fwizzer1/Fwizzer-R1-3B-RU-v2

Delete mmproj/Ministral-3-3B-Instruct-2512.F16-mmproj.gguf with huggingface_hub

Update chat_template.jinja with mandatory <think> prefill

Fix Fwizzer-R1.preset.json encoding, system prompt, and reasoning parser

Fix links in README to match root GGUF filenames

Rename GGUF files to include -RU- prefix for consistency

Update jan-model.json to Fwizzer-R1-3B-RU-v2

Update Modelfile to Fwizzer-R1-3B-RU-v2

Update model name and repo links to Fwizzer-R1-3B-RU-v2

Move Russian GGUFs to root for LM Studio compatibility and relocate mmproj

Fix chat template: remove <think> prefill for automatic UI reasoning curtain

Restore rich trilingual flagship Model Card with ru/, en/, and zh/ links

Copy files from models/fwizzer1/Fwizzer-R1-3B-ZH-v2

Copy files from models/fwizzer1/Fwizzer-R1-3B-ZH-v2

Copy files from models/fwizzer1/Fwizzer-R1-3B-ZH-v2

Copy files from models/fwizzer1/Fwizzer-R1-3B-ZH-v2

Copy files from models/fwizzer1/Fwizzer-R1-3B-ZH-v2

Copy files from models/fwizzer1/Fwizzer-R1-3B-ZH-v2

Copy files from models/fwizzer1/Fwizzer-R1-3B-ZH-v2

Copy files from models/fwizzer1/Fwizzer-R1-3B-ZH-v2

Copy files from models/fwizzer1/Fwizzer-R1-3B-ZH-v2

Copy files from models/fwizzer1/Fwizzer-R1-3B-ZH-v2

Copy files from models/fwizzer1/Fwizzer-R1-3B-ZH-v2

Copy files from models/fwizzer1/Fwizzer-R1-3B-ZH-v2

Copy files from models/fwizzer1/Fwizzer-R1-3B-ZH-v2

Copy files from models/fwizzer1/Fwizzer-R1-3B-ZH-v2

Copy files from models/fwizzer1/Fwizzer-R1-3B-ZH-v2

Delete intermediate training checkpoints folder

Remove root Ministral-3-3B-Instruct-2512.Q8_0.gguf

Place in folder: en/Fwizzer-R1-3B-EN-Max.gguf

Remove root Ministral-3-3B-Instruct-2512.Q5_K_M.gguf

Place in folder: en/Fwizzer-R1-3B-EN-Balanced.gguf

Remove root Ministral-3-3B-Instruct-2512.Q4_K_M.gguf

Place in folder: en/Fwizzer-R1-3B-EN-Speed.gguf

Add README

Export EN GGUFs

Export EN GGUFs

Export EN GGUFs

Export EN GGUFs

Update active EN LoRA adapter (step 2500/2500)

Auto-checkpoint EN step 2500/2500

Update active EN LoRA adapter (step 2400/2500)

Auto-checkpoint EN step 2400/2500

Update en/README.md with comprehensive English model card

Update comprehensive flagship Model Card (RU/Multilingual)

Update active EN LoRA adapter (step 2300/2500)

Auto-checkpoint EN step 2300/2500

Update active EN LoRA adapter (step 2200/2500)

Auto-checkpoint EN step 2200/2500

Add Modelfile directly inside ru/ directory

Add en/generation_config.json with optimal sampling defaults (temp=0.6, top_p=0.95, rep_penalty=1.15)