Upload folder using huggingface_hub

wordpiece32000minipile/raw_tok_path.txt

@@ -1 +1 @@
-/home/pl487/rdd/outputs/tok_train/wordpiece_minipile_2025-02-03T10-57-36
+/home/pl487/rdd/outputs/tok_train/wordpiece_minipile_2025-02-03T12-10-57

wordpiece32000minipile/special_tokens_map.json

@@ -1,3 +1,4 @@
 {
-  "eos_token": "<|endoftext|>"
+  "eos_token": "<|endoftext|>",
+  "unk_token": "<|unk|>"
 }

wordpiece32000minipile/tokenizer_config.json

@@ -7,10 +7,19 @@
       "rstrip": false,
       "single_word": false,
       "special": true
+    },
+    "1": {
+      "content": "<|unk|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
     }
   },
   "clean_up_tokenization_spaces": true,
   "eos_token": "<|endoftext|>",
   "model_max_length": 1000000000000000019884624838656,
-  "tokenizer_class": "PreTrainedTokenizerFast"
+  "tokenizer_class": "PreTrainedTokenizerFast",
+  "unk_token": "<|unk|>"
 }