[Usage] ImportError: cannot import name 'LlavaLlamaForCausalLM' from 'llava.model'

[Pompey21]

Describe the issue

Issue:
Hi, I've been trying to setup the project on my local machine and after downloading/installing all necessary pip libraries I tried running the exemplary test script (see below) but am running into problems.
I created a new python file in the home directory of the project - where llava folder is also present.

Also please note that I made a fork and names it deep-llava. Nothing has been changed yet.

Command:

from llava.model.builder import load_pretrained_model
from llava.mm_utils import get_model_name_from_path
from llava.eval.run_llava import eval_model

model_path = "liuhaotian/llava-v1.5-7b"

tokenizer, model, image_processor, context_len = load_pretrained_model(
    model_path=model_path,
    model_base=None,
    model_name=get_model_name_from_path(model_path)
)

Log:

Traceback (most recent call last):
  File "/path-on-my-machine/deep-llava/test.py", line 1, in <module>
    from llava.model.builder import load_pretrained_model
  File "/path-on-my-machine/deep-llava/llava/__init__.py", line 1, in <module>
    from .model import LlavaLlamaForCausalLM
ImportError: cannot import name 'LlavaLlamaForCausalLM' from 'llava.model' (/path-on-my-machine/deep-llava/llava/model/__init__.py)

[zhaohm14]

As discussed in issue #1101, I've tried the following package versions and they seem to work fine in my environment:

• transformers==4.37.2
• accelerate==0.28.0
• torch==2.1.2

[Pompey21]

Thank you, this was really helpful :)

[wentaoyuan]

As discussed in issue #1101, I've tried the following package versions and they seem to work fine in my environment:

• transformers==4.37.2
• accelerate==0.28.0
• torch==2.1.2

I have exactly the same versions but still running into ImportError: cannot import name 'LlavaLlamaForCausalLM' from 'llava.model'

[zhaohm14]

@wentaoyuan Maybe you can comment out the try...except... block in llava/model/__init__.py

# try:
from .language_model.llava_llama import LlavaLlamaForCausalLM, LlavaConfig
from .language_model.llava_mpt import LlavaMptForCausalLM, LlavaMptConfig
from .language_model.llava_mistral import LlavaMistralForCausalLM, LlavaMistralConfig
# except:
#     pass

to find out more detailed error information.
#1101 (comment)