prithivMLmods/gemma-4-31B-it-qat-FP8

Quantization vs google/gemma-4-31B-it-qat-q4_0-unquantized. prithivMLmods/gemma-4-31B-it-qat-FP8 is a quantized version of the other (F8_E4M3/BF16).

On the Hub · Downloads on Model Pulse