|
Download README.md from OrionLLM/OpenAgentInstruct: direct link, hf CLI and curl.
- Browser
- Download file 662 Bytes
-
https://huggingface.co/datasets/OrionLLM/OpenAgentInstruct/resolve/main/README.md
- Command line
-
hf download hf://datasets/OrionLLM/OpenAgentInstruct/README.md
-
curl -L -o README.md https://huggingface.co/datasets/OrionLLM/OpenAgentInstruct/resolve/main/README.md
662 Bytes
| dataset_info: | |
| features: | |
| - name: conversations | |
| list: | |
| - name: content | |
| dtype: string | |
| - name: role | |
| dtype: string | |
| splits: | |
| - name: train | |
| num_bytes: 430031471 | |
| num_examples: 15209 | |
| download_size: 430095010 | |
| dataset_size: 430031471 | |
| configs: | |
| - config_name: default | |
| data_files: | |
| - split: train | |
| path: data/train-* | |
| license: apache-2.0 | |
| size_categories: | |
| - 10K<n<100K | |
| # Welcome to OpenAgentInstruct | |
| **OpenAgentInstruct** is a chat-style **Supervised Fine-Tuning (SFT)** dataset built for **terminal / command-line response**. | |
| - **Rows:** 15,209 (train split only) | |
| - **Size:** ~430 MB | |
| - **Language:** English (primarily) |