Gemma-3-12B
Requested: Gemma-3-12B / Current WebLLM build maps to Gemma-2-9B fallback.
TransPackgenerator
WebGPU + WebLLM Edition
Upload .jar/.zip and build a translated Minecraft resource pack fully in-browser.
Step 1: Select .jar / .zip / .json
Drag and drop mod(.jar), shader pack(.zip), or lang json(.json)
Step 2: Translation target language
Step 3: Translation settings
For low-end PCs, 30 lines / 1 job is recommended.
Step 4: Model selection
Model download progress
Not started
Translation progress
0 / 0 lines
Task status
Waiting...
Model guide (recommended specs)
Gemma-2-2B
Estimated download: ~1.9GB / Recommended VRAM: 4-6GB / RAM: 16GB+
Gemma-4-9B
Requested: Gemma-4-9B / Current WebLLM build maps to Gemma-4-E2B fallback.
Gemma-4-E2B
Estimated download: ~7.2GB / Recommended VRAM: 10GB+ / RAM: 32GB+
Qwen-3.5-14B
Requested: Qwen-3.5-14B / Current WebLLM build maps to Qwen3.5-9B fallback.
Gemma-3-4B-IT
Estimated download: ~2.2GB / Recommended VRAM: 4-6GB / RAM: 16GB+
Llama-3.2-3B-Instruct
Estimated download: ~1.8GB / Recommended VRAM: 3-5GB / RAM: 16GB+
Qwen3-4b-2507
Estimated download: ~2.5GB / Recommended VRAM: 4-7GB / RAM: 16GB+
Qwen-2.5-7B-Instruct
Estimated download: ~4.3GB / Recommended VRAM: 7-10GB / RAM: 24GB+
Note: Actual memory usage varies by browser, driver, and context length.