Is this version still the same as previous Kimi K2 w.r.t. quantization: trained on BF16 and inference on FP8?
· Sign up or log in to comment
