Some points about your models:
1. On the Tunica Manicata, the sleeve stripes were never at the end of the sleeve. They were always at least about 1 centimeter from the edge. You couldn't do the decoration right if they were on the edge of the sleeve anyways, due to the way the tunics were made.
2. The Spathae you have made, which come from Attila which are based off of Deepeeka's AH2002N "Late Late Roman Spatha", are bad representations of Late 2nd to Early 4th century Lauriacum-Hromovka type blades. Notably, the one Deepeeka's is based off of is a late 3rd century or early 4th century example found at Cologne.
More appropriate blades for this time period would be the Feltwell Blade, or the Pouan Blade. Also the Trier Blade.
Trier, c.a. 380-400 AD:
The Deepeeka Feltwell, which I helped develop, c.a. 400-425 AD:
**The blade is Hexagonal, not fullered.
Pouan and belt fittings, with neck Torque, ring, and wrist Armilla, Gothic or Burgundian, c.a. 470 AD:
Reconstruction here:
http://en.ordalies.com/replicas.php
This man here, the mid-5th century Kemathen Warrior, has a Germanic Type-I Spatha:
Altlussheim Spatha, mid 5th-early 6th century. Hunnic. I can provide further images of a historically accurate reconstruction (not including the gold scabbard foil pieces) but will take some time to upload.
3. Your Plumbata is based off the Deepeeka one, which although improved is still rather overweight. The Strategikon's suggestion that they were carried in quivers is more practical, although both reconstructions are valid.
4. The Spiculum and Verrutum both had Barbed points on short iron shafts. The Lancea was an alternate name for the Verrutum, the spear itself was called a Contus and was about 8 foot long for infantry, 12 foot for cavalry.
Verruta with leaf-bladed points. Leaf-bladed and trilobate points are both validated by the archaeological record, the latter particularly for the Huns.
A Spiculum
Spicula heads, based on the find from Arbeia.
A Spiculum head from Nydam.
5. Your oval shield should be edged with rawhide or leather, stitched on. They were typically dished, although some flat examples are supported by shield boss shapes. By the time of Attila, pointed shield bosses were more common than domed ones. Ovoid shields were also very wide, oftentimes almost circular. All oval late Roman shields measured about 39-42 inches in length and 34-38 in width, based on the seven or so examples from Dura Europos. Large circular shields were also in use. Smaller, lenticular circular shields existed too of course, primarily for cavalry and standard-bearer usage.
Otherwise, keep up the good work.
I can provide tons more about the timeframe from 395-455 in terms of military equipment. It's the little things, like belt fittings, scabbard fittings, and boot styles, that can make the difference between it being an early 4th century and late 4th century impression.
Also, sextapartite helmets fell out of use in the 5th century (think the Berkasova type). However, Bipartite helmets continued in use, particularly ones with earguards and Berkasova-style cheekpieces, as well as more classic Intercisa-style cheekpieces (sometimes with earguards as well). See the Koblenz helmet, for example.
However, the Concesti helmet (425 AD, Romano-Hunnic) is an exception, being of sextapartite bowl construction. It is notably lacking a neckguard, suggesting it may have had a maille coif instead.
EDIT: I reenact the Roman army of 425-454 AD, by the way.
https://www.facebook.com/Placidi-Val...7288724006375/