Download and inspect one local model
Download and inspect one local model — Local AI with Ollama. A tag is part of the experiment; Pull then show; Separate disk from runtime memory. Practical assignment: Lab evidence.
Demo edition. The materials are undergoing editorial review.
Topics covered
- A tag is part of the experiment
- Pull then show
- Separate disk from runtime memory
A tag is part of the experiment
The example uses llama3.2:1b as an explicit small-model tag, not a quality recommendation. Confirm that it remains available and fits your machine. If unavailable, choose a documented local replacement and record it everywhere the example names the model.
Keep a small lab notebook containing the operating system, Ollama version, exact model tag, download size and task. A model tag identifies a particular available artifact, not a promise about quality or a universal memory requirement. Start with harmless invented text. Read the model license before redistributing weights or using them commercially. Keep installation and inference costs separate.
Pull then show
Download once, list the installed models, then inspect the selected entry. Read the license and parameters shown by the current tool. A successful download only proves that the artifact arrived; a generation request is still needed to test execution.
The first generation loads weights into working memory; listing installed models reads an inventory on disk. If listing succeeds but generation fails, preserve the error and inspect loaded allocation. This explains why repeating the download is usually the wrong first response to memory pressure.
Run the example only after the previous check succeeds. The commands below target the local service; they do not authorize publishing a network port. Preserve the response and exit status before editing parameters. Change one input at a time so a later improvement can be attributed to an actual change. An empty response, HTTP error and slow generation require different diagnoses.
ollama pull llama3.2:1b
ollama list
ollama show llama3.2:1b
Separate disk from runtime memory
Run one short prompt and inspect loaded models in a second terminal. Download size and runtime memory are different quantities. If memory pressure rises, stop this model before trying another; deleting unrelated files will not fix an oversized active context.
Measure a cold request separately from a second request with the model already loaded. Record whether the answer is correct against the supplied fictional facts, rather than judging only fluency. Keep private files out of prompts until the endpoint and logging behavior are verified. Local inference reduces one data transfer path, but downloads, integrations, backups and browser extensions still deserve inspection.
ollama run llama3.2:1b "Reply with one short greeting."
ollama ps
ollama stop llama3.2:1b
Practical assignment
Lab evidence
Continue to learning