Public/automatic
1
0
Fork 0
mirror of https://github.com/vladmandic/automatic synced 2026-09-19 01:04:32 +02:00
Code Issues Packages Projects Releases Wiki Activity

Fix TI loading from legacy formats

Browse Source
This commit is contained in:
AI-Casanova
2024-01-25 22:48:50 -06:00
committed by Vladimir Mandic
parent e924cc9ecd
commit 1440e07f02
1 changed files with 12 additions and 12 deletions
Download Patch File Download Diff File Expand all files Collapse all files
modules/textual_inversion/textual_inversion.py
+12 -12
View File
@@ -162,14 +162,14 @@ class EmbeddingDatabase:
if not isinstance(filename, list)
else filename
)
exts = [".SAFETENSORS", ".PT"]
exts = [".SAFETENSORS", '.BIN', '.PT', '.PNG', '.WEBP', '.JXL', '.AVIF'] # SDXL only uses safetensors
filename_paths = zip(filenames, len(filenames) * [path] if (isinstance(path, str) or path is None) else path)
model_type = None
if clip_g is None and tokenizer_2 is None:
model_type = 'SD'
elif clip_g and tokenizer_2:
model_type = 'SD-XL'
model_type = 'SDXL'
else:
model_type = 'UNDEFINED'
try:
@@ -213,17 +213,17 @@ class EmbeddingDatabase:
with safetensors.torch.safe_open(embedding.filename, framework="pt") as f: # type: ignore
for k in f.keys():
embeddings_dict[k] = f.get_tensor(k)
else:
raise NotImplementedError(f'extension {ext} not supported')
else: # fallback for sd1.5 pt embeddings
embeddings_dict["clip_l"] = self.load_from_file(embedding.filename, embedding.filename)
if 'clip_l' not in embeddings_dict:
raise ValueError('Invalid Embedding, dict missing required key `clip_l`')
if 'clip_g' in embeddings_dict:
embedding_type = 'SD-XL'
embedding_type = 'SDXL'
else:
embedding_type = 'SD'
if embedding_type != model_type:
raise ValueError(f'Unable to load `{embedding_type}` Embedding into `{model_type}` Model')
raise ValueError(f'Unable to load {embedding_type} Embedding "{embedding.name}" into {model_type} Model')
_tokens_to_add = {}
for i in range(len(embeddings_dict["clip_l"])):
if len(clip_l.get_input_embeddings().weight.data[0]) == len(embeddings_dict["clip_l"][i]):
@@ -246,14 +246,14 @@ class EmbeddingDatabase:
if len(tokens_to_add) > 0:
tokenizer.add_tokens(list(tokens_to_add.keys()))
clip_l.resize_token_embeddings(len(tokenizer))
if model_type == 'SD-XL':
if model_type == 'SDXL':
tokenizer_2.add_tokens(list(tokens_to_add.keys())) # type: ignore
clip_g.resize_token_embeddings(len(tokenizer_2)) # type: ignore
for token, data in tokens_to_add.items():
token_id = tokenizer.convert_tokens_to_ids(token)
if token_id > unk_token_id:
clip_l.get_input_embeddings().weight.data[token_id] = data.clip_l
if model_type == 'SD-XL':
if model_type == 'SDXL':
clip_g.get_input_embeddings().weight.data[token_id] = data.clip_g # type: ignore
except Exception as e:
errors.display(e, 'Embedding Load Failure')
@@ -270,16 +270,13 @@ class EmbeddingDatabase:
continue
self.skipped_embeddings[embedding.name] = embedding
debug(f"TI Loading: Text Encoder total embeddings={shared.sd_model.text_encoder.get_input_embeddings().weight.data.shape[0]}")
if model_type == 'SD-XL':
if model_type == 'SDXL':
debug(f"TI Loading: Text Encoder 2 total embeddings={shared.sd_model.text_encoder_2.get_input_embeddings().weight.data.shape[0]}")
return len(self.word_embeddings) - _loaded_pre
def load_from_file(self, path, filename):
name, ext = os.path.splitext(filename)
ext = ext.upper()
if shared.backend == shared.Backend.DIFFUSERS:
self.load_diffusers_embedding(filename, path)
return
if ext in ['.PNG', '.WEBP', '.JXL', '.AVIF']:
if '.preview' in filename.lower():
@@ -315,6 +312,9 @@ class EmbeddingDatabase:
else:
raise RuntimeError(f"Couldn't identify {filename} as textual inversion embedding")
if shared.backend == shared.Backend.DIFFUSERS:
return emb
vec = emb.detach().to(devices.device, dtype=torch.float32)
# name = data.get('name', name)
embedding = Embedding(vec=vec, name=name, filename=path)