Update perplexity evaluator

* Introduce results channel
* Introduce jobs channel
* Generalize evaluation via callback

1 files changed, 53 insertions, 0 deletions

diff --git a/llm/cmd/eval/main.go b/llm/cmd/eval/main.go
new file mode 100644
index 0000000..05b3f96
--- /dev/null
+++ b/llm/cmd/eval/main.go
@@ -0,0 +1,53 @@
+package main
+
+import (
+	"log"
+
+	"github.com/jonasknobloch/mbpe"
+
+	"go.jknobloc.com/x/dataset"
+	"go.jknobloc.com/x/gpt2"
+)
+
+func main() {
+	perplexity()
+}
+
+func data() *dataset.ParquetReader {
+	var miniPile *dataset.ParquetReader
+
+	if r, err := dataset.NewParquetReader("dataset/cmd/dataset/tmp/minipile/validation"); err != nil {
+		log.Fatal(err)
+	} else {
+		miniPile = r
+	}
+
+	return miniPile
+}
+
+func model() *gpt2.Model {
+	m := gpt2.NewModel("gpt2/models/base/model.onnx", "0")
+
+	if err := m.Init(); err != nil {
+		log.Fatal(err)
+	}
+
+	return m
+}
+
+func tokenizer() *mbpe.Tokenizer {
+	m := mbpe.NewMBPE()
+
+	if err := m.Load("gpt2/models/base/vocab.json", "gpt2/models/base/merges.txt"); err != nil {
+		log.Fatal(err)
+	}
+
+	t := mbpe.NewTokenizer(m)
+
+	byteLevel := mbpe.NewByteLevel(false)
+
+	t.SetPreTokenizer(byteLevel)
+	t.SetDecoder(byteLevel)
+
+	return t
+}