Add DBRX GGUF Support Error

[cjfghk5697]

System Info

• transformers version: 4.46.0.dev0
• Platform: Linux-6.1.85+-x86_64-with-glibc2.35
• Python version: 3.10.12
• Huggingface_hub version: 0.24.7
• Safetensors version: 0.4.5
• Accelerate version: 0.34.2
• Accelerate config: not found
• PyTorch version (GPU?): 2.4.1+cu121 (True)
• Tensorflow version (GPU?): 2.17.0 (True)
• Flax version (CPU?/GPU?/TPU?): 0.8.5 (gpu)
• Jax version: 0.4.33
• JaxLib version: 0.4.33
• Using distributed or parallel set-up in script?:
• Using GPU in script?:
• GPU type: Tesla T4

Who can help?

No response

Information

• The official example scripts
• My own modified scripts

Tasks

• An officially supported task in the examples folder (such as GLUE/SQuAD, ...)
• My own task or dataset (give details below)

Reproduction

Check it in Colab
Draft Code #33977

Expected behavior

Hi, I'm working on dbrx to support GGUF architecture(#33260). ggml.py has been modified for this purpose, but I'm encountering the following error: "TypeError: expected str, bytes or os.PathLike object, not NoneType".

It seems the error is caused by gguf_checkpoint_path being defined as None in the load_gguf_checkpoint function in modeling_gguf_pytorch_utils.py. I'm currently tracking down why it is being set to NoneType.

Upon investigation, I found that the value is being passed as NoneType at line 2441 in tokenization_utils_base.py. I need to understand why this value, which previously worked correctly, is now being set to NoneType and address this issue.

try:
    tokenizer = cls(*init_inputs, **init_kwargs)

It seems that the initial value for the GGUF file is being set to NoneType.

Here's the code for reference: Check it in Colab

Any help would be greatly appreciated.

[ArthurZucker]

cc @SunMarc