From db6db25bdfd6e1984cd74fd075e06d01ce3a12ff Mon Sep 17 00:00:00 2001 From: Yeaju Kim Date: Tue, 21 Jul 2026 15:38:58 -0400 Subject: [PATCH 1/2] updating documentation - diana k Updated model parameter in transcribe command and added notes for future updates. switching --model tiny to --model small --- docs/mkdocs/tutorials/walkthrough.md | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/docs/mkdocs/tutorials/walkthrough.md b/docs/mkdocs/tutorials/walkthrough.md index c41f5e1..c7fc145 100644 --- a/docs/mkdocs/tutorials/walkthrough.md +++ b/docs/mkdocs/tutorials/walkthrough.md @@ -59,7 +59,7 @@ its inputs from the directory layout; you never pass individual file paths. Whisper speech-recognition model. ```bash -hypline transcribe data/ --audio-ext .wav --model tiny +hypline transcribe data/ --audio-ext .wav --model small ``` ```text @@ -79,7 +79,7 @@ and voice-activity-detection messages.) learning the workflow rather than analyzing the transcripts. For a real analysis, omit `--model` to use the default `large-v2`, which is far more accurate but a multi-GB download and much slower on CPU (pass `--device cuda` if you have a - GPU). + GPU). -- to be updated Only four files are transcribed, not one per run. These are the reading-condition subset from [step 1](#1-get-the-example-dataset): each run's @@ -446,7 +446,7 @@ choices side by side. demonstrate the *effect*, since two runs of `--model tiny` transcripts are far too little data for the cross-brain scores to mean anything. On a full study they carry the shared-representation signal; here they only confirm the - command runs end to end. + command runs end to end. -- to be updated !!! success "Check" From 61f8c75566ae33bbb901f6f595cfbc1cbb06de21 Mon Sep 17 00:00:00 2001 From: Sangyoon Park Date: Tue, 21 Jul 2026 15:43:24 -0400 Subject: [PATCH 2/2] Update title from 'Test' to 'Interpret' in analyze-evals.md blah blah --- docs/mkdocs/how-to/analyze-evals.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/docs/mkdocs/how-to/analyze-evals.md b/docs/mkdocs/how-to/analyze-evals.md index 1b4ce5b..0608357 100644 --- a/docs/mkdocs/how-to/analyze-evals.md +++ b/docs/mkdocs/how-to/analyze-evals.md @@ -1,4 +1,4 @@ -# Test the cross-brain effect from evals +# Interpret the cross-brain effect from evals You have run [`encoding analyze`](../reference/encoding.md) a few times and have a handful of eval files. This guide turns them into an answer to the question