#3746 introduces error in convert-mpt-hf-to-gguf.py

[maddes8cht]

Expected Behavior

In release b1412 i can successfully run the convert-mpt-hf-to-gguf.py script.

Current Behavior

With the changes introduced with #3746, I get an error with each mpt model:

python convert-mpt-hf-to-gguf.py e:\hf\mpt-7b-storywriter\
gguf: loading model mpt-7b-storywriter
gguf: found 2 model parts
This gguf file is for Little Endian only
gguf: get model metadata
gguf: get tokenizer metadata
gguf: get gpt2 tokenizer vocab
Traceback (most recent call last):
  File "e:\hf\llama.cpp\convert-mpt-hf-to-gguf.py", line 140, in <module>
    if tokenizer.added_tokens_decoder[i].special:
AttributeError: 'GPTNeoXTokenizerFast' object has no attribute 'added_tokens_decoder'

Environment and Context

Windows 10, running convert scripts in a conda environment with python 3.10.13

[Galunid]

You need to update your transformers package to at least v4.34.0. #3586 (comment)
Can be closed.

[maddes8cht]

Ah thanks, that solved the problem.
I have my (right now: almost) complete set of MPT gguf models now available at https://huggingface.co/maddes8cht ( alongside the falcons)

[github-actions]

This issue was closed because it has been inactive for 14 days since being marked as stale.