Dataset · newfacade
Python LeetCode Dataset for LLMs
newfacade/LeetCodeDataset
Python LeetCode problems for training and evaluating code LLMs; English text, 1k-10k examples.
What's inside
A set of LeetCode problems in Python for training and evaluating large language models. The card references papers on time estimation and efficient training of code LLMs, as well as policy filtering for RLHF. Language is English, modality is text, and size ranges from 1k to 10k examples.
What you can do with it
- Train code LLMs to solve Python problems
- Evaluate code generation on LeetCode
- Use for RLHF and response filtering
- Tasks
- Text generation
- Languages
- English
- Size
- 1K–10K rows
- Downloads this month
- 17.2K
- Likes
- 87
Models trained on it
- mradermacher/Frank-26B-A4B-i1-GGUF 2.3K
- mradermacher/Frank-26B-A4B-GGUF 1.1K
- TendieLabs/Frank-26B-A4B-GGUFsText generation 845
- F-A-I-L/mini-agent-v2-GGUFText generation 526
- ahmetggg/Drzeon-granite-4-2-8b-stem-finetuneText generation 473
- F-A-I-L/mini-agent-v2Text generation 135
- gianniiii/Qwen2.5-Coder-7B-Instruct-GRPO-LeetCodeText generation 47
- TendieLabs/Frank-26B-A4BText generation 28
How to load it
from datasets import load_dataset
ds = load_dataset("newfacade/LeetCodeDataset")