Skip to content

Fix text_llm_runner kv cache pos count and use it for generate() #15286

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
merged 2 commits into from
Oct 21, 2025
Merged

Fix text_llm_runner kv cache pos count and use it for generate() #15286

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
7ebb46b
Fix text_llm_runner kv cache pos count
kirklandsign Oct 20, 2025
12545a3
Should use pos_ to text_token_generator
kirklandsign Oct 21, 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
4 changes: 3 additions & 1 deletion extension/llm/runner/text_llm_runner.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -190,7 +190,7 @@ Error TextLLMRunner::generate(
// Generate max_new_tokens - 1 because prefill already generated 1 token.
auto generate_result = text_token_generator_->generate(
prompt_tokens,
num_prompt_tokens,
pos_,
max_new_tokens - 1,
temperature_ == -1.0f ? config.temperature : temperature_,
wrapped_callback);
Expand All @@ -199,6 +199,8 @@ Error TextLLMRunner::generate(
}
int64_t num_generated_tokens = generate_result.get();

pos_ += num_generated_tokens;
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

you don't need this anymore right? pos_ is being updated.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

No :(

https://github.com/pytorch/executorch/blob/main/extension/llm/runner/text_token_generator.h#L53

https://github.com/pytorch/executorch/blob/main/extension/llm/runner/multimodal_runner.cpp#L180-L190


stats_->inference_end_ms = time_in_ms();
if (!config.warming) {
printf("\n");
Expand Down
Loading