Make --novram completely offload weights.

This flag is mainly used for testing the weight offloading, it shouldn't
actually be used in practice.

Remove useless import.

Author: comfyanonymous

comfy/ldm/pixart/blocks.py

@@ -6,7 +6,6 @@
 import torch.nn.functional as F
 from einops import rearrange
 
-from comfy import model_management
 from comfy.ldm.modules.diffusionmodules.mmdit import TimestepEmbedder, Mlp, timestep_embedding
 from comfy.ldm.modules.attention import optimized_attention
 

comfy/model_management.py

@@ -521,7 +521,7 @@ def load_models_gpu(models, memory_required=0, force_patch_weights=False, minimu
                 lowvram_model_memory = 0
 
         if vram_set_state == VRAMState.NO_VRAM:
-            lowvram_model_memory = 64 * 1024 * 1024
+            lowvram_model_memory = 0.1
 
         loaded_model.model_load(lowvram_model_memory, force_patch_weights=force_patch_weights)
         current_loaded_models.insert(0, loaded_model)