Comment on DeepSeek Coder: Let the Code Write ItselfparentComments−SparkyMcUnicorn2yLooks like even deepseek-coder 1.3b benchmarks higher than CodeLlama 70b.https://evalplus.github.io/leaderboard.html
Comments
Looks like even deepseek-coder 1.3b benchmarks higher than CodeLlama 70b.
https://evalplus.github.io/leaderboard.html