Yes, I am currently using BodyMedia to measure my progress at the gym. My workout routine consists of cardio, plyometrics and weights lifting (all done separately), and biking. The BM captures caloric expenditure of all these types of workout. The interactive BM dashboard even lets you isolate specific time periods of the day, so you can gauge the efficiency of each workout (e.g., I do cardio during lunch, and plyometrics after work. Not sure if you can isolate the effects of individual exercises (unless you do 3 sets in a row, and time each exercise so you are able to isolate it on the dashboard later?). I am also trying to figure out if BM captures the “afterburn” effect, but I will have to design experiment for that. Not sure about the accuracy of the estimates, though. The published accuracy is 90%, so I always assume that BM underestimates and round the caloric expenditure down 10%. But I did compare my heart rate across different routines and BM estimates seem to correlate with it: it reports higher caloric loss estimates for those workouts when my heart rate is higher.
I don’t find my BMF captures a lot of reps during weight training (compared to energy expenditure estimates from my suunto memory belt or my body measurements taken by hand over the last 2 decades ).
I have not had long enough with my BMF to get a big enough dataset to separately analyse the various raw sensor data against different workouts, but intuitively I would expect under-reporting from the accelerometer on those exercises with little arm movement eg squats. That said, IME the heat flux data seems to be damn good and responds quite well to things such as energy loss from fever, so it may well be that the heat from the large compound movements such as squats is accounted for by weighting given to the heat flux data.