I present my Illustrious character LoRA for Viago de Riva from Dragon Age: The Veilguard.<hr /><hr />V2 is my second version and attempt to see if I could refine the detail in generations byt altering my training settings. The generations do look a bit cleaner, nothing drastic but it's still a noticeable improvement.Recommended LoRA strength: 0.7 Also, I use an adetailer in my generations for this LoRA.V2 main tags are as follows:<pre><code>dav_viago, 1boy, mature male, black hair, short hair, dark skin, beard, mustache, blue eyes</code></pre>V2 outfit tags are as follows:<pre><code>leather armor, hooded cloak, single pauldron, leather gloves, leather belt, leather pants, leather boots</code></pre>The settings I used for my generation testing were 3-5 CFG, 20 steps, DPM++ 2M SDE - Karras on Forge with an adetailer (face_yolov8s).<hr /><hr />V1 is my initial version and produces decent results.Recommended LoRA strength: 0.6 Also, I use an adetailer in my generations for this LoRA.V1 main tags are as follows:<pre><code>dav_viago, 1boy, mature male, black hair, short hair, dark skin, beard, mustache, blue eyes</code></pre>V1 outfit tags are as follows:<pre><code>leather armor, hooded cloak, single pauldron, leather gloves, leather belt, leather pants, leather boots</code></pre>The settings I used for my generation testing were 3-5 CFG, 20 steps, DPM++ 2M SDE - Karras on Forge with an adetailer (face_yolov8s).

Dragon Age The Veilguard - Viago de Riva - Illustrious

V1 is my initial version and produces decent results. Recommended LoRA strength: 0.6 Also, I use an adetailer in my generations for this LoRA.

Viago.zip

Dragon Age Veilguard - Viago de Riva - Illustrious-000004.safetensors

I present another addition to my collection of Dragon Age men, Viago de Riva.<hr />V2 is my revised version of this LoRA since I felt his face looked a bit off and his hair messier than I wanted so I added more training images and tried to refine the scenery a bit more too in the training data. I also increased the dim/alpha on this version from 16/8 to 32/16 so it could learn more detail than V1 which means it is around 197 MB vs 99 MB in size.I do recommend an adetailer (or face fix) if the shots are more zoomed out.V2 main tags:<pre><code>dav_viago, 1boy, black hair, short hair, dark skin, beard, mustache, blue eyes</code></pre>V2 outfit tags:<pre><code>leather armor, hooded cloak, single pauldron, leather gloves, leather belt, leather pants, leather boots</code></pre>Note: the single pauldron on his right upper arm/shoulder doesn't like generating consistently. You'll see it on occasion but I got tired of fighting with it so it is what it is at this point The gloves also like to randomize on how they look.Recommended LoRA strength: 0.7 - 0.8All of my testing was done with Forge on 4-6 CFG with DPM++ 2M SDE - Karras at 20 Steps.<hr />V1 is my initial version of this LoRA. The generations were fairly consistent (aside from the armor being a bit random here and there) and especially looked great on close-ups and NSFW.I do recommend an adetailer (or face fix) if the shots are more zoomed out.V1 main tags:<pre><code>dav_viago, 1boy, black hair, short hair, dark skin, beard, mustache, blue eyes</code></pre>V1 outfit tags:<pre><code>leather armor, hooded cloak, single pauldron, leather gloves, leather belt, leather pants, leather boots</code></pre>Note: the single pauldron on his right upper arm/shoulder doesn't like generating consistently. You'll see it on occasion but I got tired of fighting with it so it is what it is at this point. The gloves also like to randomize on how they look.Recommended LoRA strength: 0.5 - 0.7All of my testing was done with Forge on 3-5 CFG with DPM++ 2M SDE - Karras at 20 Steps.<hr />optional tag (for location-specific scenery):<pre><code>dav_treviso</code></pre>Note: I tried training scenery shots of Treviso into the LoRA so this tag makes the scenery look closer to Treviso.

Dragon Age The Veilguard - Viago de Riva - Pony

V2 is my revised version of this LoRA since I felt his face looked a bit off and his hair messier than I wanted so I added more training images and tried to refine the scenery a bit more too in the training data. I also increased the dim/alpha on this version from 16/8 to 32/16 so it could learn more detail than V1 which means it is around 197 MB vs 99 MB in size.I do recommend an adetailer (or face fix) if the shots are more zoomed out.

Files with SHA256: f4318047fa0ec482e8f49c40c5ed8adbc97d1fff718211ab57198b4e241c2d5c