Consider tokenizer a type of text

This commit is contained in:
oobabooga 2023-03-10 00:45:28 -03:00
parent ec3de0495c
commit 875847bf88

View File

@ -108,7 +108,7 @@ def get_download_links_from_huggingface(model, branch):
is_pytorch = re.match("pytorch_model.*\.bin", fname) is_pytorch = re.match("pytorch_model.*\.bin", fname)
is_safetensors = re.match("model.*\.safetensors", fname) is_safetensors = re.match("model.*\.safetensors", fname)
is_tokenizer = re.match("tokenizer.*\.model", fname) is_tokenizer = re.match("tokenizer.*\.model", fname)
is_text = re.match(".*\.(txt|json)", fname) is_text = re.match(".*\.(txt|json)", fname) or is_tokenizer
if any((is_pytorch, is_safetensors, is_text, is_tokenizer)): if any((is_pytorch, is_safetensors, is_text, is_tokenizer)):
if is_text: if is_text: