Fenon engineering / draft template
Measure the work. Not the demo.
Use this template for a benchmark or evaluation article. No measured performance is asserted here.
Define the question.
Write the hypothesis and describe a representative workload. Name the hardware, model revision, input sizes and deployment conditions.
Plot the distribution.
Use labelled axes, real units and a clear legend. Show the appropriate distribution instead of a single unexplained number. Always cite the data source.
Explain the trade-off.
Describe what improved, what did not, and where the result may not generalize. Keep limitations next to the conclusion.