3 ms·
I uploaded 4bit bitsandbytes, GGUFs and original 16bit weights to https://huggingface.co/unsloth https://huggingface.co/unsloth for those interested! You can al
by danielhanchen 2y ago
I uploaded 4bit bitsandbytes, GGUFs and original 16bit weights to https://huggingface.co/unsloth https://huggingface.co/unsloth for those interested! You can also finetune Llama 3.3 70B in under 48GB of VRAM and 2x faster and use 70% less memory with Unsloth!