coffee
18 lines
759 B
Python
Executable File
18 lines
759 B
Python
Executable File
from transformers import AutoTokenizer,XLMRobertaTokenizer
|
|
#由spm转为hf_tokenizer
|
|
tokenizer = XLMRobertaTokenizer.from_pretrained("tokenizer_file/dolly_tokenizer_spm/tokenizer.model")
|
|
text = "A Frenchman from Paris"
|
|
encoded_input = tokenizer.encode(text)
|
|
print(encoded_input)
|
|
decoded_output = tokenizer.decode(encoded_input)
|
|
print(decoded_output)
|
|
tokenizer.save_pretrained('tokenizer_file/dolly_tokenizer_hf')
|
|
|
|
#转换为常见的tokenzier.json格式
|
|
tokenizer = AutoTokenizer.from_pretrained("tokenizer_file/dolly_tokenizer_hf")
|
|
text = "A Frenchman from Paris"
|
|
encoded_input = tokenizer.encode(text)
|
|
print(encoded_input)
|
|
decoded_output = tokenizer.decode(encoded_input)
|
|
print(decoded_output)
|
|
tokenizer.save_pretrained('tokenizer_file/dolly_tokenizer_hf2') |