SelfAttentionGuidance updates (#1725)

* SAG updates

* restore settings from infotext
* bypass if model is Flux (would error during generation)
* bypass if CFG == 1 (would error during generation)
* add control of blur mask threshold (previously hardcoded)
* change to InputAccordion

* forge_sag.py: update title

extensions-builtin/sd_forge_sag/scripts/forge_sag.py

@@ -6,7 +6,8 @@ from backend.sampling.sampling_function import calc_cond_uncond_batch
 from backend import attention, memory_management
 from torch import einsum
 from einops import rearrange, repeat
-from modules import scripts
+from modules import scripts, shared
+from modules.ui_components import InputAccordion
 
 
 attn_precision = memory_management.force_upcast_attention_dtype()
@@ -99,7 +100,7 @@ def gaussian_blur_2d(img, kernel_size, sigma):
 
 
 class SelfAttentionGuidance:
-    def patch(self, model, scale, blur_sigma):
+    def patch(self, model, scale, blur_sigma, threshold):
         m = model.clone()
 
         attn_scores = None
@@ -129,7 +130,7 @@ class SelfAttentionGuidance:
 
             sag_scale = scale
             sag_sigma = blur_sigma
-            sag_threshold = 1.0
+            sag_threshold = threshold
             model = args["model"]
             uncond_pred = args["uncond_denoised"]
             uncond = args["uncond"]
@@ -163,35 +164,52 @@ class SAGForForge(scripts.Script):
     sorting_priority = 12.5
 
     def title(self):
-        return "SelfAttentionGuidance Integrated"
+        return "SelfAttentionGuidance Integrated (SD 1.x, SD 2.x, SDXL)"
 
     def show(self, is_img2img):
         return scripts.AlwaysVisible
 
     def ui(self, *args, **kwargs):
-        with gr.Accordion(open=False, label=self.title()):
+        with InputAccordion(False, label=self.title()) as enabled:
-            enabled = gr.Checkbox(label='Enabled', value=False)
             scale = gr.Slider(label='Scale', minimum=-2.0, maximum=5.0, step=0.01, value=0.5)
             blur_sigma = gr.Slider(label='Blur Sigma', minimum=0.0, maximum=10.0, step=0.01, value=2.0)
+            threshold = gr.Slider(label='Blur mask threshold', minimum=0.0, maximum=4.0, step=0.01, value=1.0)
 
-        return enabled, scale, blur_sigma
+        self.infotext_fields = [
+            (enabled, lambda d: d.get("sag_enabled", False)),
+            (scale,         "sag_scale"),
+            (blur_sigma,    "sag_blur_sigma"),
+            (threshold,     "sag_threshold"),
+        ]
+
+        return enabled, scale, blur_sigma, threshold
 
     def process_before_every_sampling(self, p, *script_args, **kwargs):
-        enabled, scale, blur_sigma = script_args
+        enabled, scale, blur_sigma, threshold = script_args
 
         if not enabled:
             return
 
+        #   not for FLux
+        if not shared.sd_model.is_webui_legacy_model():     #   ideally would be is_flux
+            gr.Info ("Self Attention Guidance is not compatible with Flux")
+            return
+        #   Self Attention Guidance errors if CFG is 1
+        if p.cfg_scale == 1:
+            gr.Info ("Self Attention Guidance requires CFG > 1")
+            return
+
         unet = p.sd_model.forge_objects.unet
 
-        unet = opSelfAttentionGuidance.patch(unet, scale, blur_sigma)[0]
+        unet = opSelfAttentionGuidance.patch(unet, scale, blur_sigma, threshold)[0]
 
         p.sd_model.forge_objects.unet = unet
 
         p.extra_generation_params.update(dict(
-            sag_enabled=enabled,
+            sag_enabled     = enabled,
-            sag_scale=scale,
+            sag_scale       = scale,
-            sag_blur_sigma=blur_sigma
+            sag_blur_sigma  = blur_sigma,
+            sag_threshold   = threshold,
         ))
 
         return