Skip to content

fix: detect Ollama vision models from naming conventions #345

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
manavshrivastavagit wants to merge 1 commit into
base: main
Choose a base branch
from
Open

fix: detect Ollama vision models from naming conventions #345

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
5 changes: 4 additions & 1 deletion coworker/providers/capabilities.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -24,9 +24,12 @@ def capabilities_for(model: str) -> ModelCapabilities:

# Ollama (local) models vary widely and many fake/mishandle parallel tool calls — assume
# tools work (we only point at tool-capable models) but stay conservative otherwise.
# Vision is detected from common model naming conventions (-vl, vision, llava, etc.).
if provider == "ollama":
_vision_patterns = ("-vl", "vision", "llava", "bakllava", "cogvlm", "minicpm-v")
has_vision = any(p in name for p in _vision_patterns)
return ModelCapabilities(
tools=True, vision=False, parallel_tool_calls=False, streaming=True
tools=True, vision=has_vision, parallel_tool_calls=False, streaming=True
)

# Cloud-account providers (custom-added ids; curated ones answered from the matrix).
Expand Down