Spaces:

upstage

/

open-ko-llm-leaderboard

like 487

Running on CPU Upgrade

Hello, Thanks reporting bugs. we create the failed category.
Thanks

@choco9966 Is there any way to find the fail reason for evaluation?
I checked this code work with my model.

config = AutoConfig.from_pretrained("your model name", revision=revision)

model = AutoModel.from_pretrained("your model name", revision=revision)

tokenizer = AutoTokenizer.from_pretrained("your model name", revision=revision)

but my model(The-matt/llama2_ko-7b_sandy-fire-170_1530) is checked as failed

@The-matt
안녕하세요. 처음 실패 이유는 위의 코드를 불러오는 과정에서 Tokenizer 파일이 제대로 들어있지 않아서로 보입니다.

OSError: Can't load tokenizer for 'The-matt/llama2_ko-7b_sandy-fire-170_1530'. If you were trying to load it from 'https://huggingface.co/models', make sure you don't have a local directory with the same name. Otherwise, make sure 'The-matt/llama2_ko-7b_sandy-fire-170_1530' is the correct path to a directory containing all relevant files for a LlamaTokenizerFast tokenizer.

다만, 최근에 해당 부분 수정해서 push 해주신걸로 보여서 다시 평가할 수 있도록 수정해두겠습니다.
감사합니다.

Hello. The reason for the initial failure is that the Tokenizer file is not properly contained in the process of loading the above code, so we can see each other.

OSError: Can't load tokenizer for 'The-matt/llama2_ko-7b_sandy-fire-170_1530'. If you were trying to load it from 'https://huggingface.co/models', make sure you don't have a local directory with the same name. Otherwise, make sure 'The-matt/llama2_ko-7b_sandy-fire-170_1530' is the correct path to a directory containing all relevant files for a LlamaTokenizerFast tokenizer.

It seems that you recently revised the part and pushed it, so I will revise it so that it can be evaluated again.
Thank you.

@choco9966 답변 감사드립니다! 좋은하루되세요~!