Added IQ1_S version to Ollama
#4
by
Muhammadreza
- opened
This is the Ollama merge: https://ollama.com/haghiri/DeepSeek-V3-0324
Works fine on H200. If you tested it on another devices, please let me know.
Why only IQ1_S? and not other quantization?
Why only IQ1_S? and not other quantization?
im guessing its because it's too big
@shimmyshimmer
is right.
I merged this quant because it is good and fitting.