pass trigger_text as arg

trainer/utils/inference.py

@@ -11,7 +11,7 @@ from .prompt import prepare_prompt_for_lora
 from .io import make_validation_img_grid
 
 @torch.no_grad()
-def render_images(training_pipeline, render_size, lora_path, train_step, seed, is_lora, pretrained_model, lora_scale = 0.7, n_steps = 25, n_imgs = 4, device = "cuda:0", verbose: bool = True):
+def render_images(training_pipeline, render_size, lora_path, train_step, seed, is_lora, pretrained_model, trigger_text: str, lora_scale = 0.7, n_steps = 25, n_imgs = 4, device = "cuda:0", verbose: bool = True):
 
     random.seed(seed)
 
@@ -55,7 +55,7 @@ def render_images(training_pipeline, render_size, lora_path, train_step, seed, i
         training_scheduler = pipeline.scheduler
     
     pipeline.scheduler = EulerDiscreteScheduler.from_config(pipeline.scheduler.config)
-    validation_prompts = [prepare_prompt_for_lora(prompt, lora_path, verbose=verbose) for prompt in validation_prompts_raw]
+    validation_prompts = [prepare_prompt_for_lora(prompt, lora_path, verbose=verbose, trigger_text=trigger_text) for prompt in validation_prompts_raw]
     generator = torch.Generator(device=device).manual_seed(0)
     pipeline_args = {
                 "negative_prompt": "nude, naked, poorly drawn face, ugly, tiling, out of frame, extra limbs, disfigured, deformed body, blurry, blurred, watermark, text, grainy, signature, cut off, draft", 

trainer/utils/prompt.py

@@ -14,7 +14,7 @@ def replace_in_string(s, replacements):
             break
     return s
 
-def prepare_prompt_for_lora(prompt, lora_path, interpolation=False, verbose=True):
+def prepare_prompt_for_lora(prompt, lora_path, trigger_text: str, interpolation=False, verbose=True):
     if "_no_token" in lora_path:
         return prompt
         
@@ -35,10 +35,10 @@ def prepare_prompt_for_lora(prompt, lora_path, interpolation=False, verbose=True
     try:
         lora_name = str(training_args["name"])
     except: # fallback for old loras that dont have the name field:
-        return training_args["trigger_text"] + ", " + prompt
+        return trigger_text + ", " + prompt
 
     lora_name_encapsulated = "<" + lora_name + ">"
-    trigger_text = training_args["trigger_text"]
+    trigger_text = trigger_text
 
     try:
         mode = training_args["concept_mode"]