s1gnature/ gemma4-mtp:e2b-q4_k_m

1,083 2 weeks ago

Gemma 4 E4B & E2B (Q4_K_M) with MTP drafter. 32k context, up to 80 TPS on RTX 3050 8GB.

vision tools thinking
ollama run s1gnature/gemma4-mtp:e2b-q4_k_m

Details

2 weeks ago

b53f2f48c387 · 4.3GB

gemma4
·
4.65B
·
Q4_K_M
clip
·
476M
·
F16
Think step by step in a reasoning block, then act. Use the provided tools when they help. Be concise
{ "draft_num_predict": 2, "num_ctx": 32000, "stop": [ "<turn|>" ] }
gemma4-assistant
·
77.2M
·
F16
{{ if .System }}<bos><|turn>system {{ .System }}<turn|> {{ end }}{{ if .Prompt }}<|turn>user {{ .Pro

Readme

Gemma 4 E4B & E2B MTP (Q4_K_M)

Gemma 4 E4B & E2B builds with MTP drafter for faster inference.

Specs

  • Quantizations: Q4_K_M, Q8_0
  • Context length: 32,000 tokens
  • Speed: Up to 80 TPS on RTX 3050 8GB OC

Credits