THE IMAGE ABOVE IS NOT MADE WITH THE MODEL :p read and you know why

Alpha in-training model need you guys help testing, this model create walking/chilling in the city street result, mostly contain pretty girls going somewhere at night often wearing night club style clothing, and pretty girls in the streets in summer walking (no ?????? although some outfit are borderline scandalous) it also contain men, just in lesser quantity, no NSFW material was used in the training (again some outfit are borderline but not that much)

Model was trained with two lighting condition only, At night in a well lit city (mostly giving that yellowish tone due to city lamp post) and during day in sunny day in LA, no rain, no wind was trained into it (but hunyuan could probably pull it)

I need you guys to test by doing this prompt : str3etLIF3 video scene featuring a group of young people walking on the sidewalk in front of a night club, outdoor at night in a well lit city of Los Angeles, two women and two men are the main focus, they are walking toward the camera at center of the frame, one of the woman is a beautiful blonde chewing gum, one man is holding his cellphone as he talk to someone else, cars can be seen passing on the road left side of the frame, dynamic, 4k ultra quality, color graded mastery.

then if possible just start another test by starting with : str3etLIF3

and then go creative and tell me what you think it's worth (if it's worth anything, meaning if it's worth training or straight garbage material, which it might be...)

all of the training material used in this video are first person of someone either walking, or standing in the street of a city, so prompt accordingly, but go creative and maybe try 3rd person stuff too, Hunyuan is quite surprising at what it can pull sometimes (and sometimes less... hence the need to test if it's working) most of the training video used consisted of people walking on the sidewalk, so it should be it's strong point (I hope) some clips consisted of peoples just standing or sitting in the street, waiting and chilling, some other clip consisted of people walking past and then the viewer panning left or right to track a particular person (to keep him/her in frame) prompt along those line first to see if the model have ANY capacity, and if it does, try be more creative and see if it can pull something less mundane maybe

What I don't want to see (and what I fear I would see) is intersecting people, fusing in the middle of the walk as they walk into each other, disapearing limbs galore, appearing limbs galore, model confusing people going forward and or backward, and model confusion people facing the camera and suddenly turning their back to it that kind of odd stuff, what I want to see is well, just normal walking animation (no fancy move, nothing of that in this one) so depending of what your test bring, I will know if it's worth continuing or not. I appreciate anyone taking the time to help me by testing, and if no one decide to help well not the end of the world, but I might stop at epoch 20 anyway because this thing takes too long

it is only epoch 10 (out of 40) and my training time is expected to last 13 total hours (more or less) I think due to the nature of this lora, even epoch 10 can already give a decent feel about whether it's worth continuing training, or maybe not (or maybe if stopping at epoch 20 might be enough given the results)

I can't test it right now it would crash the training, it's a quick test