Ollama drops LoRA adapter support, users must merge or use llama.cpp

Ollama has deprecated support for LoRA adapters starting with version 0.34.1, released in September 2026. This change means that Modelfiles containing an ADAPTER line will now result in an error, and Ollama no longer handles safetensors conversion to GGUF or GGUF quantization. Users now have two primary methods to utilize LoRA fine-tunes: either merge the adapter directly into the base model and import the resulting single GGUF file into Ollama, or manage the adapter separately and serve it using llama.cpp's llama-server. Testing indicates that merging the adapter into the model can lead to approximately 13% faster generation times on CPU. AI

IMPACT Users must adapt workflows for local model fine-tuning and deployment with Ollama.

RANK_REASON Change in functionality for an existing tool, not a new model release or research.

Read on dev.to — LLM tag →

• adapter_model.safetensors
• GGUF
• llama.cpp
• LoRA
• Modelfile
• Ollama
• peft
• PR #14969
• Qwen3 14B
• safetensors

• model release
• product

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →