Unfortunately, a single NVIDIA GeForce RTX 4090 24GB GDDR6X just wont cut it for DeepSeek-V3. The VRAM demands are too high for quality quantization.Y...
Honestly, for Spanish and Japanese, my current setup is basically just messing with LLM prompts. I tried the dedicated tutor apps, but they get pricey...