I present my Illustrious character LoRA for Genn Greymane from World of Warcraft.<hr /><hr />V2 is my second version and attempt to increase detail consistency and quality. The facial hair is more accurate (now where the beard isn't connected to the sideburns) with moderate generation consistency. I also got better quality training images (using his most recent character model) for the outfit details and cut down on some of the training tags for eye/hair color, body hair, and such.Recommended LoRA strength: 0.7 Also, I use an adetailer in my generations for this LoRA.V2 main tags are as follows:<pre><code>genng, 1boy, mature, short hair, mustache, goatee beard, sideburns</code></pre>V2 outfit tags are as follows:<pre><code>brown overcoat, white two-tone tunic, white shirt, brown gloves, brown pants, brown boots</code></pre>The settings I used for my generation testing were 3-4 CFG, 20 steps, DPM++ 2M SDE - Karras on Forge with an adetailer (face_yolov8s).<hr /><hr />V1 is my initial version and produces very nice results. Recommended LoRA strength: 0.7 Also, I use an adetailer in my generations for this LoRA.V1 main tags are as follows:<pre><code>genn_g, 1boy, mature male, grey hair, very short hair, mustache, beard, blue eyes</code></pre>V1 outfit tags are as follows:<pre><code>brown overcoat, white jacket, white shirt, brown gloves, brown belt, brown pants, brown 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).

World of Warcraft - Genn Greymane - Illustrious

V1 is my initial version and produces very nice results.Recommended LoRA strength: 0.7 Also, I use an adetailer in my generations for this LoRA.

Genn Greymane 3.zip

World of Warcraft - Genn Greymane - Illustrious-000005.safetensors

This is my modest attempt at a Genn Greymane (that isn't his worgen form) LoRA for Pony.<hr />V3 is my third version and an attempt to clean up my training tags and images and improve the quality generation results. I got consistent results in my testing but I do recommend using an adetailer to clean up the face on more zoomed out shots.Recommended LoRA strength: 0.7-0.8V3 main tags I used are as follows:<pre><code>genn_g, mature male, grey hair, blue eyes, mustache, beard</code></pre>V3 outfit is tagged as such:<pre><code>brown overcoat, white shirt, white jacket, brown belt, brown gloves, brown pants, brown boots</code></pre>All of my V3 testing was done on Forge at 4-6 CFG on Euler A-Automatic with 20 steps.<hr />V2 (clip skip 1) is my second attempt (after learning a LOT more about training tags) so I did a better job at identifying indoors/outdoors (on the training images) as well as labeling specific parts of the armor and such. The faces on V2 generate fairly well but I recommend using an adetailer to clean it up.I changed to clip skip 1 because I've been finding a lot of my realistic LoRAs don't look very clean on clip skip 2.V2 main tags I used are as follows:<pre><code>genn_g, mature male, grey hair, blue eyes, mustache, beard</code></pre>V2 outfit is tagged as such:<pre><code>long coat, white shirt, brown belt, brown gloves, brown pants, brown boots</code></pre>All of my V2 testing was done on Forge at 5-7 CFG on DPM++ 2M SDE-Karras with 30 steps.<hr />V1 was my first attempt at making this LoRA. The likeness is pretty spot-on but the SFW prompt pose changes may be a bit finnicky at times unless you get really specific with your prompts. I had no issues posing him for NSFW, the prompt changes were very responsive. I recommend keeping the LoRA strength between 0.7 and 0.8 and this was trained on clip skip 2. I also need to note that I was using an ADetailer for all of my test generations.V1 main tags I used are as follows:<pre><code>genn_g, male, grey hair, blue eyes, mustache, beard</code></pre><ul><li>optionally, can add "short hair" if the generations have long hair in the back and you don't want that</li></ul>V1 outfit is tagged as such:<pre><code>coat, white shirt, brown belt, brown gloves, brown pants, brown boots</code></pre>All of my V1 testing was done on Forge at 6-8 CFG on DPM++ 2M SDE-Karras with 30 steps.

World of Warcraft - Genn Greymane - Pony

V3 is my third attempt to clean up my training tags and images and improve the quality generation results. I got consistent results in my testing but I do recommend using an adetailer to clean up the face on more zoomed out shots.Recommended LoRA strength: 0.6-0.7

Files with SHA256: 2342b32ea0628df2b6022463e2664ad66396bde3bbaa7de3e11656286375413d