Merge pull request #6667 from Shondoit/zero-vector-ti

Allow creation of zero vectors for TI

javascript/hints.js

@@ -92,6 +92,7 @@ titles = {
     "Weighted sum": "Result = A * (1 - M) + B * M",
     "Add difference": "Result = A + (B - C) * M",
 
+	"Initialization text": "If the number of tokens is more than the number of vectors, some may be skipped.\nLeave the textbox empty to start with zeroed out vectors",
     "Learning rate": "How fast should training go. Low values will take longer to train, high values may fail to converge (not generate accurate results) and/or may break the embedding (This has happened if you see Loss: nan in the training info textbox. If this happens, you need to manually restore your embedding from an older not-broken backup).\n\nYou can set a single numeric value, or multiple learning rates using the syntax:\n\n   rate_1:max_steps_1, rate_2:max_steps_2, ...\n\nEG:   0.005:100, 1e-3:1000, 1e-5\n\nWill train with rate of 0.005 for first 100 steps, then 1e-3 until 1000 steps, then 1e-5 for all remaining steps.",
 
     "Clip skip": "Early stopping parameter for CLIP model; 1 is stop at last layer as usual, 2 is stop at penultimate layer, etc.",

modules/textual_inversion/textual_inversion.py

@@ -248,11 +248,14 @@ def create_embedding(name, num_vectors_per_token, overwrite_old, init_text='*'):
     with devices.autocast():
         cond_model([""])  # will send cond model to GPU if lowvram/medvram is active
 
-    embedded = cond_model.encode_embedding_init_text(init_text, num_vectors_per_token)
+    #cond_model expects at least some text, so we provide '*' as backup.
+    embedded = cond_model.encode_embedding_init_text(init_text or '*', num_vectors_per_token)
     vec = torch.zeros((num_vectors_per_token, embedded.shape[1]), device=devices.device)
 
-    for i in range(num_vectors_per_token):
-        vec[i] = embedded[i * int(embedded.shape[0]) // num_vectors_per_token]
+    #Only copy if we provided an init_text, otherwise keep vectors as zeros
+    if init_text:
+        for i in range(num_vectors_per_token):
+            vec[i] = embedded[i * int(embedded.shape[0]) // num_vectors_per_token]
 
     # Remove illegal characters from name.
     name = "".join( x for x in name if (x.isalnum() or x in "._- "))