From fa4e25b4992c25cdec41bdccb8af909b93fc9c39 Mon Sep 17 00:00:00 2001 From: Jonas Knobloch Date: Tue, 31 Mar 2026 17:05:43 +0200 Subject: Count skipped texts --- llm/run.go | 16 ++++++++++++++-- 1 file changed, 14 insertions(+), 2 deletions(-) diff --git a/llm/run.go b/llm/run.go index df48ff6..227a365 100644 --- a/llm/run.go +++ b/llm/run.go @@ -52,13 +52,25 @@ func (e *Evaluator[R]) Run(title string, data dataset.Reader, window, stride int return int(e.completed.Load()) }) - defer pb.Close() + s := 0 // skipped + n := 0 // total - n := 0 + defer func() { + fmt.Printf("skipped %d of %d texts\n", s, n) + }() + + defer pb.Close() for d := range data.Texts("text") { tokens := toInt64(e.tokenizer.Tokenize(d)) + if len(tokens) == 0 { + n++ + s++ + + continue + } + e.schedule(n, tokens, window, stride, e.batchSize) pb.SetTotal(pb.Total() + e.estimateJobs(tokens, window, stride)) -- cgit v1.2.3