TechxGenus/LeetCode-Contest

LeetCode Contest Benchmark A new benchmark for evaluating Code LLMs proposed by DeepSeek-Coder, which consists of the latest algorithm problems of different difficulties. Usage git clone https://github.com/deepseek-ai/DeepSeek-Coder.git cd Evaluation/LeetCode # Set the model or path here MODEL="deepseek-ai/deepseek-coder-7b-instruct" python vllm_inference.py --model_name_or_path $MODEL --saved_path output/20240121-Jul.deepseek-coder-7b-instruct.jsonl python… See the full description on the dataset page: https://huggingface.co/datasets/TechxGenus/LeetCode-Contest.