Reblacking

Signed-off-by: Davis Wertheimer <[email protected]>

Author: daviswer

fms_fsdp/config/training.py

@@ -15,9 +15,7 @@ class train_config:
     file_type: str = "arrow"
     col_name: str = "tokens"
     tokenizer_path: str = "/fsx/tokenizer"
-    datasets: str = (
-        "lang=en/dataset=commoncrawl,lang=en/dataset=webhose,lang=en/dataset=github_clean,lang=de/dataset=wikipedia,lang=es/dataset=wikipedia,lang=fr/dataset=wikipedia,lang=ja/dataset=wikipedia,lang=pt/dataset=wikipedia,lang=en/dataset=wikimedia,lang=en/dataset=uspto,lang=en/dataset=pubmedcentral,lang=en/dataset=arxiv,lang=en/dataset=stackexchange"
-    )
+    datasets: str = "lang=en/dataset=commoncrawl,lang=en/dataset=webhose,lang=en/dataset=github_clean,lang=de/dataset=wikipedia,lang=es/dataset=wikipedia,lang=fr/dataset=wikipedia,lang=ja/dataset=wikipedia,lang=pt/dataset=wikipedia,lang=en/dataset=wikimedia,lang=en/dataset=uspto,lang=en/dataset=pubmedcentral,lang=en/dataset=arxiv,lang=en/dataset=stackexchange"
     weights: str = "7725,500,550,28,17,22,25,8,100,500,175,250,100"
     seq_length: int = 4096
     vocab_size: int = 32000