File size: 46,325 Bytes
db16e05 50b030f db16e05 50b030f db16e05 50b030f db16e05 50b030f db16e05 50b030f db16e05 50b030f db16e05 50b030f db16e05 50b030f db16e05 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 |
import gradio as gr
import torch
import os
import gc
import shutil
import requests
import json
import struct
import numpy as np
import re
import yaml
import subprocess
from pathlib import Path
from typing import Dict, Any, Optional, List
from huggingface_hub import HfApi, hf_hub_download, list_repo_files, login
from safetensors.torch import load_file, save_file
from tqdm import tqdm
# --- Memory Efficient Safetensors ---
class MemoryEfficientSafeOpen:
def __init__(self, filename):
self.filename = filename
self.file = open(filename, "rb")
self.header, self.header_size = self._read_header()
def __enter__(self):
return self
def __exit__(self, exc_type, exc_val, exc_tb):
self.file.close()
def keys(self) -> list[str]:
return [k for k in self.header.keys() if k != "__metadata__"]
def metadata(self) -> Dict[str, str]:
return self.header.get("__metadata__", {})
def get_tensor(self, key):
if key not in self.header:
raise KeyError(f"Tensor '{key}' not found in the file")
metadata = self.header[key]
offset_start, offset_end = metadata["data_offsets"]
self.file.seek(self.header_size + 8 + offset_start)
tensor_bytes = self.file.read(offset_end - offset_start)
return self._deserialize_tensor(tensor_bytes, metadata)
def _read_header(self):
header_size = struct.unpack("<Q", self.file.read(8))[0]
header_json = self.file.read(header_size).decode("utf-8")
return json.loads(header_json), header_size
def _deserialize_tensor(self, tensor_bytes, metadata):
dtype_map = {
"F32": torch.float32, "F16": torch.float16, "BF16": torch.bfloat16,
"I64": torch.int64, "I32": torch.int32, "I16": torch.int16, "I8": torch.int8,
"U8": torch.uint8, "BOOL": torch.bool
}
dtype = dtype_map[metadata["dtype"]]
shape = metadata["shape"]
return torch.frombuffer(tensor_bytes, dtype=torch.uint8).view(dtype).reshape(shape)
# --- Constants & Setup ---
try:
TempDir = Path("/tmp/temp_tool")
os.makedirs(TempDir, exist_ok=True)
except:
TempDir = Path("./temp_tool")
os.makedirs(TempDir, exist_ok=True)
api = HfApi()
def cleanup_temp():
if TempDir.exists():
shutil.rmtree(TempDir)
os.makedirs(TempDir, exist_ok=True)
gc.collect()
def get_key_stem(key):
key = key.replace(".weight", "").replace(".bias", "")
key = key.replace(".lora_down", "").replace(".lora_up", "")
key = key.replace(".lora_A", "").replace(".lora_B", "")
key = key.replace(".alpha", "")
prefixes = [
"model.diffusion_model.", "diffusion_model.", "model.",
"transformer.", "text_encoder.", "lora_unet_", "lora_te_", "base_model.model."
]
changed = True
while changed:
changed = False
for p in prefixes:
if key.startswith(p):
key = key[len(p):]
changed = True
return key
# =================================================================================
# TAB 1: MERGE & RESHARD (Legacy Logic)
# =================================================================================
def parse_hf_url(url):
if "huggingface.co" in url and "resolve" in url:
try:
parts = url.split("huggingface.co/")[-1].split("/")
repo_id = f"{parts[0]}/{parts[1]}"
filename = "/".join(parts[4:]).split("?")[0]
return repo_id, filename
except:
return None, None
return None, None
def download_lora_smart(input_str, token):
local_path = TempDir / "adapter.safetensors"
if local_path.exists(): os.remove(local_path)
repo_id, filename = parse_hf_url(input_str)
if repo_id and filename:
try:
hf_hub_download(repo_id=repo_id, filename=filename, token=token, local_dir=TempDir)
found = list(TempDir.rglob(filename.split("/")[-1]))[0]
if found != local_path: shutil.move(found, local_path)
return local_path
except: pass
try:
if ".safetensors" in input_str and input_str.count("/") >= 2:
parts = input_str.split("/")
repo_id = f"{parts[0]}/{parts[1]}"
filename = "/".join(parts[2:])
hf_hub_download(repo_id=repo_id, filename=filename, token=token, local_dir=TempDir)
found = list(TempDir.rglob(filename.split("/")[-1]))[0]
if found != local_path: shutil.move(found, local_path)
return local_path
candidates = ["adapter_model.safetensors", "model.safetensors"]
files = list_repo_files(repo_id=input_str, token=token)
target = next((f for f in files if f in candidates), None)
if not target:
safes = [f for f in files if f.endswith(".safetensors")]
if safes: target = safes[0]
if not target: raise ValueError("No safetensors found")
hf_hub_download(repo_id=input_str, filename=target, token=token, local_dir=TempDir)
found = list(TempDir.rglob(target.split("/")[-1]))[0]
if found != local_path: shutil.move(found, local_path)
return local_path
except Exception as e:
if input_str.startswith("http"):
try:
headers = {"Authorization": f"Bearer {token}"} if token else {}
r = requests.get(input_str, stream=True, headers=headers, timeout=60)
r.raise_for_status()
with open(local_path, 'wb') as f:
for chunk in r.iter_content(chunk_size=8192): f.write(chunk)
return local_path
except: pass
raise e
def load_lora_to_memory(lora_path, precision_dtype=torch.bfloat16):
state_dict = load_file(lora_path, device="cpu")
pairs = {}
alphas = {}
for k, v in state_dict.items():
stem = get_key_stem(k)
if "alpha" in k:
alphas[stem] = v.item() if isinstance(v, torch.Tensor) else v
else:
if stem not in pairs: pairs[stem] = {}
if "lora_down" in k or "lora_A" in k:
pairs[stem]["down"] = v.to(dtype=precision_dtype)
pairs[stem]["rank"] = v.shape[0]
elif "lora_up" in k or "lora_B" in k:
pairs[stem]["up"] = v.to(dtype=precision_dtype)
for stem in pairs:
pairs[stem]["alpha"] = alphas.get(stem, float(pairs[stem].get("rank", 1.0)))
return pairs
class ShardBuffer:
def __init__(self, max_size_gb, output_dir, output_repo, subfolder, hf_token, filename_prefix="model"):
self.max_bytes = int(max_size_gb * 1024**3)
self.output_dir = output_dir
self.output_repo = output_repo
self.subfolder = subfolder
self.hf_token = hf_token
self.filename_prefix = filename_prefix
self.buffer = []
self.current_bytes = 0
self.shard_count = 0
self.index_map = {}
self.total_size = 0
def add_tensor(self, key, tensor):
if tensor.dtype == torch.bfloat16:
raw_bytes = tensor.view(torch.int16).numpy().tobytes()
dtype_str = "BF16"
elif tensor.dtype == torch.float16:
raw_bytes = tensor.numpy().tobytes()
dtype_str = "F16"
else:
raw_bytes = tensor.numpy().tobytes()
dtype_str = "F32"
size = len(raw_bytes)
self.buffer.append({"key": key, "data": raw_bytes, "dtype": dtype_str, "shape": tensor.shape})
self.current_bytes += size
self.total_size += size
if self.current_bytes >= self.max_bytes: self.flush()
def flush(self):
if not self.buffer: return
self.shard_count += 1
filename = f"{self.filename_prefix}-{self.shard_count:05d}.safetensors"
path_in_repo = f"{self.subfolder}/{filename}" if self.subfolder else filename
header = {"__metadata__": {"format": "pt"}}
current_offset = 0
for item in self.buffer:
header[item["key"]] = {"dtype": item["dtype"], "shape": item["shape"], "data_offsets": [current_offset, current_offset + len(item["data"])]}
current_offset += len(item["data"])
self.index_map[item["key"]] = filename
header_json = json.dumps(header).encode('utf-8')
out_path = self.output_dir / filename
with open(out_path, 'wb') as f:
f.write(struct.pack('<Q', len(header_json)))
f.write(header_json)
for item in self.buffer: f.write(item["data"])
api.upload_file(path_or_fileobj=out_path, path_in_repo=path_in_repo, repo_id=self.output_repo, token=self.hf_token)
os.remove(out_path)
self.buffer = []
self.current_bytes = 0
gc.collect()
def streaming_copy_structure(token, src_repo, dst_repo, ignore_prefix=None, is_root_merge=False):
try:
files = api.list_repo_files(repo_id=src_repo, token=token)
for f in tqdm(files, desc="Copying Structure"):
if ignore_prefix and f.startswith(ignore_prefix): continue
if is_root_merge:
if any(f.endswith(ext) for ext in ['.safetensors', '.bin', '.pt', '.pth']): continue
try:
local = hf_hub_download(repo_id=src_repo, filename=f, token=token, local_dir=TempDir)
api.upload_file(path_or_fileobj=local, path_in_repo=f, repo_id=dst_repo, token=token)
if os.path.exists(local): os.remove(local)
except: pass
except: pass
def task_merge(hf_token, base_repo, base_subfolder, lora_input, scale, precision, shard_size, output_repo, structure_repo, private, progress=gr.Progress()):
cleanup_temp()
if not hf_token: return "Error: HF Token required."
login(hf_token.strip())
try: api.create_repo(repo_id=output_repo, private=private, exist_ok=True, token=hf_token)
except Exception as e: return f"Error creating repo: {e}"
output_subfolder = base_subfolder if base_subfolder else ""
if structure_repo:
ignore = output_subfolder if output_subfolder else None
streaming_copy_structure(hf_token, structure_repo, output_repo, ignore_prefix=ignore, is_root_merge=not bool(output_subfolder))
progress(0.1, desc="Downloading Input Model...")
files = list_repo_files(repo_id=base_repo, token=hf_token)
input_shards = []
for f in files:
if f.endswith(".safetensors"):
if output_subfolder and not f.startswith(output_subfolder): continue
local = TempDir / "inputs" / os.path.basename(f)
os.makedirs(local.parent, exist_ok=True)
hf_hub_download(repo_id=base_repo, filename=f, token=hf_token, local_dir=local.parent, local_dir_use_symlinks=False)
found = list(local.parent.rglob(os.path.basename(f)))
if found: input_shards.append(found[0])
if not input_shards: return "No safetensors found."
input_shards.sort()
filename_prefix = "diffusion_pytorch_model" if (output_subfolder in ["transformer", "unet"] or "diffusion_pytorch_model" in os.path.basename(input_shards[0])) else "model"
index_filename = f"{filename_prefix}.safetensors.index.json"
dtype = torch.bfloat16 if precision == "bf16" else torch.float16 if precision == "fp16" else torch.float32
try:
progress(0.15, desc="Downloading LoRA...")
lora_path = download_lora_smart(lora_input, hf_token)
lora_pairs = load_lora_to_memory(lora_path, precision_dtype=dtype)
except Exception as e: return f"Error loading LoRA: {e}"
buffer = ShardBuffer(shard_size, TempDir, output_repo, output_subfolder, hf_token, filename_prefix=filename_prefix)
for i, shard_file in enumerate(input_shards):
progress(0.2 + (0.7 * i / len(input_shards)), desc=f"Processing {os.path.basename(shard_file)}")
with MemoryEfficientSafeOpen(shard_file) as f:
for k in f.keys():
v = f.get_tensor(k)
base_stem = get_key_stem(k)
match = lora_pairs.get(base_stem)
if not match:
if "to_q" in base_stem: match = lora_pairs.get(base_stem.replace("to_q", "qkv"))
elif "to_k" in base_stem: match = lora_pairs.get(base_stem.replace("to_k", "qkv"))
elif "to_v" in base_stem: match = lora_pairs.get(base_stem.replace("to_v", "qkv"))
if match:
down, up = match["down"], match["up"]
scaling = scale * (match["alpha"] / match["rank"])
if len(v.shape) == 4 and len(down.shape) == 2:
down, up = down.unsqueeze(-1).unsqueeze(-1), up.unsqueeze(-1).unsqueeze(-1)
try:
delta = (up.squeeze() @ down.squeeze()).reshape(up.shape[0], down.shape[1], 1, 1) if len(up.shape) == 4 else up @ down
except: delta = up.T @ down
delta = delta * scaling
if delta.shape == v.shape: v = v.to(dtype).add_(delta.to(dtype))
del delta
buffer.add_tensor(k, v.to(dtype))
del v
os.remove(shard_file)
gc.collect()
buffer.flush()
index_data = {"metadata": {"total_size": buffer.total_size}, "weight_map": buffer.index_map}
path_in_repo = f"{output_subfolder}/{index_filename}" if output_subfolder else index_filename
with open(TempDir / index_filename, "w") as f: json.dump(index_data, f, indent=4)
api.upload_file(path_or_fileobj=TempDir / index_filename, path_in_repo=path_in_repo, repo_id=output_repo, token=hf_token)
cleanup_temp()
return f"Done! Merged {buffer.shard_count} shards to {output_repo}"
# =================================================================================
# TAB 2: EXTRACT LORA
# =================================================================================
def identify_and_download_model(input_str, token):
repo_id, filename = parse_hf_url(input_str)
if repo_id and filename:
local_path = TempDir / os.path.basename(filename)
if local_path.exists(): os.remove(local_path)
hf_hub_download(repo_id=repo_id, filename=filename, token=token, local_dir=TempDir)
return list(TempDir.rglob(os.path.basename(filename)))[0]
files = list_repo_files(repo_id=input_str, token=token)
priorities = ["transformer/diffusion_pytorch_model.safetensors", "unet/diffusion_pytorch_model.safetensors", "model.safetensors"]
target_file = next((f for f in priorities if f in files), next((f for f in files if f.endswith(".safetensors") and "lora" not in f), None))
if not target_file: raise ValueError("No model file found")
hf_hub_download(repo_id=input_str, filename=target_file, token=token, local_dir=TempDir)
return list(TempDir.rglob(os.path.basename(target_file)))[0]
def extract_lora_layer_by_layer(model_org, model_tuned, rank, clamp):
org = MemoryEfficientSafeOpen(model_org)
tuned = MemoryEfficientSafeOpen(model_tuned)
lora_sd = {}
keys = set(org.keys()).intersection(set(tuned.keys()))
for key in tqdm(keys, desc="Extracting"):
if "num_batches_tracked" in key or "running_mean" in key or "running_var" in key: continue
mat_org = org.get_tensor(key).float()
mat_tuned = tuned.get_tensor(key).float()
if mat_org.shape != mat_tuned.shape: continue
diff = mat_tuned - mat_org
if torch.max(torch.abs(diff)) < 1e-4: continue
out_dim, in_dim = diff.shape[0], diff.shape[1] if len(diff.shape) > 1 else 1
r = min(rank, in_dim, out_dim)
is_conv = len(diff.shape) == 4
if is_conv: diff = diff.flatten(start_dim=1)
elif len(diff.shape) == 1: diff = diff.unsqueeze(1)
U, S, V = torch.svd_lowrank(diff, q=r+4, niter=4)
Vh = V.t()
U, S, Vh = U[:, :r], S[:r], Vh[:r, :]
U = U @ torch.diag(S)
dist = torch.cat([U.flatten(), Vh.flatten()])
hi_val = torch.quantile(torch.abs(dist), clamp)
if hi_val > 0:
U, Vh = U.clamp(-hi_val, hi_val), Vh.clamp(-hi_val, hi_val)
if is_conv:
U = U.reshape(out_dim, r, 1, 1)
Vh = Vh.reshape(r, in_dim, mat_org.shape[2], mat_org.shape[3])
else:
U = U.reshape(out_dim, r)
Vh = Vh.reshape(r, in_dim)
stem = key.replace(".weight", "")
lora_sd[f"{stem}.lora_up.weight"] = U.contiguous()
lora_sd[f"{stem}.lora_down.weight"] = Vh.contiguous()
lora_sd[f"{stem}.alpha"] = torch.tensor(r).float()
out = TempDir / "extracted.safetensors"
save_file(lora_sd, out)
return str(out)
def task_extract(hf_token, org, tun, rank, out):
cleanup_temp()
if hf_token: login(hf_token.strip())
try:
p1 = identify_and_download_model(org, hf_token)
p2 = identify_and_download_model(tun, hf_token)
f = extract_lora_layer_by_layer(p1, p2, int(rank), 0.99)
api.create_repo(repo_id=out, exist_ok=True, token=hf_token)
api.upload_file(path_or_fileobj=f, path_in_repo="extracted_lora.safetensors", repo_id=out, token=hf_token)
return "Done! Extracted to " + out
except Exception as e: return f"Error: {e}"
# =================================================================================
# TAB 3: MERGE ADAPTERS
# =================================================================================
def load_full_state_dict(path):
raw = load_file(path, device="cpu")
cleaned = {}
for k, v in raw.items():
if "lora_A" in k: new_k = k.replace("lora_A", "lora_down")
elif "lora_B" in k: new_k = k.replace("lora_B", "lora_up")
else: new_k = k
cleaned[new_k] = v.float()
return cleaned
def task_merge_adapters_advanced(hf_token, inputs_text, method, weight_str, beta, sigma_rel, target_rank, out_repo, private):
cleanup_temp()
if hf_token: login(hf_token.strip())
urls = [line.strip() for line in inputs_text.replace(" ", "\n").split('\n') if line.strip()]
if len(urls) < 2: return "Error: Please provide at least 2 adapters."
try:
weights = [float(w.strip()) for w in weight_str.split(',')] if weight_str.strip() else [1.0] * len(urls)
if len(weights) < len(urls): weights += [1.0] * (len(urls) - len(weights))
except: return "Error parsing weights."
paths = []
try:
for url in tqdm(urls, desc="Downloading Adapters"): paths.append(download_lora_smart(url, hf_token))
except Exception as e: return f"Download Error: {e}"
merged = None
if "Iterative EMA" in method:
base_sd = load_file(paths[0], device="cpu")
for k in base_sd:
if base_sd[k].dtype.is_floating_point: base_sd[k] = base_sd[k].float()
gamma = None
if sigma_rel > 0:
t_val = sigma_rel**-2
roots = np.roots([1, 7, 16 - t_val, 12 - t_val])
gamma = roots[np.isreal(roots) & (roots.real >= 0)].real.max()
for i, path in enumerate(paths[1:]):
current_beta = (1 - 1 / (i + 1)) ** (gamma + 1) if gamma is not None else beta
curr = load_file(path, device="cpu")
for k in base_sd:
if k in curr and "alpha" not in k:
base_sd[k] = base_sd[k] * current_beta + curr[k].float() * (1 - current_beta)
merged = base_sd
else:
states = [load_full_state_dict(p) for p in paths]
merged = {}
all_stems = set()
for s in states:
for k in s:
if "lora_" in k: all_stems.add(k.split(".lora_")[0])
for stem in tqdm(all_stems):
down_list, up_list = [], []
alpha_sum = 0.0
total_delta = None
for i, state in enumerate(states):
w = weights[i]
dk, uk, ak = f"{stem}.lora_down.weight", f"{stem}.lora_up.weight", f"{stem}.alpha"
if dk in state and uk in state:
d, u = state[dk], state[uk]
alpha_sum += state[ak].item() if ak in state else d.shape[0]
if "Concatenation" in method:
down_list.append(d)
up_list.append(u * w)
elif "SVD" in method:
rank, alpha = d.shape[0], state[ak].item() if ak in state else d.shape[0]
scale = (alpha / rank) * w
delta = ((u.flatten(1) @ d.flatten(1)).reshape(u.shape[0], d.shape[1], d.shape[2], d.shape[3]) if len(d.shape)==4 else u @ d) * scale
total_delta = delta if total_delta is None else total_delta + delta
if "Concatenation" in method and down_list:
merged[f"{stem}.lora_down.weight"] = torch.cat(down_list, dim=0).contiguous()
merged[f"{stem}.lora_up.weight"] = torch.cat(up_list, dim=1).contiguous()
merged[f"{stem}.alpha"] = torch.tensor(alpha_sum)
elif "SVD" in method and total_delta is not None:
tr = int(target_rank)
flat = total_delta.flatten(1) if len(total_delta.shape)==4 else total_delta
try:
U, S, V = torch.svd_lowrank(flat, q=tr + 4, niter=4)
Vh = V.t()
U, S, Vh = U[:, :tr], S[:tr], Vh[:tr, :]
U = U @ torch.diag(S)
if len(total_delta.shape) == 4:
U = U.reshape(total_delta.shape[0], tr, 1, 1)
Vh = Vh.reshape(tr, total_delta.shape[1], total_delta.shape[2], total_delta.shape[3])
else:
U, Vh = U.reshape(total_delta.shape[0], tr), Vh.reshape(tr, total_delta.shape[1])
merged[f"{stem}.lora_down.weight"] = Vh.contiguous()
merged[f"{stem}.lora_up.weight"] = U.contiguous()
merged[f"{stem}.alpha"] = torch.tensor(tr).float()
except: pass
out = TempDir / "merged_adapters.safetensors"
save_file(merged, out)
api.create_repo(repo_id=out_repo, private=private, exist_ok=True, token=hf_token)
api.upload_file(path_or_fileobj=out, path_in_repo="merged_adapters.safetensors", repo_id=out_repo, token=hf_token)
return f"Success! Merged to {out_repo}"
# =================================================================================
# TAB 4: RESIZE
# =================================================================================
def task_resize(hf_token, lora_input, new_rank, dynamic_method, dynamic_param, out_repo):
cleanup_temp()
if hf_token: login(hf_token.strip())
path = download_lora_smart(lora_input, hf_token)
state = load_file(path, device="cpu")
new_state = {}
groups = {}
for k in state:
simple = k.split(".lora_")[0]
if simple not in groups: groups[simple] = {}
if "lora_down" in k or "lora_A" in k: groups[simple]["down"] = state[k]
if "lora_up" in k or "lora_B" in k: groups[simple]["up"] = state[k]
if "alpha" in k: groups[simple]["alpha"] = state[k]
target_rank_limit = int(new_rank)
for stem, g in tqdm(groups.items()):
if "down" in g and "up" in g:
down, up = g["down"].float(), g["up"].float()
merged = (up.squeeze() @ down.squeeze()).reshape(up.shape[0], down.shape[1], down.shape[2], down.shape[3]) if len(down.shape)==4 else up @ down
flat = merged.flatten(1)
U, S, V = torch.svd_lowrank(flat, q=target_rank_limit + 32)
Vh = V.t()
calc_rank = target_rank_limit
if dynamic_method == "sv_ratio":
calc_rank = int(torch.sum(S > (S[0] / dynamic_param)).item())
elif dynamic_method == "sv_cumulative":
calc_rank = int(torch.searchsorted(torch.cumsum(S, 0) / torch.sum(S), dynamic_param)) + 1
elif dynamic_method == "sv_fro":
calc_rank = int(torch.searchsorted(torch.cumsum(S.pow(2), 0) / torch.sum(S.pow(2)), dynamic_param**2)) + 1
final_rank = max(1, min(calc_rank, target_rank_limit, S.shape[0]))
U = U[:, :final_rank] @ torch.diag(S[:final_rank])
Vh = Vh[:final_rank, :]
if len(down.shape) == 4:
U = U.reshape(up.shape[0], final_rank, 1, 1)
Vh = Vh.reshape(final_rank, down.shape[1], down.shape[2], down.shape[3])
new_state[f"{stem}.lora_down.weight"] = Vh.contiguous()
new_state[f"{stem}.lora_up.weight"] = U.contiguous()
new_state[f"{stem}.alpha"] = torch.tensor(final_rank).float()
out = TempDir / "shrunken.safetensors"
save_file(new_state, out)
api.create_repo(repo_id=out_repo, exist_ok=True, token=hf_token)
api.upload_file(path_or_fileobj=out, path_in_repo="shrunken.safetensors", repo_id=out_repo, token=hf_token)
return "Done"
# =================================================================================
# NEW HELPERS: MERGEKIT CLI WRAPPER
# =================================================================================
def run_mergekit_cli(config_str, output_path, hf_token):
# This replaces the Python API call to avoid 'unexpected keyword' errors
# Writes config to file -> runs `mergekit-yaml` subprocess -> returns path
config_file = TempDir / "config.yaml"
with open(config_file, "w") as f:
f.write(config_str)
# Ensure token is in env for subprocess
env = os.environ.copy()
if hf_token:
env["HF_TOKEN"] = hf_token.strip()
cmd = [
"mergekit-yaml",
str(config_file),
str(output_path),
"--allow-crimes", # Allows mixing architectures if needed
"--lazy-unpickle", # Memory optimization
"--copy-tokenizer"
]
# Run
print(f"Running command: {' '.join(cmd)}")
result = subprocess.run(cmd, env=env, capture_output=True, text=True)
if result.returncode != 0:
raise RuntimeError(f"MergeKit CLI Failed:\n{result.stderr}")
return str(output_path)
def upload_folder_to_hf(folder, repo_id, token, private=True):
api.create_repo(repo_id=repo_id, private=private, exist_ok=True, token=token)
api.upload_folder(folder_path=folder, repo_id=repo_id, token=token)
return f"Success! Uploaded to {repo_id}"
# =================================================================================
# TAB 5: WEIGHTED & SPARSIFIED (Linear, Ties, Dare)
# =================================================================================
def task_mergekit_weighted(hf_token, models_text, method, dtype, base_model, weights, density, normalize, out_repo, private):
cleanup_temp()
if not hf_token: return "Error: Token required"
login(hf_token.strip())
model_list = [m.strip() for m in models_text.split('\n') if m.strip()]
if not model_list: return "Error: No models listed"
# Build Config
config = {}
if method == "linear":
# Linear/Model Stock usually structure:
# models:
# - model: x
# parameters:
# weight: 1.0
c_models = []
w_list = [float(x) for x in weights.split(',')] if weights.strip() else [1.0] * len(model_list)
if len(w_list) < len(model_list): w_list += [1.0] * (len(model_list) - len(w_list))
for i, m in enumerate(model_list):
c_models.append({"model": m, "parameters": {"weight": w_list[i]}})
config = {"models": c_models, "merge_method": method, "dtype": dtype}
else:
# TIES / DARE / ETC
c_models = []
w_list = [float(x) for x in weights.split(',')] if weights.strip() else [1.0] * len(model_list)
for i, m in enumerate(model_list):
item = {"model": m, "parameters": {"weight": w_list[i] if i < len(w_list) else 1.0}}
if density and method in ["dare_ties", "dare_linear", "ties"]:
item["parameters"]["density"] = float(density)
c_models.append(item)
config = {
"models": c_models,
"merge_method": method,
"base_model": base_model if base_model else model_list[0],
"parameters": {
"normalize": normalize,
"int8_mask": True if "dare" in method else False
},
"dtype": dtype
}
yaml_str = yaml.dump(config, sort_keys=False)
out_path = TempDir / "out_merged"
try:
run_mergekit_cli(yaml_str, out_path, hf_token)
return upload_folder_to_hf(str(out_path), out_repo, hf_token, private)
except Exception as e:
return f"Error: {e}"
def task_mergekit_interp(hf_token, model_a, model_b, base_model, method, t_val, dtype, out_repo, private):
cleanup_temp()
if not hf_token: return "Error: Token required"
login(hf_token.strip())
config = {}
if method in ["slerp", "nuslerp"]:
config = {
"slices": [
{
"sources": [
{"model": model_a, "layer_range": [0, 32]}, # Default full range assumption
{"model": model_b, "layer_range": [0, 32]}
],
"parameters": {
"t": float(t_val)
}
}
],
"merge_method": method,
"base_model": model_a, # Slerp needs a base usually just for config
"dtype": dtype
}
elif method == "task_arithmetic":
config = {
"models": [
{"model": model_a, "parameters": {"weight": 1.0}},
{"model": model_b, "parameters": {"weight": -1.0}} # Simple subtraction example
],
"base_model": base_model if base_model else model_a,
"merge_method": method,
"dtype": dtype
}
# Correcting for generic usage
# If Task Arithmetic is selected, let's allow more generic standard config
if method == "task_arithmetic":
config = {
"base_model": base_model if base_model else model_a,
"merge_method": "task_arithmetic",
"models": [
{"model": model_a, "parameters": {"weight": 1.0}},
{"model": model_b, "parameters": {"weight": float(t_val)}}
],
"dtype": dtype
}
yaml_str = yaml.dump(config, sort_keys=False)
out_path = TempDir / "out_interp"
try:
run_mergekit_cli(yaml_str, out_path, hf_token)
return upload_folder_to_hf(str(out_path), out_repo, hf_token, private)
except Exception as e:
return f"Error: {e}"
def task_mergekit_moe(hf_token, base_model, experts_text, gate_mode, dtype, out_repo, private):
cleanup_temp()
if not hf_token: return "Error: Token required"
login(hf_token.strip())
experts = [e.strip() for e in experts_text.split('\n') if e.strip()]
if not experts: return "Error: No experts listed"
# Construct MoE config
formatted_experts = []
for e in experts:
formatted_experts.append({
"source_model": e,
"positive_prompts": [""] # Simplified for GUI
})
config = {
"base_model": base_model if base_model else experts[0],
"gate_mode": gate_mode,
"dtype": dtype,
"experts": formatted_experts
}
yaml_str = yaml.dump(config, sort_keys=False)
out_path = TempDir / "out_moe"
try:
run_mergekit_cli(yaml_str, out_path, hf_token)
return upload_folder_to_hf(str(out_path), out_repo, hf_token, private)
except Exception as e:
return f"Error: {e}"
# =================================================================================
# TAB 8: RAW PYTORCH (Passthrough / Non-Transformer)
# =================================================================================
def task_raw_merge(hf_token, models_text, method, dtype, out_repo, private):
cleanup_temp()
if not hf_token: return "Error: Token required"
login(hf_token.strip())
models = [m.strip() for m in models_text.split('\n') if m.strip()]
# For Raw/Passthrough, we basically treat it like linear but with passthrough method
# Or simple linear
config = {
"models": [{"model": m, "parameters": {"weight": 1.0}} for m in models],
"merge_method": method, # passthrough, linear
"dtype": dtype
}
yaml_str = yaml.dump(config, sort_keys=False)
out_path = TempDir / "out_raw"
try:
run_mergekit_cli(yaml_str, out_path, hf_token)
return upload_folder_to_hf(str(out_path), out_repo, hf_token, private)
except Exception as e:
return f"Error: {e}"
def task_dare_soonr(hf_token, base_model, ft_model, ratio, mask_rate, out_repo, private):
# Ported from the requested DARE-MERGE-SOONR implementation
cleanup_temp()
if not hf_token: return "Error: Token required"
login(hf_token.strip())
try:
print("Downloading Base...")
base_path = identify_and_download_model(base_model, hf_token)
print("Downloading FT...")
ft_path = identify_and_download_model(ft_model, hf_token)
print("Loading Tensors...")
base_sd = load_file(base_path, device="cpu")
ft_sd = load_file(ft_path, device="cpu")
merged_sd = {}
common_keys = set(base_sd.keys()).intersection(set(ft_sd.keys()))
print("Merging...")
for key in tqdm(common_keys):
base_t = base_sd[key]
ft_t = ft_sd[key]
if base_t.dtype != ft_t.dtype or base_t.shape != ft_t.shape:
merged_sd[key] = ft_t # Fallback
continue
# DARE Logic
# 1. Delta
delta = ft_t.float() - base_t.float()
# 2. Mask (Drop)
if mask_rate > 0.0:
# Bernoulli mask
mask = torch.bernoulli(torch.full_like(delta, 1.0 - mask_rate))
# Rescale
rescale_factor = 1.0 / (1.0 - mask_rate)
delta = delta * mask * rescale_factor
# 3. Apply Ratio and Add
merged_t = base_t.float() + (delta * ratio)
# Cast back
if base_t.dtype == torch.bfloat16:
merged_sd[key] = merged_t.bfloat16()
elif base_t.dtype == torch.float16:
merged_sd[key] = merged_t.half()
else:
merged_sd[key] = merged_t
# Save
out_path = TempDir / "dare_merged.safetensors"
save_file(merged_sd, out_path)
# Upload
api.create_repo(repo_id=out_repo, private=private, exist_ok=True, token=hf_token)
api.upload_file(path_or_fileobj=out_path, path_in_repo="model.safetensors", repo_id=out_repo, token=hf_token)
return f"Success! Uploaded to {out_repo}"
except Exception as e:
return f"DARE Error: {e}"
# =================================================================================
# UI
# =================================================================================
css = ".container { max-width: 1100px; margin: auto; }"
with gr.Blocks() as demo:
title = gr.HTML(
"""<h1><img src="https://huggingface.co/spaces/AlekseyCalvin/Soon_Merger/resolve/main/SMerger3.png" alt="SOONmerge®"> Transform Transformers for FREE!</h1>""",
elem_id="title",
)
gr.Markdown("# 🧰Training-Free CPU-run Model Creation Toolkit")
with gr.Tabs():
with gr.Tab("Merge into Base Model"):
with gr.Row():
t1_token = gr.Textbox(label="Token", type="password")
with gr.Row():
t1_base = gr.Textbox(label="Base Repo", value="name/repo")
t1_sub = gr.Textbox(label="Subfolder (Optional)", value="")
t1_lora = gr.Textbox(label="LoRA Direct Link or Repo", value="https://huggingface.co/GuangyuanSD/Z-Image-Re-Turbo-LoRA/resolve/main/Z-image_re_turbo_lora_8steps_rank_32_v1_fp16.safetensors")
with gr.Row():
t1_scale = gr.Slider(label="Scale", value=1.0, minimum=0, maximum=3.0, step=0.1)
t1_prec = gr.Radio(["bf16", "fp16", "float32"], value="bf16", label="Precision")
t1_shard = gr.Slider(label="Max Shard Size (GB)", value=2.0, minimum=0.1, maximum=10.0, step=0.1)
t1_out = gr.Textbox(label="Output Repo")
t1_struct = gr.Textbox(label="Extras Source (copies configs/components/etc)", value="name/repo")
t1_priv = gr.Checkbox(label="Private", value=True)
t1_btn = gr.Button("Merge")
t1_res = gr.Textbox(label="Result")
t1_btn.click(task_merge, [t1_token, t1_base, t1_sub, t1_lora, t1_scale, t1_prec, t1_shard, t1_out, t1_struct, t1_priv], t1_res)
with gr.Tab("Extract Adapter"):
t2_token = gr.Textbox(label="Token", type="password")
t2_org = gr.Textbox(label="Original Model")
t2_tun = gr.Textbox(label="Tuned or Homologous Model")
t2_rank = gr.Number(label="Extract At Rank", value=32, minimum=1, maximum=1024, step=1)
t2_out = gr.Textbox(label="Output Repo")
t2_btn = gr.Button("Extract")
t2_res = gr.Textbox(label="Result")
t2_btn.click(task_extract, [t2_token, t2_org, t2_tun, t2_rank, t2_out], t2_res)
with gr.Tab("Merge Adapters"):
gr.Markdown("### Batch Adapter Merging")
t3_token = gr.Textbox(label="Token", type="password")
t3_urls = gr.TextArea(label="Adapter URLs/Repos (one per line, or space-separated)", placeholder="user/lora1\nhttps://hf.co/user/lora2.safetensors\n...")
with gr.Row():
t3_method = gr.Dropdown(
["Iterative EMA (Linear w/ Beta/Sigma coefficient)", "Concatenation (MOE-like weights-stack)", "SVD Fusion (Task Arithmetic/Compressed)"],
value="Iterative EMA (Linear w/ Beta/Sigma coefficient)",
label="Merge Method"
)
with gr.Row():
t3_weights = gr.Textbox(label="Weights (comma-separated) – for Concat/SVD", placeholder="1.0, 0.5, 0.8...")
t3_rank = gr.Number(label="Target Rank – For SVD only", value=128, minimum=1, maximum=1024)
with gr.Row():
t3_beta = gr.Slider(label="Beta – for linear/post-hoc EMA", value=0.95, minimum=0.01, maximum=1.00, step=0.01)
t3_sigma = gr.Slider(label="Sigma Rel – for linear/post-hoc EMA", value=0.21, minimum=0.01, maximum=1.00, step=0.01)
t3_out = gr.Textbox(label="Output Repo")
t3_priv = gr.Checkbox(label="Private Output", value=True)
t3_btn = gr.Button("Merge")
t3_res = gr.Textbox(label="Result")
t3_btn.click(task_merge_adapters_advanced, [t3_token, t3_urls, t3_method, t3_weights, t3_beta, t3_sigma, t3_rank, t3_out, t3_priv], t3_res)
with gr.Tab("Resize Adapter"):
t4_token = gr.Textbox(label="Token", type="password")
t4_in = gr.Textbox(label="LoRA")
with gr.Row():
t4_rank = gr.Number(label="To Rank (Safety Ceiling)", value=8, minimum=1, maximum=512, step=1)
t4_method = gr.Dropdown(["None", "sv_ratio", "sv_fro", "sv_cumulative"], value="None", label="Dynamic Method")
t4_param = gr.Number(label="Dynamic Param", value=0.9)
gr.Markdown(
"""
### 📉 Dynamic Resizing Guide
These methods intelligently determine the best rank per layer.
* **sv_ratio (Relative Strength):** Keeps features that are at least `1/Param` as strong as the main feature. **Param must be >= 2**. (e.g. 2 = keep features half as strong as top).
* **sv_fro (Visual Information Density):** Preserves `Param%` of the total information content (Frobenius Norm) of the layer. **Param between 0.0 and 1.0** (e.g. 0.9 = 90% info retention).
* **sv_cumulative (Cumulative Sum):** Preserves weights that sum up to `Param%` of the total strength. **Param between 0.0 and 1.0**.
* **⚠️ Safety Ceiling:** The **"To Rank"** slider acts as a hard limit. Even if a dynamic method wants a higher rank, it will be cut down to this number to keep file sizes small.
"""
)
t4_out = gr.Textbox(label="Output")
t4_btn = gr.Button("Resize")
t4_res = gr.Textbox(label="Result")
t4_btn.click(task_resize, [t4_token, t4_in, t4_rank, t4_method, t4_param, t4_out], t4_res)
with gr.Tab("Stir/Tie Bases"):
gr.Markdown("### Linear, TIES, dare-TIES, Model Stock")
t5_token = gr.Textbox(label="HF Token", type="password")
with gr.Row():
t5_method = gr.Dropdown(["linear", "ties", "dare_ties", "dare_linear", "model_stock"], value="linear", label="Method")
t5_dtype = gr.Dropdown(["float16", "bfloat16", "float32"], value="bfloat16", label="Dtype")
t5_models = gr.TextArea(label="Models (one per line)")
with gr.Row():
t5_base = gr.Textbox(label="Base Model (Optional)")
t5_weights = gr.Textbox(label="Weights (comma sep)", placeholder="1.0, 0.5")
with gr.Row():
t5_density = gr.Textbox(label="Density (for DARE/TIES)", placeholder="0.5")
t5_norm = gr.Checkbox(label="Normalize", value=True)
t5_out = gr.Textbox(label="Output Repo")
t5_priv = gr.Checkbox(label="Private", value=True)
t5_btn = gr.Button("Run MergeKit (CLI)")
t5_res = gr.Textbox(label="Result")
t5_btn.click(task_mergekit_weighted, [t5_token, t5_models, t5_method, t5_dtype, t5_base, t5_weights, t5_density, t5_norm, t5_out, t5_priv], t5_res)
with gr.Tab("Amphinterpolative"):
gr.Markdown("### Slerp, Task Arithmetic, NuSlerp")
t6_token = gr.Textbox(label="HF Token", type="password")
with gr.Row():
t6_model_a = gr.Textbox(label="Model A")
t6_model_b = gr.Textbox(label="Model B")
with gr.Row():
t6_method = gr.Dropdown(["slerp", "nuslerp", "task_arithmetic"], value="slerp", label="Method")
t6_dtype = gr.Dropdown(["float16", "bfloat16"], value="bfloat16", label="Dtype")
t6_t = gr.Textbox(label="t (Interpolation factor)", value="0.5")
t6_base = gr.Textbox(label="Base Model (for Task Arithmetic)", placeholder="Same as A usually")
t6_out = gr.Textbox(label="Output Repo")
t6_priv = gr.Checkbox(label="Private", value=True)
t6_btn = gr.Button("Run MergeKit (CLI)")
t6_res = gr.Textbox(label="Result")
t6_btn.click(task_mergekit_interp, [t6_token, t6_model_a, t6_model_b, t6_base, t6_method, t6_t, t6_dtype, t6_out, t6_priv], t6_res)
with gr.Tab("MoEr"):
gr.Markdown("### Mixture of Experts Construction")
t7_token = gr.Textbox(label="HF Token", type="password")
t7_base = gr.Textbox(label="Base Model")
t7_experts = gr.TextArea(label="Experts (one per line)")
with gr.Row():
t7_gate = gr.Dropdown(["cheap_embed", "random", "hidden"], value="cheap_embed", label="Gate Mode")
t7_dtype = gr.Dropdown(["float16", "bfloat16"], value="bfloat16", label="Dtype")
t7_out = gr.Textbox(label="Output Repo")
t7_priv = gr.Checkbox(label="Private", value=True)
t7_btn = gr.Button("Build MoE (CLI)")
t7_res = gr.Textbox(label="Result")
t7_btn.click(task_mergekit_moe, [t7_token, t7_base, t7_experts, t7_gate, t7_dtype, t7_out, t7_priv], t7_res)
with gr.Tab("Rawer"):
gr.Markdown("### Raw PyTorch MergeKit / Non-pipeline-classed")
t8_token = gr.Textbox(label="HF Token", type="password")
t8_models = gr.TextArea(label="Models (one per line)")
t8_method = gr.Dropdown(["linear", "passthrough"], value="linear", label="Method")
t8_dtype = gr.Dropdown(["float32", "float16", "bfloat16"], value="float32", label="dtype")
t8_out = gr.Textbox(label="Output Repo")
t8_priv = gr.Checkbox(label="Private", value=True)
t8_btn = gr.Button("Merge")
t8_res = gr.Textbox(label="Result")
t8_btn.click(task_raw_merge, [t8_token, t8_models, t8_method, t8_dtype, t8_out, t8_priv], t8_res)
with gr.Tab("Mario,DARE!"):
gr.Markdown("### From sft-merger by [Martyn Garcia](https://github.com/martyn)")
t9_token = gr.Textbox(label="HF Token", type="password")
with gr.Row():
t9_base = gr.Textbox(label="Base Model")
t9_ft = gr.Textbox(label="Fine-Tuned Model")
with gr.Row():
t9_ratio = gr.Slider(0, 2, 1.0, label="Ratio")
t9_mask = gr.Slider(0, 0.99, 0.5, label="Mask Rate (Drop)")
t9_out = gr.Textbox(label="Output Repo")
t9_priv = gr.Checkbox(label="Private", value=True)
t9_btn = gr.Button("Run DARE Custom")
t9_res = gr.Textbox(label="Result")
t9_btn.click(task_dare_soonr, [t9_token, t9_base, t9_ft, t9_ratio, t9_mask, t9_out, t9_priv], t9_res)
if __name__ == "__main__":
demo.queue().launch(css=css, ssr_mode=False) |