Hi dear
Thank you for such excellent work and I am very excited to follow it。But I have a problem when I follow quickstart in README.
Specifically, when running to model = AutoModel.from_pretrained(ckpt_path, trust_remote_code=True).cuda().half(), the following error Caught signal 11 (Segmentation fault: address not mapped to object at address (nil)) will be reported and blocked.
Hi dear Thank you for such excellent work and I am very excited to follow it。But I have a problem when I follow quickstart in README. Specifically, when running to
model = AutoModel.from_pretrained(ckpt_path, trust_remote_code=True).cuda().half()
, the following errorCaught signal 11 (Segmentation fault: address not mapped to object at address (nil))
will be reported and blocked.my env bulit with
conda create -n songcomposer python=3.9.19 conda activate songsomposer pip install torch==2.1.0 torchvision==0.16.0 torchaudio==2.1.0 --index-url https://download.pytorch.org/whl/cu121 pip install transformers==4.31.0 pip install sentencepiece==0.2.0
A800 NVIDIA-SMI 525.60.13 Driver Version: 525.60.13 CUDA Version: 12.2 thx