neurojenml-inference — Merged Model for Inference
This is the base model with the neurojenml LoRA adapter merged in. Use this for inference — the HF Router can serve this directly.
Base: google/medgemma-4b-it
Adapter: sciedoc/neurojenml
Train loss: 0.2150
Examples: 763
Local Usage (Ollama)
To use with Ollama, convert to GGUF locally:
# Clone the merged model
huggingface-cli download sciedoc/neurojenml-inference --local-dir ./neurojen-merged
# Create Ollama model (if you have Modelfile.ollama)
ollama create neurojen -f Modelfile.ollama