Rank 128 LoRA extraction of <a target="_blank" rel="ugc" href="https://civarchive.com/models/2600510/juggernaut-z">https://civarchive.com/models/2600510/juggernaut-z</a> againt <a target="_blank" rel="ugc" href="https://civarchive.com/models/2342797/z-image-base">Z-Image-Base</a> in order to be able to use it with <a target="_blank" rel="ugc" href="https://civarchive.com/models/2168935/z-image-turbo">Z-Image-Turbo</a>. Mainly done because <a target="_blank" rel="ugc" href="https://github.com/leejet/stable-diffusion.cpp/issues/1488">my setup (sd.cpp) don't work with quantized version of ZiB</a> and derivatives 😭All showcases pictures are using Euler+Smoothstep (~Beta), 8 steps, CFG 1, Flow shift 3, with a Q4_K quantized version of both ZiT and Qwen3-4B and were made on a GTX 1060 (6GB VRAM) using Stable-diffusion.cpp.<hr />Full ownership of this LoRA stays with <a target="_blank" rel="ugc" href="https://civarchive.com/user/KandooAI">https://civarchive.com/user/KandooAI</a>, merge are forbidden, this is for your personal use with ZiT on-site or locally only.<hr />The extraction was done using this script adapted to be able to handle ANY model as long as they have the same keys (which means i had to re-key Juggernaut Z to unfuse the qkv layers).<pre><code>import sys, re

# Very basic args check
if len(sys.argv) &lt; 4:
 print(f"Usage: {sys.argv[0]} origin.safetensors tuned.safetensors lora.safetensors")
 exit(1)

# Importing the libs
from safetensors.torch import load_file, save_file
import torch
from tqdm import tqdm

# No gradient needed
torch.set_grad_enabled(False)


# The most BASIC svd extraction of lora weight
def extract_lora_weight(m, r):
 u, s, vh = torch.linalg.svd(m.cuda())
 B = u[:, :r] @ torch.diag(s[:r])
 A = vh[:r, :]
 return (B.cpu().contiguous(), A.cpu().contiguous())


# Load models
print(f"Loading {sys.argv[1]}...")
orig = load_file(sys.argv[1])
print(f"Loading {sys.argv[2]}...")
tune = load_file(sys.argv[2])

wk = list(set(orig) &amp; set(tune))

# Get ready
lora = {}
rank = 128
count = 0

# Extract LoRA
print("Extracting...")
for k in tqdm(wk):
 diff = tune[k].to(dtype=torch.float) - orig[k].to(dtype=torch.float)
 if torch.mean(torch.abs(diff)) &lt; 0.0001 or len(diff.shape) not in [1,2]:
 # Either not the right shape or not enough difference, let's skip it
 del diff
 continue
 if len(diff.shape) == 1:
 count +=1
 if k.endswith(".bias"):
 lora[k.replace(".bias", ".diff_b")] = diff.to(dtype=torch.bfloat16)
 else:
 lora[k.replace(".weight", ".diff")] = diff.to(dtype=torch.bfloat16)
 del diff
 else:
 if diff.shape[0] &lt;= rank or diff.shape[1] &lt;= rank:
 # Too small, let's skip it for now
 del diff
 continue
 count +=1
 B, A = extract_lora_weight(diff, rank)
 del diff
 base_key = k.replace(".weight", "")
 lora[f"{base_key}.lora_B.weight"] = B.to(dtype=torch.bfloat16)
 lora[f"{base_key}.lora_A.weight"] = A.to(dtype=torch.bfloat16)

# Save LoRA
print(f"Stored {count}/{len(wk)} keys to LoRA.")
print(f"Saving {sys.argv[3]} ...")
save_file(lora, sys.argv[3])
print("Done!")
exit(0)</code></pre>

Juggernaut Z for ZiT

Use and upload of this model are following the source model authorization model. Merge or commercial use of this LoRA is forbidden, for commercial use of Juggernaut Z please contact original author.

Files with SHA256: 70c1ade091e3bc36a6bead405b6b9556d12ef64a1cfbe5c20d086b8b4a12afe9