mattbucci
/

gemma-4-26B-AWQ

4-bit precision

Model card Files Files and versions

gemma-4-26B-AWQ

16.8 GB

Ctrl+K

Ctrl+K

1 contributor

History: 12 commits

mattbucci's picture

Fix chat_template: close model turn after tool_call + tool_response (agentic tool-call fix)

cfe71af verified 4 months ago

.gitattributes

1.57 kB
Gemma 4 26B MoE AWQ: forced-routing GPTQ calibration for all 128 experts, 30 tok/s on RDNA4 5 months ago
README.md

1.99 kB
Update vision status: untestable due to server crash 5 months ago
chat_template.jinja

17.3 kB
Fix chat_template: close model turn after tool_call + tool_response (agentic tool-call fix) 4 months ago
config.json

4.08 kB
add quantization_config.ignore=1 (downstream audit fix) 5 months ago
generation_config.json

203 Bytes
Gemma 4 26B MoE AWQ: forced-routing GPTQ calibration for all 128 experts, 30 tok/s on RDNA4 5 months ago
model-00001-of-00001.safetensors

15.7 GB
xet

Replace INT4 vision with BF16 originals (model-00001-of-00001.safetensors) 5 months ago
model-vision.safetensors

1.15 GB
xet

Replace INT4 vision with BF16 originals (model-vision.safetensors) 5 months ago
model.safetensors.index.json

3.64 MB
Replace INT4 vision with BF16 originals (model.safetensors.index.json) 5 months ago
processor_config.json

1.69 kB
Add files using upload-large-folder tool 5 months ago
tokenizer.json

32.2 MB
xet

Gemma 4 26B MoE AWQ: forced-routing GPTQ calibration for all 128 experts, 30 tok/s on RDNA4 5 months ago
tokenizer_config.json

2.69 kB
Gemma 4 26B MoE AWQ: forced-routing GPTQ calibration for all 128 experts, 30 tok/s on RDNA4 5 months ago