 into main Oct 5, 2026
6 checks passed

EricNGOntos deleted the fix/wuchengke/use-deepseek-vlm branch October 5, 2026 01:56