fix embeddings load

CHANGELOG.md

@@ -7,7 +7,6 @@
 - PixArt-Σ
 - IP adapter masking
 - InstantStyle
-- <https://github.com/huggingface/diffusers/issues/7902>
 
 ### Note
 
@@ -166,6 +165,7 @@ it is time to give credit to original [author](https://github.com/auTOMATIC1111)
   - Control module restore button full functionality
   - Control improved handling with multiple control units and different init images
   - Control add correct metadata to image
+  - Time embeddings load part of model load
   - A1111 update OptionInfo properties
   - MOTD exception handling
   - Notifications not triggering

modules/sd_models.py

@@ -1129,12 +1129,6 @@ def load_diffuser(checkpoint_info=None, already_loaded_state_dict=None, timer=No
             sd_model.set_progress_bar_config(bar_format='Progress {rate_fmt}{postfix} {bar} {percentage:3.0f}% {n_fmt}/{total_fmt} {elapsed} {remaining}', ncols=80, colour='#327fba')
 
         sd_unet.load_unet(sd_model)
-
-        from modules.textual_inversion import textual_inversion
-        sd_model.embedding_db = textual_inversion.EmbeddingDatabase()
-        sd_model.embedding_db.add_embedding_dir(shared.opts.embeddings_dir)
-        sd_model.embedding_db.load_textual_inversion_embeddings(force_reload=True)
-
         set_diffuser_options(sd_model, vae, op)
 
         if op == 'refiner' and shared.opts.diffusers_move_refiner:
@@ -1160,8 +1154,14 @@ def load_diffuser(checkpoint_info=None, already_loaded_state_dict=None, timer=No
         model_data.sd_refiner = sd_model
     else:
         model_data.sd_model = sd_model
-
     timer.record("load")
+
+    from modules.textual_inversion import textual_inversion
+    sd_model.embedding_db = textual_inversion.EmbeddingDatabase()
+    sd_model.embedding_db.add_embedding_dir(shared.opts.embeddings_dir)
+    sd_model.embedding_db.load_textual_inversion_embeddings(force_reload=True)
+    timer.record("embeddings")
+
     devices.torch_gc(force=True)
     if shared.cmd_opts.profile:
         errors.profile(pr, 'Load')