Sign inGet started
Dataset · Hugging Face

ar-mine/deepstack-post-second-timing-handoff-20261021

DeepStack 后置第二路门控实验 这个公开交付包用于复现“E5+C3 已训练完成后,再训练投影+第二路门控”。源端四张 A100 的 8 步 smoke 已通过并独立审计:作业 25895355,固定 E5+C3 25723372/step17180。接收端 agent 请先读 AGENT_TASK.md,再用 START_HERE.md 的固定 commit、SHA 和完整下载命令部署。 本包提供部署脚本、严格参考张量、源端验收证据和训练入口。公共代码、冻结转换数据与父权重从 dependencies.json 指定的不可变公开版本下载,避免重复上传同一份大文件;接收不需要 NSCC 账号。公共基础包原来的“三路训练任务说明”不适用于本任务,接收端只运行本包覆盖后的 WORK/tools/run.py。 实验设置 项目 固定配置 父模型 E5+C3 25723372,step17180,精确公开 revision 5d14687980321be699cc8d748b25b8195ea80f08… See the full description on the dataset page: https://huggingface.co/datasets/ar-mine/deepstack-post-second-timing-handoff-20261021.

Author / organizationar-mine
Downloads0 (as reported by Hugging Face)
Likes0
Created21 October 2026
Last updated at source21 October 2026
roboticsreproducibilitytraining-handoff
Embed this entry
<iframe src="https://silytics.si/embed/item/ecbcd456-65c8-4189-b779-ba2b5b41ef01" width="420" height="190" style="border:0" loading="lazy" title="Silytics entry"></iframe>

Free, no tracking. More about feeds and embeds.

Recent changes

  • 21 Oct 2026 Description: none → DeepStack 后置第二路门控实验 这个公开交付包用于复现“E5+C3 已训练完成后,再训练投影+第二路门控”。源端四张 A100 的 8 步 smoke 已通过并独立审计:作业 25895355,固定 E5+C3 25723372/step17180。接收端 agent 请先读 AGENT_TASK.md,再用 START_HERE.md 的固定 commit、SHA 和完整下载命令部署。 本包提供部署脚本、严格参考张量、源端验收证据和训练入口。公共代码、冻结转换数据与
  • 21 Oct 2026 Topics added robotics, reproducibility, training-handoff

Changes detected by Silytics between refreshes of the source.

More from ar-mine

Silytics shows public metadata from Hugging Face and links to the original. Information may be incomplete or out of date. Last refreshed 21 October 2026. Licenses and terms of the original apply to the work itself. Not financial, legal or professional advice.