Skip to content

84 possible refactor#92

Open
AymanBx wants to merge 23 commits into
mainfrom
84-possible-refactor
Open

84 possible refactor#92
AymanBx wants to merge 23 commits into
mainfrom
84-possible-refactor

Conversation

@AymanBx

@AymanBx AymanBx commented Jun 10, 2026

Copy link
Copy Markdown
Contributor
  • Runner has run method that is called by task
    • run logic moved to runner
    • Runner object has param dict as an attribute
    • runner now creates a dict called response info that includes token count and stop reason in addition to response
  • Logger class
  • Logger objects will either be initialized in bench.run or task.run
  • Logger object passed between bench, task and runner
    • Each will log its info, then runner logs interaction (each logging step has a method)
  • Logging file structure modified
    • Bench has its own run_id that is included in task's run_id
  • Bench_score works with the new structure.
    • Collate option removed
  • Task has score method renamed to score_prompt
  • Added demo files to demonstrate the new changes

@AymanBx
AymanBx requested a review from brownsarahm June 10, 2026 18:41
@AymanBx AymanBx linked an issue Jun 10, 2026 that may be closed by this pull request
This was linked to issues Jun 18, 2026
@@ -1,4 +1,4 @@
runner_type: ollama
model:
models:

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

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

enable specifying the parameters under here

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

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

combine this with runnery.yml under 'model_params' key and then pass that whole sub-dictionary to the model

Comment thread benchtools/benchmark.py Outdated
for task_dict in task_list:
tasks.append(Task.from_dict(task_dict,source_path=benchmark_path))

runners = Bench.load_runners(benchmark_path)

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

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

call the from_file on benchlist instead and in the bench constructor

@AymanBx
AymanBx marked this pull request as ready for review July 13, 2026 06:26
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

possible refactor add token counts improve log structure

2 participants