2、配置镜像 export HF_ENDPOINT=hf-mirror.com 3、下载模型到缓存 huggingface-cli download repo_id 4、下载模型到指定目录 huggingface-cli download repo_id --local-dir /path/to/model 5、扫描缓存 huggingface-cli scan-cache 6、清理缓存 huggingface-cli delete-cache ...
Each option can be implemented separately in a different PR. CLI implementation can be found in./commands/delete_cache.pywhile the cache scan tool itself is in./utils/_cache_manager.py We are interested in cleaning cached models which were not accessed in the past N days. Would it make se...
--local-dir-use-symlinks False表示不在默认路径~/.cache/huggingface/hub中保存模型,如果设为true,则保存在默认路径,指定目录只是超链接。 bug:huggingface-cli: error: argument {env,login,whoami,logout,repo,lfs-enable-largefiles,lfs-multipart-upload,scan-cache,delete-cache}: invalid choice: 'download'...
As an example here's the last two checkpoints when i run the huggingface-cli scan-cache command: Regarding thedelete-cachecommand, this is expected since you only have 1 revision per model (snapshotf4b0486b498f84668e828044de1d0c8ba486e05bfor xl-base-1.0 and snapshot462165984030d82259a11f4...
正确的命令列表应该包括 env, login, whoami, logout, repo, lfs-enable-largefiles, lfs-multipart-upload, scan-cache, delete-cache 等。检查用户输入的命令: 你输入的命令 huggingface-cli download 是错误的,因为 download 并不是 huggingface-cli 的一个有效命令。
$ huggingface-cli upload --repo-type [repo_type] [repo_id] [path/to/local_file_or_directory] [path/to/repo_file_or_directory] Scan cache to see downloaded repositories and their disk usage $ huggingface-cli scan-cache Delete the cache interactively ...
usage: huggingface-cli [<args>] huggingface-cli: error: invalid choice: 'download' (choose from 'env', 'login', 'whoami', 'logout', 'repo', 'lfs-enable-largefiles', 'lfs-multipart-upload', 'scan-cache', 'delete-cache') @absalanYes...
huggingface-cli login 登录成功后,可以使用create_repo()函数创建新的repo。 from huggingface_hub import create_repo repo_url = create_repo(name="github-issues", repo_type="dataset") repo_url 'https://huggingface.co/datasets/lewtun/github-issues' ...
EDIT (before reading below): motivation is the same but the plan has been slightly improved (see this comment). Basically, the plan is: add delete_patterns in upload_folder implement commit_in_chunks to chained commits in a PR implement ...