SKU SFT-2002 · Sold by External
OpenHermes 2.5
Product specifications
| SKU | SFT-2002 |
|---|---|
| Data type | SFT dataset |
| Volume | 1001551 rows |
| Size on disk | Not published by source |
| Format | Parquet (Hugging Face) |
| Access model | PUBLIC LICENSE |
| Pricing | Free · open-source license |
| Quality score | — |
| License | Unverified — see source |
A large open compilation of instruction and chat data curated by Teknium to train the OpenHermes 2.5 models. It aggregates and cleans many public/synthetic sources including Airoboros, CamelAI, WizardLM Evol-Instruct, Glaive code, GPTeacher, SlimOrca, Platypus and MetaMath, spanning general chat, coding, math/chain-of-thought, roleplay and science. Rows use the ShareGPT conversations schema (list of {from, value} turns) plus source and category metadata.