Update gligen/ldm/models/diffusion/plms.py

gligen/ldm/models/diffusion/plms.py

@@ -5,7 +5,7 @@ from functools import partial
 from copy import deepcopy
 from ldm.modules.diffusionmodules.util import make_ddim_sampling_parameters, make_ddim_timesteps, noise_like
 import math
-from ldm.models.diffusion.loss import  caculate_loss_att_fixed_cnt, caculate_loss_self_att, caculate_loss_LoCo_V2
+from ldm.models.diffusion.loss import  caculate_loss_att_fixed_cnt, caculate_loss_self_att
 class PLMSSampler(object):
     def __init__(self, diffusion, model, schedule="linear", alpha_generator_func=None, set_alpha_scale=None):
         super().__init__()
@@ -57,14 +57,14 @@ class PLMSSampler(object):
 
 
     # @torch.no_grad()
-    def sample(self, S, shape, input, uc=None, guidance_scale=1, mask=None, x0=None, loss_type='LoCo'):
+    def sample(self, S, shape, input, uc=None, guidance_scale=1, mask=None, x0=None, loss_type='SAR_CAR'):
         self.make_schedule(ddim_num_steps=S)
         # import pdb; pdb.set_trace()
         return self.plms_sampling(shape, input, uc, guidance_scale, mask=mask, x0=x0, loss_type=loss_type)
 
 
     # @torch.no_grad()
-    def plms_sampling(self, shape, input, uc=None, guidance_scale=1, mask=None, x0=None, loss_type='LoCo'):
+    def plms_sampling(self, shape, input, uc=None, guidance_scale=1, mask=None, x0=None, loss_type='SAR_CAR'):
 
         b = shape[0]
         
@@ -102,16 +102,12 @@ class PLMSSampler(object):
             # three loss types
             if loss_type !=None and loss_type!='standard':
                 if input['object_position'] != []:
-                    # if loss_type=='SAR_CAR':
-                    #     x = self.update_loss_self_cross( input,i, index, ts )
-                    # elif loss_type=='SAR':
-                    #     x = self.update_only_self( input,i, index, ts )
-                    # elif loss_type=='CAR':
-                    #     x = self.update_loss_only_cross( input,i, index, ts )
-                    # elif loss_type=='LoCo':
-
-                    x = self.update_loss_only_cross( input,i, index, ts )
-                    # x = self.update_loss_LoCo( input,i, index, ts, )   
+                    if loss_type=='SAR_CAR':
+                        x = self.update_loss_self_cross( input,i, index, ts )
+                    elif loss_type=='SAR':
+                        x = self.update_only_self( input,i, index, ts )
+                    elif loss_type=='CAR':
+                        x = self.update_loss_only_cross( input,i, index, ts )
                     input["x"] = x
             img, pred_x0, e_t = self.p_sample_plms(input, ts, index=index, uc=uc, guidance_scale=guidance_scale, old_eps=old_eps, t_next=ts_next)
             input["x"] = img
@@ -120,55 +116,7 @@ class PLMSSampler(object):
                 old_eps.pop(0)
 
         return img
-
-    def update_loss_LoCo(self, input,index1, index, ts, type_loss='self_accross'):
-        
-        # loss_scale = 30
-        # max_iter = 5
-        #print('time_factor is: ', time_factor)
-        if index1 < 10:
-            loss_scale = 8
-            max_iter = 5
-        elif index1 < 20:
-            loss_scale = 5
-            max_iter = 5
-        else:
-            loss_scale = 1
-            max_iter = 1
-        loss_threshold = 0.1
-        
-        max_index = 30
-        x = deepcopy(input["x"]) 
-        iteration = 0
-        loss = torch.tensor(10000)
-        input["timesteps"] = ts
-        
-        # print("optimize", index1)
-        while loss.item() > loss_threshold and iteration < max_iter and (index1 < max_index) :
-            # print('iter', iteration)
-            x = x.requires_grad_(True)
-            # print('x shape', x.shape)
-            input['x'] = x
-            e_t,  att_first, att_second, att_third, self_first, self_second, self_third = self.model(input)
-            
-            bboxes = input['boxes']
-            object_positions = input['object_position'] 
-            loss2 = caculate_loss_LoCo_V2(att_second,att_first,att_third, bboxes=bboxes,
-                        object_positions=object_positions, t = index1)*loss_scale
-            # loss = loss2
-            # loss.requires_grad_(True)
-            #print('LoCo loss', loss) 
-
-
-            
-            hh = torch.autograd.backward(loss2,  retain_graph=True)
-            grad_cond = x.grad
-            x = x - grad_cond 
-            x = x.detach()
-            iteration += 1
-            torch.cuda.empty_cache()
-        return x
-        
+     
     def update_loss_self_cross(self, input,index1, index, ts,type_loss='self_accross' ):
         if index1 < 10:
             loss_scale = 4