5Gemma 3 QAT (Quantized Aware Training) 3x less memoryAmazing, I've been wishing for this! Do you have any estimates on how much accuracy is first lost then recovered compared to the original bf16 and the naively quantized models?Thank you so much for continuing to support Gemma 3 with these updates.
Amazing, I've been wishing for this! Do you have any estimates on how much accuracy is first lost then recovered compared to the original bf16 and the naively quantized models?
Amazing, I've been wishing for this! Do you have any estimates on how much accuracy is first lost then recovered compared to the original bf16 and the naively quantized models?
Thank you so much for continuing to support Gemma 3 with these updates.