turintech/ gemma-4-e2b-it-artemis:iq3_xxs

54 4 hours ago

Text-only Gemma 4 E2B GGUF files. Default: q3_k_l. TurinTech models, built with Artemis.

tools thinking 
ollama run turintech/gemma-4-e2b-it-artemis:iq3_xxs

Details

4 hours ago

c3df9203d0e7 · 2.1GB

model
gemma4
·
4.63B
·
Q4_0
license
Apache License Version 2.0, January 2004 http://www.apache.org/licenses/ TERMS AND CONDITIONS FOR US
license
Apache License Version 2.0, January 2004 http://www.apache.org/licenses/ TERMS AND CONDITIONS FOR US
params
{ "stop": [ "<bos>", "<|turn>", "<turn|>", "<|turn>user" ],
template
{{ if .System }}<bos><|turn>system {{ .System }}<turn|> {{ end }}{{ if .Prompt }}<|turn>user {{ .Pro

Readme

Gemma 4 E2B Artemis

Text-only GGUF files of Google’s quantization-aware (QAT) Gemma 4 E2B checkpoint. This is an independent conversion, not affiliated with or endorsed by Google.

TurinTech models, built with Artemis

Ollama

ollama run turintech/gemma-4-e2b-it-artemis:q3_k_l

latest selects q3_k_l. The explicit tags below use the same model bytes as the corresponding Hugging Face files. Existing size tags remain available.

Tag Hugging Face file Size
q3_k_l (default) gemma-4-E2B-it-Artemis-Q3_K_L.gguf 2.308 GB
iq3_xxs gemma-4-E2B-it-Artemis-IQ3_XXS.gguf 2.052 GB
q3_k_m gemma-4-E2B-it-Artemis-Q3_K_M.gguf 2.198 GB

GB is download size, not runtime memory. Quant labels are size classes for these mixed files; the model card lists tensor types.

Packages use Ollama’s Gemma 4 renderer and parser, with temperature 1, top_k 64 and top_p 0.95. Set context length for your memory budget.

Measurements and limitations

The default loses 3 of 16 maths questions in the task sample. If a 2.6 GB file fits, Unsloth’s UD-Q4_K_XL of the same checkpoint is closer to the original. Measurements used llama.cpp, not Ollama.

Full measurements and test conditions.

Licence

Apache-2.0 (licence). See the Hugging Face repository for LICENSE and NOTICE.

These files keep improving.