Gemma4-E2B-it-Deepseek-V4-8000x : GGUF

• Epochs: 2
• Method: QLoRA
• Context length: 1024
• Learning Rate: 0.0002

• Rank: 16
• Alpha: 16
• Dropout: 0.00
• Target modules: All
• LoRA

• Optimizer: Paged AdamW 8-Bit
• LR scheduler: Linear
• Batch Size: 1
• Grad Accum: 32
• Weight Decay: 0.001

• For text only LLMs: llama-cli -hf Alienstro/Gemma4-E2B-it-Deepseek-V4-8000x --jinja
• For multimodal models: llama-mtmd-cli -hf Alienstro/Gemma4-E2B-it-Deepseek-V4-8000x --jinja

Available Model files:

• gemma-4-e2b-it.Q4_K_M.gguf
• gemma-4-e2b-it.BF16-mmproj.gguf

⚠️ Ollama Note for Vision Models

1. Place the Modelfile in the same directory as the finetuned bf16 merged model
2. Run: ollama create model_name -f ./Modelfile (Replace model_name with your desired name)

This will create a unified bf16 model that Ollama can use. This was trained 2x faster with Unsloth