Hi
So my advice is start with short sessions first like 30 seconds then 1 minute and so on.I do recommend the class mode but I feel to get comfortable start with 30 seconds sessions on your own. So start with the line of action which is the general line that forms the gesture. Do this until you're happy with the general gesture you see before adding details. When you start adding details don't focus too much on details and get the most prominent part of the gesture down first.
You have a very good base from what I can see but it's also just your proportions. I think the best way forward though may be to go back to basics
Try not to focus on one part too much in the beginning. Work on everything at the same time.
I hope this helps. Good luck