Modelers 魔乐社区

modelers.cn · Storage · 1 record · JSON

Modelers 魔乐社区 (modelers.cn, Huawei/Ascend ecosystem) Storage

caveatemailno cardphoneIper-object-cap

modelers.cn · model-hub · storage/modelers · JSON · compare

Free tier	Not published. The community is an explicit Hugging-Face-shaped hub — models, datasets, and Spaces, plus its own client (openMind Hub Client) and library (openMind) targeting Ascend NPUs. Its contribution guide instructs 'git lfs track' for '100MB以上' files, which confirms a real LFS backend rather than a web-only uploader. No GB ceiling appears on the site or in the docs I could reach. unit GB; metered on unstated
Price	none stated
Per-object cap	No per-file cap, account quota, retention, or rate limit found
Retention	not stated
Egress	No egress meter documented; models download through both the web and the openMind client anonymously. CN-network served.
Identity	email email + +86 phone (Huawei-adjacent Chinese community); access token required for any write.
Limits	per-object-cap No per-file cap, account quota, retention, or rate limit found source Unverified. git-LFS 100 MB+ tracking threshold documented. No per-file cap, account quota, retention, or rate limit found. Upload requires an access token with Write scope created in the personal centre.
Binding	per-object-cap the per-object cap sets the shard size and therefore the object count; it bounds the total more tightly than any headline quota
Interface	git git
Tooling	First-party openMind Hub Client (pip) with upload/download by repo_id, plus git+LFS, plus a web UI. Documented at modelers.cn/docs/zh/openmind-hub-client. This is a genuine HF drop-in: repo types are 'model', 'dataset', 'space', same as HF.
Composes with	N repos per account, free. The strategic value is that it is the domestic Ascend/MindSpore hub, so it is the natural place a Chinese-ecosystem user would park multi-hundred-GB checkpoints.
Reliability	API surface is opaque: https://modelers.cn/api/v1/models returns 404, and model detail pages return a 1,638-byte SPA shell regardless of path — so anonymous programmatic access is effectively absent; you need the openMind client.
Notes	Not published. The community is an explicit Hugging-Face-shaped hub — models, datasets, and Spaces, plus its own client (openMind Hub Client) and library (openMind) targeting Ascend NPUs. Its contribution guide instructs 'git lfs track' for '100MB以上' files, which confirms a real LFS backend rather than a web-only uploader. No GB ceiling appears on the site or in the docs I could reach.
Evidence rung	inferred One page read moves it to documented.
Checked	2026-10-21 (inherited); stale on 2026-11-20

Evidence (3)

date	source	what it proved
2026-10-21	modelers.cn/docs/zh/best-practices/community_contribution/mo	read live: '在本地Git仓库中追踪大型文件（100MB以上）', git lfs upload flow, write-scoped access token required
2026-10-21	modelers.cn/	read live: 'Models', 'Spaces', 'openMind Library', 'openMind Hub Client — Easily search for, upload, and download pre-trained models'; dataset upload guide dated 2024-09-22
2026-10-21	modelers.cn/api/v1/models?per=2	probe: GET -> 404; GET https://modelers.cn/models/OpenMind/DeepSeek-R1-Distill-Qwen-7B -> 200 but 1,638-byte SPA shell; POST .../info/lfs/objects/batch -> 404 (returns the SPA shell, not LFS JSON)

Unknown or withheld values (4)

$.free_tier.notes process
$.per_object_cap.notes not-extracted
$.retention.notes not-extracted
$.retention.stated not-extracted

Built from commit 872967ae41, data as of 2026-10-27. Every value links to its source; freshness badges are computed in your browser. How to read this.