summaryrefslogtreecommitdiff
path: root/tokenizer/byte/tokenizer_test.go
diff options
context:
space:
mode:
authorJonas Knobloch <jonas.knobloch@t-online.de>2026-09-11 18:38:13 +0200
committerJonas Knobloch <jonas.knobloch@t-online.de>2026-09-11 18:38:13 +0200
commit75e581b3bc19a73d0c1f48b32f2e58121c9741c1 (patch)
tree897e2f24db3374d58ba36dcee67879cc7cd4956c /tokenizer/byte/tokenizer_test.go
parent330c6387962b197c5ca7a8051d4864feaecc0f25 (diff)
parentf998fee7427403698f9262d1e92d1a307c652094 (diff)
Merge remote-tracking branch 'origin/wip-entropy' into wip-frequency
Diffstat (limited to 'tokenizer/byte/tokenizer_test.go')
-rw-r--r--tokenizer/byte/tokenizer_test.go49
1 files changed, 49 insertions, 0 deletions
diff --git a/tokenizer/byte/tokenizer_test.go b/tokenizer/byte/tokenizer_test.go
new file mode 100644
index 0000000..d2c8c6c
--- /dev/null
+++ b/tokenizer/byte/tokenizer_test.go
@@ -0,0 +1,49 @@
+package byte
+
+import (
+ "fmt"
+ "os"
+ "testing"
+
+ "go.jknobloc.com/x/shelf"
+)
+
+func TestNewTokenizer_Encode(t *testing.T) {
+ var file *os.File
+
+ if f, err := os.Open(shelf.Abs("tokenizers/minipile/tokenizer_gpt2_256_m000_minipile/vocab.json")); err != nil {
+ t.Fatal(err)
+ } else {
+ file = f
+ }
+
+ defer file.Close()
+
+ tok, err := NewTokenizer(file)
+
+ if err != nil {
+ t.Fatal(err)
+ }
+
+ fmt.Println(tok.Encode(" \nabc"))
+}
+
+func TestNewTokenizer_Decode(t *testing.T) {
+ var file *os.File
+
+ if f, err := os.Open(shelf.Abs("tokenizers/minipile/tokenizer_gpt2_256_m000_minipile/vocab.json")); err != nil {
+ t.Fatal(err)
+ } else {
+ file = f
+ }
+
+ defer file.Close()
+
+ tok, err := NewTokenizer(file)
+
+ if err != nil {
+ t.Fatal(err)
+ }
+
+ fmt.Println(tok.Decode([]uint8{220, 198, 64, 65, 66}))
+}