DuxVector - Preview images V4.9 Mizuki Yukikaze (Taimanin Yukikaze) (BWC) 1274 images + some info
infoAbout This Title
Taimanin clan ninja are trained to endure all kinds of...persuasion techniques~
Instant access and no watermark version available for $10 Basic Plus+ Tier
$5 Basic Member Tier access later today~
Download Link for $10 Basic Plus+ Members
bonus content - official signature outfit, taimanin bodysuit, elbow gloves, thighhighs
scene - nighttime dungeon
Set order slightly messed up. Will accommodate for that.
A small production error happened which affected the male skin-tone quite noticeably in images with heavier moonlight. A fix has been applied already, but due to the current sets running in a special test-mode (that's why there's a lot of images), the fix couldn't be applied in any reasonable timeframe today. My apolocheese."Test-mode" will continue for 2 more sets after this, after which pose recalibrations will be done and I'll internally finish training my little VLM helper tool. Image count after that should return in the 600-750 range.
The idea behind all of this is preparation for V5.
V5 will include a significant overhaul to the current experience by trading the more volume-oriented sets for smaller, text-driven narrative experiences with a thematic focus (the fabled themed sets). While I say smaller, sets will still pack a punch. Expectations are in the 300-500 image range with a 33/66 balance between text and non-text. (Disclaimer: Early estimates.)
The initial themed sets were intended for V4.7 if I remember correctly, but the V4.5 base just isn't good enough when it comes to text generation. Preliminary V5 teasers have demonstrated that text will be improved greatly, which will allow for proper automation. A lot of the groundwork has been laid already with the scaffolding for each narative experience currently underway. "Stories" will be the main update point of V5 content as visual fidelity has reached a satisfactory point. The initial estimate is that V5 will launch with around ~40 stories with more to come. As for what the framework for stories will look like and what level of dynamic "branching paths" is to be expected, that'll be covered once we get closer to V5.
All of this is a massive engineering and creative endeavor. But as of right now there's no roadblocks to the execution. The in-house VLM and pose recalibration is all to free up as much devtime as possible. Internal alpha version already works, with the beta version likely to finalize testing by the middle of August. After that it's all about wiring each story into production. The real question is when NAI drops V5... Hopefully they skip the Curated phase or keep it short. Fingers crossed
Taimanin clan ninja are trained to endure all kinds of...persuasion techniques~
Instant access and no watermark version available for $10 Basic Plus+ Tier
$5 Basic Member Tier access later today~
Download Link for $10 Basic Plus+ Members
bonus content - official signature outfit, taimanin bodysuit, elbow gloves, thighhighs
scene - nighttime dungeon
Set order slightly messed up. Will accommodate for that.
A small production error happened which affected the male skin-tone quite noticeably in images with heavier moonlight. A fix has been applied already, but due to the current sets running in a special test-mode (that's why there's a lot of images), the fix couldn't be applied in any reasonable timeframe today. My apolocheese."Test-mode" will continue for 2 more sets after this, after which pose recalibrations will be done and I'll internally finish training my little VLM helper tool. Image count after that should return in the 600-750 range.
The idea behind all of this is preparation for V5.
V5 will include a significant overhaul to the current experience by trading the more volume-oriented sets for smaller, text-driven narrative experiences with a thematic focus (the fabled themed sets). While I say smaller, sets will still pack a punch. Expectations are in the 300-500 image range with a 33/66 balance between text and non-text. (Disclaimer: Early estimates.)
The initial themed sets were intended for V4.7 if I remember correctly, but the V4.5 base just isn't good enough when it comes to text generation. Preliminary V5 teasers have demonstrated that text will be improved greatly, which will allow for proper automation. A lot of the groundwork has been laid already with the scaffolding for each narative experience currently underway. "Stories" will be the main update point of V5 content as visual fidelity has reached a satisfactory point. The initial estimate is that V5 will launch with around ~40 stories with more to come. As for what the framework for stories will look like and what level of dynamic "branching paths" is to be expected, that'll be covered once we get closer to V5.
All of this is a massive engineering and creative endeavor. But as of right now there's no roadblocks to the execution. The in-house VLM and pose recalibration is all to free up as much devtime as possible. Internal alpha version already works, with the beta version likely to finalize testing by the middle of August. After that it's all about wiring each story into production. The real question is when NAI drops V5... Hopefully they skip the Curated phase or keep it short. Fingers crossed
groupArtist
1k+
DuxVectorverified
photo_library20 albums
visibility51.8K views
View Profile arrow_forward
Comments (0)