Hugging Face's logo

google
/
gemma-4-E4B-it

Any-to-Any
Transformers
Safetensors
gemma4
image-text-to-text
Eval Results
Model card Files
xet
 Community
43

Is there any quantization process done to this google/gemma-4-E4B-it model checkpoint?

#41
by yliu0777 - opened
Discussion
yliu0777
 Google org

Hi team,

I want to check and confirm if the model is a original bf16 model, or if there is any quantization processing done to it such as GPTQ weight quantization or quantization-aware training.
pannaga10
 Google org

Hi @yliu0777
The google/gemma-4-E4B-it repository contains the original, unquantized instruction-tuned model checkpoint.If you are looking for official Quantization-Aware Training (QAT) variants, Google has provided dedicated checkpoints for those instead. You can find the specific QAT models here .

  1. google/gemma-4-E4B-it-qat-w4a16-ct
  2. google/gemma-4-E4B-it-qat-q4_0-gguf
  3. google/gemma-4-E4B-it-qat-q4_0-unquantized

For the complete list, check out the official Hugging Face collection: https://huggingface.co/collections/google/gemma-4-qat-q4-0

Thanks

Sign up or log in to comment