Posting generated or AI-edited text is against HN guidelines.

The author didn't even take the care to format the Huggingface README properly, look at the broken output formatting in https://huggingface.co/AmSach/DeepSeek-Coder-V2-Lite-Instruc...

If you're dealing with 6GB of VRAM just go for a smaller model, nothing wrong with Qwen 3.5-4B. It'll outperform this AI psychosis project.