Quantization-Aware Training: When the Model Learns Its Own Precision
Why Kimi K2 Thinking ships in INT4 and gpt-oss and K3 in MXFP4: training with rounding in the loop, explained and measured on a small model.
qat
1 post
1 post
Why Kimi K2 Thinking ships in INT4 and gpt-oss and K3 in MXFP4: training with rounding in the loop, explained and measured on a small model.