Show HN: DeepSeek-Coder-V2-Lite Under 5GB: 65 TPS 100% In-VRAM on 6GB GPU
huggingface.co
1 thread
Looks potentially useful for low VRAM systems so I'm curious as to why the currently 2 comments here are being flagged without any bothering to justify the flagging. Maybe someone will care to educate me?
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.
noted. thanks for taking the trouble.