# Rui Zhu: Open-Weight Models and Datasets
Source: https://savrn.com/model-publishers/huthvincent
Markdown alternate of the page above; the site index is https://savrn.com/llms.txt

---

## Datasets

Dataset

### [btc15-dataset](https://savrn.com/datasets/btc15-dataset)

[Rui Zhu](https://savrn.com/model-publishers/huthvincent)

一个完整记录 Kalshi KXBTC15M 市场量化交易的公开数据集:模型对每个 15 分钟窗口的 适合研究:成交率建模(IOC vs 挂单)、预测校准、二元期权做市、sim-vs-live 差距。 按 machine= /date= /part.parquet 分片。每行 = 某秒的一次快照: tickiso, windowid, windowanchorts, secondsfromopen, btclast, btcopen, marketticker, eventticker, floorstrike, yesbid, yesask, nobid, tickts, machine, runid, kind … 含 BTC 现价/开盘价、Kalshi yes/no 最优四价、floor strike、抓取延迟(brti/kalshi/book/fetch)。 7×24 采集器额外带盘口深度列:yesbookjson / nobookjson(升序 [[价,美元量],…] 档位, 最优 bid 在末档;Kalshi 只报双边 bid,ask 用对偶价 yesask = 1 - nobestbid)、 yesbooklevels / nobooklevels(档数)、yesbidsize / nobidsize(最优档美元量)。 旧 feeder 行的深度列为空——按 yesbooklevels 非空过滤即可拿到带深度的行。 - 时间:ts 为 UTC epoch 秒(int),iso 为可读字符串。 - kind:sim(模拟,乐观成交)/ live(真实下单)/…

Publicly accessible mit

[View dataset](https://savrn.com/datasets/btc15-dataset)

## Explore More

- [All model publishers](https://savrn.com/model-publishers)
- [The model directory](https://savrn.com/models)
- [The dataset directory](https://savrn.com/datasets)

## Source

- Listed from their public repositories, read 2026-09-30.
- [Hugging Face profile](https://huggingface.co/huthvincent)
- [How the hub is built](https://savrn.com/model-hub/methodology)