Downloads · 30 days
0
openevo-recovery/openevo-h144-qwen25-3b-webshop
openevo-h144-qwen25-3b-webshop is a machine learning model from openevo-recovery. Use it for the machine learning task on the model card, and read the license before you ship it in a product. The card lists the license as other.
这是 OpenEVO 跨平台实验资产的一部分。本段只补充统一身份与导航信息,不修改仓库已有模型、数据、checkpoint、trajectory 或历史科学语义。
Downloads · 30 days
0
Access
Public
Updated Sep 28, 2026
Repo size
—
Likes
0
Public
Click a slice to open those files.
.json43.9 KB · 82%
From the Hugging Face model README
这是 OpenEVO 跨平台实验资产的一部分。本段只补充统一身份与导航信息,不修改仓库已有模型、数据、checkpoint、trajectory 或历史科学语义。
miyuki17/openevo-h144-qwen25-3b-webshopLEGACY_UNRESOLVEDlegacy-unresolvedunresolvedpubliclegacy-unresolvedPRESERVE_PENDING_RESOLUTIONmykcs/openevo-experiment@bb57c88eb5de3036dee3ea9b095bc550dd6882c8mykcs/openevo-experiment@a1994980f16a922374f25442a599031825b2cf03命名和语言约定:repo、title、canonical ID 使用英文;README 默认中文。本“中文说明(默认)”是默认阅读入口,下方已有英文说明作为 English companion / historical detail 保留。若本段与 immutable publication receipt 或精确 remote revision/hash 冲突,以后者为准。
English: This block adds cross-platform identity metadata only. Existing artifact bytes and scientific claims are unchanged.
<!-- openevo-cross-platform-identity:end -->导航 / Navigation:OpenEvo WebShop 公开科研产物
这是 H1.44-ON-3B 的公开 checkpoint 元数据/指针仓库。这里记录的是一个历史进度快照,不是当前服务器实时状态,也不是已经训练完成的独立模型。
仓库已发布的状态停在 rollout 2,976:Stage 1 与 Stage-2 blocks 0–5 都闭合为 no-update,因此这里只保留 checkpoint/recovery manifest 与 policy pointer,没有虚构或重复上传 adapter 权重。对应策略仍是固定的 Qwen/Qwen2.5-3B-Instruct 基座 revision aa8e72537993ba99e69dfaafa59ed015b17504d1;基础模型权重不在本仓库重复保存。
固定 128-task formal panel 没有用于 checkpoint 选择。这些文件只能作为历史进度与恢复证据,不能被解释为最终 adaptation efficacy、held-out transfer 或 OpenEvo-vs-SEED 结论。
This repository preserves a public checkpoint-publication snapshot for H1.44-ON-3B. It is a historical metadata/pointer record, not a statement that H1.44 is currently running.
At the snapshot published here (rollout 2,976), Stage 1 and Stage-2 blocks 0–5 had closed with no adapter update. Therefore this batch contains checkpoint/recovery manifests and policy pointers, not invented adapter weights. The policy at every published pointer is still the pinned Qwen/Qwen2.5-3B-Instruct base at revision aa8e72537993ba99e69dfaafa59ed015b17504d1; the unchanged base weights are not re-uploaded.
The formal 128-task test panel has not been used for checkpoint selection. These files are progress/recovery evidence only and do not establish adaptation efficacy, held-out transfer, a final-model result, or an OpenEvo-versus-SEED conclusion.
The experiment pins OpenEvo stable@796e6248a61112a61c880885844d3b0420c0b6ae and the H1.44 schedule SHA-256 1b3ac4252cda4d25d9542713e68e2fa4aacd90ceb377e5c03c8c1ad7c6878428. See publication-snapshot.json, each checkpoint manifest, and provenance/ for immutable source-state identities.
Qwen notice: this experiment is built with / intended to improve using Qwen. The referenced Qwen2.5-3B base remains subject to its applicable Qwen Research License terms. No Qwen base-model weights are redistributed here.