Dataset · allenai
Tulu V2 SFT Mixture
allenai/tulu-v2-sft-mixture
English instruction data mixture for SFT of Tulu language assistants.
What's inside
This is a data mixture for training Tulu language assistant models. The card lists FLAN (50k examples from FLAN v2) and another 50k for chain-of-thought reasoning; it uses the ODC-BY license, and some subsets may be non-commercial. The dataset is presented as a research artifact in English with a size of 100K–1M.
What you can do with it
- SFT of language assistant models
- Training chain-of-thought reasoning
- Research experiments with Tulu
- Tasks
- Question answering, Text generation
- Languages
- English
- Size
- 100K–1M rows
- Downloads this month
- 1.4K
- Likes
- 140
Models trained on it
- bartowski/Tralalabs_Pythia-6.9B-Instruct-v1-Merged-GGUFText generation 1.8K
- mradermacher/llama-3-tulu-2-8b-i1-GGUF 1.1K
- allenai/OLMo-7B-Instruct-hfText generation 1.1K
- allenai/tulu-2-13bText generation 1.1K
- mradermacher/llama-3-tulu-v2.5-8b-uf-mean-70b-uf-rm-mixed-prompts-i1-GGUF 1K
- mradermacher/tulu-2-70b-i1-GGUF 956
- mradermacher/pythia-1.4b-tulu-v2-mix-GGUF 920
- mradermacher/tulu-2-70b-GGUF 861
How to load it
from datasets import load_dataset
ds = load_dataset("allenai/tulu-v2-sft-mixture")