
I cannot find a crosshatch style LoRa for Flux, so I created one ?.
You will probably need to use the word "drawing" or "sketch" for the LoRA to take effect, and if you want black and white, you should use "black and white drawing/sketch". The word "crosshatch" is probably not required, but use it to indicate your intentions clearly.
You can add color descriptions to specific items in the prompt, and it should be rendered in color.
The crosshatch effect is strong even at lower weight (try 0.7 to 1.0), so you can tone it down for better prompt following or to reduce distortion.
Have fun, and if you post your first image directly here (click on "Add Post" above the gallery), I will give you 10 buzz for that image ? (If I forgot, just complain loudly ?)
描述:
Crosshatch V1 NoCD6A3Cos6 2024-12-16 00:46:49
No caption training, but manually added these 3 captions:
"two cherries", "Space marine from Warhammer 40K", and "An African baby with her left hand on her ear"
This is epoch 10, the rest of the epoch can be found here: tensor. art/models/807651957285228923/Crosshatch-V1-NoCD6A3Cos6-2024-12-16-00:46:49-Ep-10
FLUX.1 - dev-fp8
Trigger: crosshatch drawing
Repeat: 20 Epoch: 10 20 512x512 images 4000 steps total
Unet LR: 0.0006 Scheduler: cosine Optimizer: AdamW
Network Dim: 6 Alpha: 3
训练词语: crosshatch drawing
名称: crosshatch1_nocap_d6a3.safetensors
大小 (KB): 56201
类型: Model
Pickle 扫描结果: Success
Pickle 扫描信息: No Pickle imports
病毒扫描结果: Success