Qwen/Qwen2.5-VL-32B-Instruct · Loading Error: Unrecognized model type 'qwen2_5

4 days ago

I'm getting a ValueError when loading the model:
ValueError: Unrecognized configuration class <class 'transformers.models.qwen2_5_vl.configuration_qwen2_5_vl.Qwen2_5_VLConfig'> for this kind of AutoModel: AutoModelForCausalLM.

I've followed the instructions, installed transformers and accelerate from the provided GitHub repo, and set trust_remote_code=True, but the error persists.

Is there something specific I might've missed or an additional step required?

Thanks!

bluelike

Qwen org 3 days ago

Hello, Qwen2_5_VL does not belong to AutoModelForCausalLM and cannot be directly loaded via AutoModel. Instead, it should be loaded using the specific class as shown below:

from transformers import Qwen2_5_VLForConditionalGeneration, AutoTokenizer, AutoProcessor
from qwen_vl_utils import process_vision_info

# Default: Load the model on the available device(s)
model = Qwen2_5_VLForConditionalGeneration.from_pretrained(
    "Qwen/Qwen2.5-VL-32B-Instruct", torch_dtype="auto", device_map="auto"
)

# We recommend enabling flash_attention_2 for better acceleration and memory saving, especially in multi-image and video scenarios.
# model = Qwen2_5_VLForConditionalGeneration.from_pretrained(
#     "Qwen/Qwen2.5-VL-32B-Instruct",
#     torch_dtype=torch.bfloat16,
#     attn_implementation="flash_attention_2",
#     device_map="auto",
# )

kentrolla

3 days ago

Thank you!

kentrolla changed discussion status to closed 3 days ago

Qwen
/

Qwen2.5-VL-32B-Instruct

Loading Error: Unrecognized model type 'qwen2_5_vl'