Dataset · nlpai-lab
KULLM-v2 Korean instructions
nlpai-lab/kullm-v2
GPT4ALL, Dolly, and Vicuna translations into Korean for instruction-based model training.
What's inside
English instruction data from GPT4ALL, Dolly, and Vicuna translated into Korean via the DeepL API. The card specifies instruction, input, and output fields, an Apache-2.0 license, and a size ranging from 100K to 1M examples.
What you can do with it
- Supervised fine-tuning of Korean instruction models
- Training LoRA for Korean text generation
- Evaluating instruction following in Korean
- Tasks
- Text generation
- Languages
- Korean
- Size
- 100K–1M rows
- Downloads this month
- 664
- Likes
- 77
Models trained on it
- mradermacher/kosolra-kullm-LDCC-merge-i1-GGUF 887
- mradermacher/DataVortexTL-1.1B-v0.1-i1-GGUFText generation 610
- shopkeeper/kor-finewiki-1b-chatText generation 383
- mradermacher/DataVortexTL-1.1B-v0.1-GGUFText generation 265
- mradermacher/T3Q-ko-solar-sft-v2.0-i1-GGUF 234
- mradermacher/kosolra-kullm-LDCC-merge-GGUF 210
- mradermacher/T3Q-ko-solar-sft-v2.0-GGUF 154
- heegyu/llama-2-ko-7b-chatText generation 142
How to load it
from datasets import load_dataset
ds = load_dataset("nlpai-lab/kullm-v2")