LocalForge
Voltar ao guia
MODEL / GOOGLE

Gemma 3 4B

Vision support depends on runtime and projector compatibility.

Fonte primária
Parâmetros totais4B
Memória estimada~4.2 GBINT4 / GGUF Q4
Contexto128KText + Image
LicenseGemma140+ languages
01 / IDEAL PARA
  • vision
  • document Q&A
  • multilingual chat
02 / AJUSTE ESTIMADO
8 GB CPU laptopTight16 GB CPU desktopComfortableApple silicon · 16 GBComfortableNVIDIA · 8 GBComfortableNVIDIA · 12 GBComfortable
03 / RODAR COM
OllamaEasyollama run gemma3:4b
llama.cppMediumllama-server -m /path/to/verified-model.gguf --port 8080
LM StudioEasySearch “google/gemma-3-4b-it” → Load a Q4/INT4 build → Start Server
MLX LMMediummlx_lm.chat --model google/gemma-3-4b-it
Atenção

Vision support depends on runtime and projector compatibility.