Not sure what "raw numbers" you are looking at but at the referenced CDC site, there are downloadable spreadsheets with the numbers that underly the chart. There are two values for each combination of parameters (date, age group, cause of death, etc),"predicted" (modeled) and "observed" (raw numbers).
Comparison of total observed deaths per week through September:
261K deaths total have already been recorded. Given the time lag in tabulating the data, we are probably already past that as of now. Using the average number of deaths per week from the past two years to extrapolate forward yields 314K deaths, so the end result will likely be much higher than that.
To an earlier point, here are the number of deaths from heart disease per week through September. It's actually higher this year than previous years:
