Cherry
Cherry
> 没有监督数据，建议使用大语言模型构造数据,可以参考[https://docs.llamaindex.ai/en/stable/examples/finetuning/embeddings/finetune_embedding.html。只做无监督微调难以让模型更适应领域数据。](https://docs.llamaindex.ai/en/stable/examples/finetuning/embeddings/finetune_embedding.html%E3%80%82%E5%8F%AA%E5%81%9A%E6%97%A0%E7%9B%91%E7%9D%A3%E5%BE%AE%E8%B0%83%E9%9A%BE%E4%BB%A5%E8%AE%A9%E6%A8%A1%E5%9E%8B%E6%9B%B4%E9%80%82%E5%BA%94%E9%A2%86%E5%9F%9F%E6%95%B0%E6%8D%AE%E3%80%82) 链接失效了，能重新发一下吗
> 可以提供一下训练命令方便分析。 训练指令是这样的： python /home/wk/wkstation/retrival/FlagEmbedding-master/FlagEmbedding/baai_general_embedding/finetune/run.py --output_dir /home/wk/wkstation/models/BAAI/finetuned/medical_bge-large-zh-v1.5 --model_name_or_path /home/wk/wkstation/models/BAAI/bge-large-zh-v1.5/Xorbits/bge-large-zh-v1.5 --train_data /home/wk/wkstation/retrival/Multi-CPR-main/data/my_data/medical_train.jsonl --learning_rate 1e-5 --fp16 --num_train_epochs 10 --per_device_train_batch_size 256 --dataloader_drop_last True --normlized True --temperature 0.02 --query_max_len 64 --passage_max_len 256 --train_group_size 9...
**请问运行到8808报这样的错怎么改啊** Traceback (most recent call last): File "E:/12.论文代码/QASystemOnMedicalKG-master/build_medicalgraph.py", line 271, in handler.create_graphnodes() File "E:/12.论文代码/QASystemOnMedicalKG-master/build_medicalgraph.py", line 185, in create_graphnodes self.create_diseases_nodes(disease_infos) File "E:/12.论文代码/QASystemOnMedicalKG-master/build_medicalgraph.py", line 177, in create_diseases_nodes self.g.create(node) File "F:\anaconda\envs\KBQA\lib\site-packages\py2neo\graph.py", line 389,...
> Hi @chensisi0730 , `nvcc -V` and `python -c "import torch; print(torch.version.cuda)"` should give a consistent version. In your case, installing cuda 11.6 in your own virtual environment like https://anaconda.org/nvidia/cuda-toolkit...
> Hi, may I see your `colossalai check -i` output?  I encounter the same error，here is my `colossalai check -i` output
> I installed by pip install colossalai ，but when i run the example I encounter a problem “ImportError: cannot import name 'zero_model_wrapper' from 'colossalai.zero'” so I installed by git clone...