Implementation:Ollama Ollama ParseFromModel

Knowledge Sources	Ollama
Domains	Storage, Model_Management
Last Updated	2026-02-14 00:00 GMT

Overview

Concrete tool for resolving a base model's layers from its manifest provided by the server package.

Description

parseFromModel takes a model name, finds its local manifest, and returns its layers as a slice of layerGGML structs. Each layer contains the GGUF-parsed metadata and a reference to the blob file.

The companion GetModel function provides a higher-level interface, returning a fully populated Model struct with paths, template, system prompt, options, and capability detection.

Usage

Called by CreateHandler to load the base model's layers when processing a FROM directive in model creation.

Code Reference

Source Location

• Repository: ollama
• File: server/model.go (parseFromModel), server/images.go (GetModel)
• Lines: model.go:L28-77 (parseFromModel), images.go:L271-369 (GetModel)

Signature

func parseFromModel(ctx context.Context, name model.Name, fn func(api.ProgressResponse)) (layers []*layerGGML, err error)

func GetModel(name string) (*Model, error)

Import

import "github.com/ollama/ollama/server"

I/O Contract

Inputs

Name	Type	Required	Description
ctx	context.Context	Yes	Cancellation context
name	model.Name	Yes	Parsed model name (host/namespace/model:tag)
fn	func(api.ProgressResponse)	Yes	Progress callback

Outputs

Name	Type	Description
layers	[]*layerGGML	Slice of GGUF layers from the base model (weights, config, template, system, params)
error	error	Non-nil if model not found or GGUF parsing fails

Usage Examples

Internal Usage in CreateHandler

// From server/create.go
layers, err := parseFromModel(ctx, baseName, fn)
if err != nil {
    return err
}
// layers now contains the base model's weights and metadata
// Override with Modelfile directives...

Related Pages

Implements Principle

• Principle:Ollama_Ollama_Base_Model_Resolution

Requires Environment

• Environment:Ollama_Ollama_Go_Runtime