Update to bloom tokenizer

tokenizer_config.json

@@ -1 +1 @@
-{"unk_token": "<unk>", "eos_token": "</s>", "bos_token": "<s>", "pad_token": "<pad>", "name_or_path": "/home/opc/sgpt/biencoder/nli_msmarco/sentence-transformers/bloom-1b3", "special_tokens_map_file": null, "tokenizer_class": "PreTrainedTokenizerFast"}
+{"unk_token": "<unk>", "eos_token": "</s>", "bos_token": "<s>", "pad_token": "<pad>", "name_or_path": "bigscience/tokenizer", "special_tokens_map_file": null, "tokenizer_class": "BloomTokenizerFast"}