Modelpublic

henrique-simoes/gemma-4-e2b-it-istara-ux-research-MLX-F16

sourceHugging Facegemmaupdated 5mo agoView on Hugging Face

1likes184downloads

Model Card

gemma-4-e2b-it-istara-ux-research MLX F16

Apple MLX F16 conversion of henrique-simoes/gemma-4-e2b-it-istara-ux-research, a merged Gemma 4 E2B instruction model adapted for UX research workflows in Istara.

This build materializes Gemma 4 shared K/V tensors for compatibility with stricter MLX loaders such as LM Studio's bundled MLX runtime.

Use

bash

mlx_lm.generate \
  --model henrique-simoes/gemma-4-e2b-it-istara-ux-research-MLX-4bit \
  --prompt "Draft a UX research plan for onboarding drop-off." \
  --max-tokens 256

Smoke Test

This artifact loaded locally with mlx-lm 0.31.2, generated successfully on Apple Silicon, and used about 9.4 GB peak memory in a short smoke run.