search

SkunkworksAI / BakLLaVA

Apache License 2.0
702 stars 46 forks source link

Cannot import name '_expand_mask' from 'transformers.models.bloom.modeling_bloom' #7

Open ArturFormella opened 1 year ago

ArturFormella commented 1 year ago

I have the following error:

[2023-10-28 15:58:46,969] [INFO] [real_accelerator.py:110:get_accelerator] Setting ds_accelerator to cuda (auto detect)
Traceback (most recent call last):
  File "/home/artur/anaconda3/envs/llava/lib/python3.10/runpy.py", line 187, in _run_module_as_main
    mod_name, mod_spec, code = _get_module_details(mod_name, _Error)
  File "/home/artur/anaconda3/envs/llava/lib/python3.10/runpy.py", line 110, in _get_module_details
    __import__(pkg_name)
  File "/T8/ai_soft/BakLLaVA/llava/__init__.py", line 1, in <module>
    from .model import LlavaLlamaForCausalLM
  File "/T8/ai_soft/BakLLaVA/llava/model/__init__.py", line 3, in <module>
    from .language_model.llava_mpt import LlavaMPTForCausalLM, LlavaMPTConfig
  File "/T8/ai_soft/BakLLaVA/llava/model/language_model/llava_mpt.py", line 26, in <module>
    from .mpt.modeling_mpt import MPTConfig, MPTForCausalLM, MPTModel
  File "/T8/ai_soft/BakLLaVA/llava/model/language_model/mpt/modeling_mpt.py", line 19, in <module>
    from .hf_prefixlm_converter import add_bidirectional_mask_if_missing, convert_hf_causal_lm_to_prefix_lm
  File "/T8/ai_soft/BakLLaVA/llava/model/language_model/mpt/hf_prefixlm_converter.py", line 15, in <module>
    from transformers.models.bloom.modeling_bloom import _expand_mask as _expand_mask_bloom
ImportError: cannot import name '_expand_mask' from 'transformers.models.bloom.modeling_bloom' (/T8/ai_soft/BakLLaVA/.venv/lib/python3.10/site-packages/transformers/models/bloom/modeling_bloom.py)

I think it's related to these changes: https://github.com/huggingface/transformers/commit/ac5893756bafcd745d93a442cf36f984545dbad8

Steps:

conda create -n llava python=3.10 -y conda activate llava pip install --upgrade pip # enable PEP 660 support pip install -e . pip uninstall transformers pip install git+https://github.com/huggingface/transformers ls .venv/lib/python3.10/site-packages/transformers/models/bloom/modeling_bloom.py .venv/lib/python3.10/site-packages/transformers/models/bloom/modeling_bloom.py

python -m llava.serve.model_worker --host 0.0.0.0 --controller http://localhost:10000 --port 40001 --worker http://localhost:40001 --model-path /T8/ai_files/gpt_models/SkunkworksAI_BakLLaVA-1

nvidia-smi -L GPU 0: NVIDIA GeForce RTX 3090 (UUID: GPU-c632cd8d-fdb2-5295-e950-ddb18636c332)

Ubuntu 22.10 VM on Proxmox

ArturFormella commented 1 year ago 
> pip install transformers==4.31.0
> python -m llava.serve.model_worker --host 0.0.0.0 --controller http://localhost:10000 --port 40001 --worker http://localhost:40001 --model-path /T8/ai_files/gpt_models/SkunkworksAI_BakLLaVA-1
[2023-10-28 16:22:15,589] [INFO] [real_accelerator.py:110:get_accelerator] Setting ds_accelerator to cuda (auto detect)
Traceback (most recent call last):
  File "/home/artur/anaconda3/envs/llava/lib/python3.10/runpy.py", line 187, in _run_module_as_main
    mod_name, mod_spec, code = _get_module_details(mod_name, _Error)
  File "/home/artur/anaconda3/envs/llava/lib/python3.10/runpy.py", line 110, in _get_module_details
    __import__(pkg_name)
  File "/T8/ai_soft/BakLLaVA/llava/__init__.py", line 1, in <module>
    from .model import LlavaLlamaForCausalLM
  File "/T8/ai_soft/BakLLaVA/llava/model/__init__.py", line 2, in <module>
    from .language_model.llava_mistral import LlavaMistralForCausalLM, LlavaConfig
  File "/T8/ai_soft/BakLLaVA/llava/model/language_model/llava_mistral.py", line 22, in <module>
    from transformers import AutoConfig, AutoModelForCausalLM, \
ImportError: cannot import name 'MistralConfig' from 'transformers' (/T8/ai_soft/BakLLaVA/.venv/lib/python3.10/site-packages/transformers/__init__.py)

I have no idea what to do.

ArturFormella commented 1 year ago

It looks like release 4.34.0 is the latest working with BakLLaVA:

pip install transformers==4.34.0