cleanup

Signed-off-by: Vladimir Mandic <mandic00@live.com>

installer.py

@@ -1010,7 +1010,6 @@ def ensure_base_requirements():
     setuptools_version = '69.5.1'
 
     def update_setuptools():
-        # print('Install base requirements')
         global pkg_resources, setuptools, distutils # pylint: disable=global-statement
         # python may ship with incompatible setuptools
         subprocess.run(f'"{sys.executable}" -m pip install setuptools=={setuptools_version}', shell=True, check=False, env=os.environ, stdout=subprocess.PIPE, stderr=subprocess.PIPE)

modules/instantir/ip_adapter/ip_adapter.py

@@ -169,8 +169,6 @@ class IPAdapterPlus(torch.nn.Module):
         if "latents" in state_dict["image_proj"] and "latents" in self.image_proj.state_dict():
             # Check if the shapes are mismatched
             if state_dict["image_proj"]["latents"].shape != self.image_proj.state_dict()["latents"].shape:
-                print(f"Shapes of 'image_proj.latents' in checkpoint {ckpt_path} and current model do not match.")
-                print("Removing 'latents' from checkpoint and loading the rest of the weights.")
                 del state_dict["image_proj"]["latents"]
                 strict_load_image_proj_model = False
 

modules/linfusion/linfusion.py

@@ -89,9 +89,7 @@ class LinFusion(ModelMixin, ConfigMixin):
                 pipe_name_path = pipe_name_path or pipeline._internal_dict._name_or_path # pylint: disable=protected-access
                 pretrained_model_name_or_path = model_dict.get(pipe_name_path, None)
                 if pretrained_model_name_or_path:
-                    print(
+                    pass
-                        f"Matching LinFusion '{pretrained_model_name_or_path}' for pipeline '{pipe_name_path}'."
-                    )
                 else:
                     raise RuntimeError(
                         f"LinFusion not found for pipeline [{pipe_name_path}], please provide the path."

modules/lora/lora_convert.py

@@ -205,8 +205,6 @@ def _convert_to_ai_toolkit_cat(sds_sd, ait_sd, sds_key, ait_keys, dims=None):
                     up_weight[i : i + dims[j], k * ait_rank : (k + 1) * ait_rank] == 0
                 )
             i += dims[j]
-        # if is_sparse:
-        #     print(f"weight is sparse: {sds_key}")
 
     # make ai-toolkit weight
     ait_down_keys = [k + ".lora_down.weight" for k in ait_keys]

modules/meissonic/transformer.py

@@ -670,15 +670,11 @@ class Simple_UVitBlock(nn.Module):
             self.upsample = None
 
     def forward(self, x):
-        # print("before,", x.shape)
         if self.downsample is not None:
-            # print('downsample')
             x = self.downsample(x)
 
         if self.upsample is not None:
-            # print('upsample')
             x = self.upsample(x)
-        # print("after,", x.shape)
         return x
 
 

modules/omnigen/model.py

@@ -259,7 +259,6 @@ class OmniGen(nn.Module, PeftAdapterMixin):
         left = (self.pos_embed_max_size - width) // 2
         spatial_pos_embed = self.pos_embed.reshape(1, self.pos_embed_max_size, self.pos_embed_max_size, -1)
         spatial_pos_embed = spatial_pos_embed[:, top : top + height, left : left + width, :]
-        # print(top, top + height, left, left + width, spatial_pos_embed.size())
         spatial_pos_embed = spatial_pos_embed.reshape(1, -1, spatial_pos_embed.shape[-1])
         return spatial_pos_embed
 

modules/pixelsmith/vae.py

@@ -681,10 +681,6 @@ class VectorQuantizer(nn.Module):
             if self.unknown_index == "extra":
                 self.unknown_index = self.re_embed
                 self.re_embed = self.re_embed + 1
-            print(
-                f"Remapping {self.n_e} indices to {self.re_embed} indices. "
-                f"Using {self.unknown_index} for unknown indices."
-            )
         else:
             self.re_embed = n_e
 

modules/pulid/eva_clip/hf_model.py

@@ -222,7 +222,6 @@ class HFTextEncoder(nn.Module):
 
         encoder = self.transformer.encoder if hasattr(self.transformer, 'encoder') else self.transformer
         layer_list = getattr(encoder, arch_dict[self.config.model_type]["config_names"]["layer_attr"])
-        print(f"Unlocking {unlocked_layers}/{len(layer_list) + 1} layers of hf model")
         embeddings = getattr(
             self.transformer, arch_dict[self.config.model_type]["config_names"]["token_embeddings_attr"])
         modules = [embeddings, *layer_list][:-unlocked_layers]

modules/schedulers/scheduler_dc.py

@@ -820,7 +820,6 @@ class DCSolverMultistepScheduler(SchedulerMixin, ConfigMixin):
             loss.backward()
             optimizer.step()
             ratio_bound = bound_func(ratio_param)
-            print(f'iter [{iter_}]', ratio_bound.item(), loss.item())
 
         torch.cuda.empty_cache()
         return ratio_bound.data.detach().item()

modules/schedulers/scheduler_tdd.py

@@ -117,8 +117,7 @@ class TDDScheduler(DPMSolverSinglestepScheduler):
         timesteps = tcd_origin_timesteps[inference_indices]
         if self.special_jump:
             if self.tdd_train_step == 50:
-                #timesteps = np.array([999., 879., 759., 499., 259.])
+                pass
-                print(timesteps)
             elif self.tdd_train_step == 250:
                 if num_inference_steps == 5:
                     timesteps = np.array([999., 875., 751., 499., 251.])
@@ -203,7 +202,6 @@ class TDDScheduler(DPMSolverSinglestepScheduler):
 
         sigmas_s = np.array(((1 - self.alphas_cumprod) / self.alphas_cumprod) ** 0.5)
         if self.config.use_karras_sigmas:
-            print("have not write")
             pass
         else:
             sigmas_s = np.interp(timesteps_s, np.arange(0, len(sigmas_s)), sigmas_s)

modules/schedulers/scheduler_vdm.py

@@ -355,7 +355,6 @@ class VDMScheduler(SchedulerMixin, ConfigMixin):
             )
 
         # 3. Clip or threshold "predicted x_0"
-        # print({ 'timestep': timestep.item(), 'min': pred_original_sample.min().item(), 'max': pred_original_sample.max().item(), 'alpha': alpha.item(), 'sigma': sigma.item() })
         if self.config.thresholding:
             pred_original_sample = self._threshold_sample(pred_original_sample)
         elif self.config.clip_sample:

modules/todo/todo_merge.py

@@ -25,7 +25,6 @@ def init_generator(device: torch.device, fallback: torch.Generator = None):
     """
     Forks the current default random generator given device.
     """
-    print(f"init_generator device = {device}")
     if device.type == "cpu":
         return torch.Generator(device="cpu").set_state(torch.get_rng_state())
     elif device.type == "cuda":