Models, artifacts, and runtimes · Lesson 15 of 16 · 6 min

Build a model baseline

A serving baseline includes model, tokenizer, runtime, quantization and workload. Record these together before tuning so changes can be attributed rather than guessed.

Open the lab Course map

Mastery 0/3Run itExplain itApply it

learner@clairos:~ $

↑ history · Tab complete · Ctrl+C cancel · Esc, Tab to leaveExit 0

Your objectives

Hints are always available. Wrong turns never cost XP.

1
Read the case evidence
Read /opt/ai-lab/cases/aiops-models-runtimes-015.json. Separate the recorded facts from your assumptions.
Hint & worked example
A serving baseline includes model, tokenizer, runtime, quantization and workload. Record these together before tuning so changes can be attributed rather than guessed.
The complete case JSON is read successfully.
```
cat /opt/ai-lab/cases/aiops-models-runtimes-015.json
```
Common mistake to avoid: Keep only the model display name as the baseline.
2
Isolate the deciding signal
Select the workload field from the case JSON. What does it establish—and what does it not establish?
Hint & worked example
Capture the complete model/runtime/workload tuple before changing a parameter.
A successful JSON query isolates workload.
```
jq .workload /opt/ai-lab/cases/aiops-models-runtimes-015.json
```
Common mistake to avoid: Using a different case or an unrelated signal as the justification.
3
Record your operational decision
Choose the justified code: capture-model-baseline, change-without-evidence, or ignore-the-signal. Save that one code followed by a newline in /home/learner/lab/aiops-models-runtimes-015-decision.txt.
Hint & worked example
Capture the complete model/runtime/workload tuple before changing a parameter.
The decision artifact contains the correct code, not just a matching command in history.
```
printf '%s\n' 'capture-model-baseline' > /home/learner/lab/aiops-models-runtimes-015-decision.txt
```
Common mistake to avoid: Keep only the model display name as the baseline.

Understand the pattern

A serving baseline includes model, tokenizer, runtime, quantization and workload. Record these together before tuning so changes can be attributed rather than guessed.

Interpret the evidence

Capture the complete model/runtime/workload tuple before changing a parameter.

A tempting mistake

Keep only the model display name as the baseline.

All names, measurements, prices, versions, and thresholds in this exercise are fictional training data. The evidence supports this bounded scenario, not a claim about a live provider.

Worked example and expected result

$ cat /opt/ai-lab/cases/aiops-models-runtimes-015.json

{
  "fixture": "Fictional training data; not a live measurement.",
  "model": "m3",
  "tokenizer": "tok-v2",
  "runtime": "r2",
  "quantization": "q8",
  "workload": "gold-v4"
}

Exit status: 0. The command completed successfully in the training fixture.

$ jq .workload /opt/ai-lab/cases/aiops-models-runtimes-015.json

"gold-v4"

Exit status: 0. The command completed successfully in the training fixture.

$ printf '%s\n' 'capture-model-baseline' > /home/learner/lab/aiops-models-runtimes-015-decision.txt

(no output; exit 0)

Exit status: 0. The command completed successfully in the training fixture.

Try a variation

Name the assumption that would have to change for “capture-model-baseline” to stop being the justified action. Do not silently replace missing evidence with a guess.

Explain it

What does the evidence in “Build a model baseline” justify?Keep only the model display name as the baseline.Make the change now and gather the missing evidence afterward.Capture the complete model/runtime/workload tuple before changing a parameter.

Apply it elsewhere

A colleague concludes: “Keep only the model display name as the baseline.” What should you do?Treat this fixture as a live measurement.Reject that conclusion; return to the case evidence and its stated limits.Accept it because the training command completed.

Keep the momentum.

Complete the lab, then explain and apply the pattern.

Next: Models and runtimes checkpoint Try an incident drill

Anonymous lesson and mastery progress lasts for this session. Sign in to keep synced lesson XP between sessions; mastery stays device-local. No speed bonus, lost lives, or hint penalties.

Keep learning

Case methodology: evidence, limits, decision