search

jingyaogong / minimind-v

「大模型」3小时从0训练27M参数的视觉多模态VLM，个人显卡即可推理训练！
https://jingyaogong.github.io/minimind-v
Apache License 2.0
250 stars 27 forks source link

关于预训练报错没有正确安装或初始化 Git LFS，但是已经安装并且初始化了 #12

Open Aether-star opened 1 week ago

Aether-star commented 1 week ago

(minimindv) root@autodl-container-651142bf34-3b6ccf84:~/minimind-v# sudo apt-get install git-lfs Reading package lists... Done Building dependency tree... Done Reading state information... Done git-lfs is already the newest version (3.0.2-1ubuntu0.2). 0 upgraded, 0 newly installed, 0 to remove and 99 not upgraded.

(minimindv) root@autodl-container-651142bf34-3b6ccf84:~/minimind-v# git lfs install Updated git hooks. Git LFS initialized.

(minimindv) root@autodl-container-651142bf34-3b6ccf84:~/minimind-v# git lfs pull

(minimindv) root@autodl-container-651142bf34-3b6ccf84:~/minimind-v# python 1-pretrain_vlm.py 模型可学习参数: 27.272192 百万 = 0.027272192 B (Billion) Traceback (most recent call last): File "/root/miniconda3/envs/minimindv/lib/python3.9/site-packages/transformers/modeling_utils.py", line 575, in load_state_dict return torch.load( File "/root/miniconda3/envs/minimindv/lib/python3.9/site-packages/torch/serialization.py", line 1027, in load raise pickle.UnpicklingError(UNSAFE_MESSAGE + str(e)) from None _pickle.UnpicklingError: Weights only load failed. Re-running torch.load with weights_only set to False will likely succeed, but it can result in arbitrary code execution.Do it only if you get the file from a trusted source. WeightsUnpickler error: Unsupported operand 118

During handling of the above exception, another exception occurred:

Traceback (most recent call last): File "/root/minimind-v/1-pretrain_vlm.py", line 197, in model, tokenizer, (vision_model, preprocess) = init_model(lm_config) File "/root/minimind-v/1-pretrain_vlm.py", line 131, in init_model (vision_model, preprocess) = get_vision_model() File "/root/minimind-v/model/vision_utils.py", line 13, in get_vision_model model = CLIPModel.from_pretrained(model_path) File "/root/miniconda3/envs/minimindv/lib/python3.9/site-packages/transformers/modeling_utils.py", line 3738, in from_pretrained state_dict = load_state_dict(resolved_archive_file) File "/root/miniconda3/envs/minimindv/lib/python3.9/site-packages/transformers/modeling_utils.py", line 585, in load_state_dict raise OSError( OSError: You seem to have cloned a repository without having git-lfs installed. Please install git-lfs and run git lfs install followed by git lfs pull in the folder you cloned.

jingyaogong commented 1 week ago

https://github.com/jingyaogong/minimind-v/issues/10

参考这个issue解决，试试是否可行😊

Aether-star commented 1 week ago

感谢博主的回复 已经解决了 问题出在clip模型的比较大，导致clone不下来，手动下载后出现了一些bug，于是我使用git lfs clone https://hf-mirror.com/openai/clip-vit-base-patch32命令，成功将clip模型克隆下来就好了。

jingyaogong commented 1 week ago

感谢博主的回复 已经解决了 问题出在clip模型的比较大，导致clone不下来，手动下载后出现了一些bug，于是我使用git lfs clone https://hf-mirror.com/openai/clip-vit-base-patch32命令，成功将clip模型克隆下来就好了。

ok, no thanks