Upload tokenizer_specs_16k.json with huggingface_hub

tokenizer_specs_16k.json

@@ -0,0 +1,9 @@
+{
+  "tokenizers": [
+    {
+      "name": "sp_bpe_16384",
+      "dataset_suffix": "sp16384",
+      "vocab_size": 16384
+    }
+  ]
+}