[BUG/Help] <title>一加载python就崩溃 #1511
Description
Is there an existing issue for this?
- I have searched the existing issues
Current Behavior
2025-08-26 17:36:55,253 [INFO] AD_portal.ai_utils: 开始加载 ChatGLM 模型...
loading file tokenizer.model
loading file added_tokens.json
loading file special_tokens_map.json
loading file tokenizer_config.json
loading file tokenizer.json
loading file chat_template.jinja 一运行到这里就弹出对话框python遇到问题，我本地模型文件夹内容如下2025/08/26 17:14
2025/08/26 17:14 ..
2025/08/26 17:14 298 chat_template.jinja
2025/08/20 09:25 1,317 config.json
2025/08/21 10:43 2,332 configuration_chatglm.py
2025/08/20 10:24 1,827,774,160 model-00001-of-00007.safetensors
2025/08/20 10:58 1,968,291,888 model-00002-of-00007.safetensors
2025/08/20 11:17 1,927,406,936 model-00003-of-00007.safetensors
2025/08/20 10:08 1,815,217,640 model-00004-of-00007.safetensors
2025/08/20 11:37 1,968,291,920 model-00005-of-00007.safetensors
2025/08/20 15:35 1,927,406,960 model-00006-of-00007.safetensors
2025/08/20 15:19 1,052,805,400 model-00007-of-00007.safetensors
2025/08/20 09:30 21,249 model.safetensors.index.json
2025/08/20 09:27 56,521 modeling_chatglm.py
2025/08/20 09:27 14,687 quantization.py
2025/08/20 09:26 3 special_tokens_map.json
2025/08/20 09:25 13,096 tokenization_chatglm.py
2025/08/20 09:26 1,018,370 tokenizer.model
2025/08/21 11:00 1,395 tokenizer_config.json
17 个文件 12,488,324,172 字节
2 个目录 13,331,349,504 可用字节
在CPU环境下运行的。 用T5模型没有崩溃
Expected Behavior
No response
Steps To Reproduce
windows server 2016
Python 3.11.2
faiss-cpu 1.7.4
sentence-transformers 5.1.0
torch 2.8.0
transformers 4.55.4
CPU环境下
Environment
- OS:
- Python:
- Transformers:
- PyTorch:
- CUDA Support (`python -c "import torch; print(torch.cuda.is_available())"`) :Anything else?
No response