ValueError: Architecture deepseek2 not supported

[czq99972]

System Info

The current Transformers framework doesn't support the gguf quantized model files from deepseek2. Can you please advise when this support might be added? @SunMarc @MekkCyber

Who can help?

@SunMarc @MekkCyber

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

File "/home/work/miniforge3/envs/vllm/lib/python3.11/site-packages/transformers/models/auto/configuration_auto.py", line 1006, in from_pretrained
config_dict, unused_kwargs = PretrainedConfig.get_config_dict(pretrained_model_name_or_path, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/work/miniforge3/envs/vllm/lib/python3.11/site-packages/transformers/configuration_utils.py", line 570, in get_config_dict
config_dict, kwargs = cls._get_config_dict(pretrained_model_name_or_path, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/work/miniforge3/envs/vllm/lib/python3.11/site-packages/transformers/configuration_utils.py", line 661, in _get_config_dict
config_dict = load_gguf_checkpoint(resolved_config_file, return_tensors=False)["config"]
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/work/miniforge3/envs/vllm/lib/python3.11/site-packages/transformers/modeling_gguf_pytorch_utils.py", line 103, in load_gguf_checkpoint
raise ValueError(f"Architecture {architecture} not supported")
ValueError: Architecture deepseek2 not supported

Expected behavior

1

[VladOS95-cyber]

Hey @czq99972, @SunMarc, @MekkCyber! I can take it as soon as I finish current implementation for Mamba arch, but it wouldn't be so long. I think I will be able to start working on deepseek2 on this week. Link to main issue thread: #33260

[SunMarc]

Hey ! Deepspeedv2 gguf can be supported with gguf files once it is integrated in transformers: #31976 !

[wavy-jung]

Any update on deepseek v2 support? @VladOS95-cyber

[VladOS95-cyber]

hey @wavy-jung, I see that Deepspeedv2 architecture is not supported yet, this PR #31976 is still in progress

[github-actions]

This issue has been automatically marked as stale because it has not had recent activity. If you think this still needs to be addressed please comment on this thread.

Please note that issues that do not follow the contributing guidelines are likely to be ignored.