Skip to content

fix Qwen/QwQ-32B #79

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

Merged
ochafik merged 2 commits into from
Aug 7, 2025
Merged

fix Qwen/QwQ-32B #79

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
48fa1a8
fix Qwen/QwQ-32B
ochafik Aug 7, 2025
2d4dd4b
Merge branch 'main' into fix-qwq
ochafik Aug 7, 2025
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
11 changes: 9 additions & 2 deletions include/minja/chat-template.hpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -162,8 +162,15 @@ class chat_template {
}), false);
caps_.supports_tools = contains(out, "some_tool");

auto out_empty = try_raw_render(json::array({dummy_user_msg, {{"role", "assistant"}, {"content", ""}}}), {}, false);
auto out_null = try_raw_render(json::array({dummy_user_msg, {{"role", "assistant"}, {"content", nullptr}}}), {}, false);
const auto render_with_content = [&](const json & content) {
const json assistant_msg {{"role", "assistant"}, {"content", content}};
// Render two assistant messages as some templates like QwQ-32B are handling
// the content differently depending on whether it's the last message or not
// (to remove the <think> tag in all but the last message).
return try_raw_render(json::array({dummy_user_msg, assistant_msg, dummy_user_msg, assistant_msg}), {}, false);
};
auto out_empty = render_with_content("");
auto out_null = render_with_content(json());
caps_.requires_non_null_content = contains(out_empty, user_needle) && !contains(out_null, user_needle);

json j_null;
Expand Down
2 changes: 1 addition & 1 deletion tests/CMakeLists.txt
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -323,9 +323,9 @@ set(MODEL_IDS
Qwen/Qwen3-235B-A22B-Instruct-2507
Qwen/Qwen3-235B-A22B-Thinking-2507
Qwen/Qwen3-Coder-30B-A3B-Instruct
Qwen/QwQ-32B

# Broken, TODO:
# Qwen/QwQ-32B
# ai21labs/AI21-Jamba-1.5-Large # https://github.com/google/minja/issues/8
# Almawave/Velvet-14B
# deepseek-ai/DeepSeek-R1
Expand Down
2 changes: 1 addition & 1 deletion tests/test-capabilities.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -80,7 +80,7 @@ TEST(CapabilitiesTest, QwQ32B) {
EXPECT_TRUE(caps.supports_system_role);
EXPECT_TRUE(caps.supports_tools);
EXPECT_TRUE(caps.supports_tool_calls);
EXPECT_FALSE(caps.supports_tool_responses);
EXPECT_TRUE(caps.supports_tool_responses);
EXPECT_TRUE(caps.supports_parallel_tool_calls);
EXPECT_TRUE(caps.requires_object_arguments);
// EXPECT_TRUE(caps.requires_non_null_content);
Expand Down
Loading