Laxhar/noobai-XL-1.0
Full fine-tune vs Laxhar/noobai-XL-0.77. Fine-tune: 100% of the weights changed, across 11 of 12 layers.
Tensors changed: Embeddings 571/590, other 662/662, Attention 716/716, Norms 546/546
10 derivatives checked
- Laxhar/noobai-XL-1.1: Partial fine-tune
- John6666/one-obsession-17-red-sdxl: Different architecture
- Anzhc/Noobai11-CLIP-L-and-BigG-Anime-Text-Encoders: Different architecture
- snuperh/AllInOne-XL-V12.5: Full fine-tune
- John6666/mol-keun-mix-v3noobaixl-sdxl: Different architecture
- Laxhar/noobai-XL-Vpred-0.5: Partial fine-tune
- John6666/perle-mix-wai-jru-pam-perlev10-sdxl: Different architecture
- John6666/mol-keun-mix-kemomolnoobaixl-sdxl: Different architecture
- John6666/one-obsession-v16-noobai-sdxl: Different architecture
- John6666/anillustrious-v5vae-sdxl: Different architecture
3 other repos hold byte-for-byte identical weight files.