Add Pile-train

constants.py

@@ -3,6 +3,7 @@ import os
 # options
 CORPUS_BY_DESC = {
     'RedPajama (LLaMA tokenizer), 1.4T tokens': 'v3_rpj_llama_c4',
+    'Pile-train (LLaMA tokenizer), 380B tokens': 'v4_piletrain_llama',
     'Pile-val (LLaMA tokenizer), 390M tokens': 'v3_pileval_llama',
     'Pile-val (GPT-2 tokenizer), 380M tokens': 'v3_pileval_gpt2',
     'Dolma-sample (OLMo tokenizer), 8.0B tokens': 'v4_dolmasample_olmo',