too slow for model build for large files

[tianweidut]

Describe the bug
128GB files use 50mins.

[2023-07-19 22:44:00.996599] 👾 verbosity: 3, log level: DEBUG
[2023-07-19 22:44:01.012767] 🚧 start to build model bundle...
[2023-07-19 22:44:01.014042] ❓ |WARN| refine resource[ResourceType.model] llama2-70b-chat/version/3jbcmkiboq6l2r4dejtuvqxextquggrt7gz4kxzf failed: Can not find the exact match item 3jbcmkiboq6l2r4dejtuvqxextquggrt7gz4kxzf, found: []
[2023-07-19 22:44:01.015152] 👷 uri local/project/self/model/llama2-70b-chat/version/3jbcmkiboq6l2r4dejtuvqxextquggrt7gz4kxzf
[2023-07-19 22:44:01.017158] 🔈 |DEBUG| 🆕 version 3jbcmkiboq6l
[2023-07-19 22:44:01.017947] 📁 workdir: /home/liutianwei/.starwhale/.tmp/tmp4zfohi_0
[2023-07-19 22:44:01.018512] 🦚 copy source code files: /mnt/data/tianwei/code/starwhale/example/LLM/llama2 -> /home/liutianwei/.starwhale/.tmp/tmp4zfohi_0/src
[2023-07-19 22:44:01.019161] 🔈 |DEBUG| copy dir: /mnt/data/tianwei/code/starwhale/example/LLM/llama2 -> /home/liutianwei/.starwhale/.tmp/tmp4zfohi_0/src, excludes: ['*/.git/*', '', '*/adapter-7b/checkpoint*', '*/adapter-7b/runs', '*/base-13b/*', '*/adapter-13b/*', '*/base-30b/*', '*/adapter-30b/*', '*/base-65b/*', '*/adapter-65b/*', '*/llama-2-7b/*', '*/llama-2-7b-chat/*', '*/llama-2-13b/*', '*/llama-2-13b-chat/*', '*/llama-2-70b/*', '__pycache__/', '*.py', '*$py.class']
[2023-07-19 23:15:59.146552] 📁 source code files size: 128.50GB
[2023-07-19 23:15:59.148892] 🔈 |DEBUG| generating model serving config for local/project/self/model/llama2-70b-chat/version/3jbcmkiboq6l2r4dejtuvqxextquggrt7gz4kxzf ...
[2023-07-19 23:15:59.515018] 🚀 generate jobs yaml from modules: ['evaluation'] , package rootdir: /mnt/data/tianwei/code/starwhale/example/LLM/llama2
[2023-07-19 23:34:39.556491] 💡 |INFO| 🧺 resource files size: 128.50GB
[2023-07-19 23:34:39.595956] 💯 finish gen resource @ /home/liutianwei/.starwhale/self/model/llama2-70b-chat/3j/3jbcmkiboq6l2r4dejtuvqxextquggrt7gz4kxzf.swmp