RAM optimizations and faster launch

README.md

@@ -255,8 +255,9 @@ You can define multiple lines of macros. If there is only one macro line, the ap
 --check-loras : filter loras that are incompatible (will take a few seconds while refreshing the lora list or while starting the app)\
 --advanced : turn on the advanced mode while launching the app\
 --i2v-settings : path to launch settings for i2v\
---t2v-settings : path to launch settings for t2v
---listen : make server accessible on network
+--t2v-settings : path to launch settings for t2v\
+--listen : make server accessible on network\
+--gpu device : run Wan on device for instance "cuda:1"
 
 ### Profiles (for power users only)
 You can choose between 5 profiles, but two are really relevant here :

gradio_server.py

@@ -14,7 +14,7 @@ import json
 import wan
 from wan.configs import MAX_AREA_CONFIGS, WAN_CONFIGS, SUPPORTED_SIZES
 from wan.utils.utils import cache_video
-from wan.modules.attention import get_attention_modes
+from wan.modules.attention import get_attention_modes, get_supported_attention_modes
 import torch
 import gc
 import traceback
@@ -23,7 +23,7 @@ import asyncio
 from wan.utils import prompt_parser
 PROMPT_VARS_MAX = 10
 
-target_mmgp_version = "3.3.1"
+target_mmgp_version = "3.3.2"
 from importlib.metadata import version
 mmgp_version = version("mmgp")
 if mmgp_version != target_mmgp_version:
@@ -169,6 +169,12 @@ def _parse_args():
         default="",
         help="Server name"
     )
+    parser.add_argument(
+        "--gpu",
+        type=str,
+        default="",
+        help="Default GPU Device"
+    )
 
     parser.add_argument(
         "--open-browser",
@@ -262,12 +268,11 @@ def get_lora_dir(i2v):
                 return lora_dir_14B
     return root_lora_dir    
 
-attention_modes_supported = get_attention_modes()
-
+attention_modes_installed = get_attention_modes()
+attention_modes_supported = get_supported_attention_modes()
 args = _parse_args()
 args.flow_reverse = True
 
-
 lock_ui_attention = False
 lock_ui_transformer = False
 lock_ui_compile = False
@@ -375,6 +380,7 @@ vae_config = server_config.get("vae_config", 0)
 if len(args.vae_config) > 0:
     vae_config = int(args.vae_config)
 
+reload_needed = False
 default_ui = server_config.get("default_ui", "t2v") 
 metadata = server_config.get("metadata_type", "metadata")
 save_path = server_config.get("save_path", os.path.join(os.getcwd(), "gradio_outputs"))
@@ -474,7 +480,7 @@ def download_models(transformer_filename, text_encoder_filename):
     from huggingface_hub import hf_hub_download, snapshot_download    
     repoId = "DeepBeepMeep/Wan2.1" 
     sourceFolderList = ["xlm-roberta-large", "",  ]
-    fileList = [ [], ["Wan2.1_VAE.pth", "models_clip_open-clip-xlm-roberta-large-vit-huge-14.pth" ] + computeList(text_encoder_filename) + computeList(transformer_filename) ]   
+    fileList = [ [], ["Wan2.1_VAE_bf16.safetensors", "models_clip_open-clip-xlm-roberta-large-vit-huge-14-bf16.safetensors" ] + computeList(text_encoder_filename) + computeList(transformer_filename) ]   
     targetRoot = "ckpts/" 
     for sourceFolder, files in zip(sourceFolderList,fileList ):
         if len(files)==0:
@@ -491,6 +497,14 @@ def download_models(transformer_filename, text_encoder_filename):
 
 
 offload.default_verboseLevel = verbose_level
+to_remove = ["models_clip_open-clip-xlm-roberta-large-vit-huge-14.pth", "Wan2.1_VAE.pth"]
+for file_name in to_remove:
+    file_name = os.path.join("ckpts",file_name)
+    if os.path.isfile(file_name):
+        try:
+            os.remove(file_name)
+        except:
+            pass
 
 download_models(transformer_filename_i2v if use_image2video else transformer_filename_t2v, text_encoder_filename) 
 
@@ -663,6 +677,8 @@ def load_models(i2v):
     elif profile == 3:
         kwargs["budgets"] = { "*" : "70%" }
     offloadobj = offload.profile(pipe, profile_no= profile, compile = compile, quantizeTransformer = quantizeTransformer, loras = "transformer", **kwargs)  
+    if len(args.gpu) > 0:
+        torch.set_default_device(args.gpu)
 
     return wan_model, offloadobj, pipe["transformer"] 
 
@@ -702,8 +718,10 @@ def generate_header(model_filename, compile, attention_mode):
 
     header += model_name 
     header += " (attention mode: " + (attention_mode if attention_mode!="auto" else "auto/" + get_auto_attention() )
-    if attention_mode not in attention_modes_supported:
+    if attention_mode not in attention_modes_installed:
         header += " -NOT INSTALLED-"
+    elif attention_mode not in attention_modes_supported:
+        header += " -NOT SUPPORTED-"
 
     if compile:
         header += ", pytorch compilation ON"
@@ -765,11 +783,7 @@ def apply_changes(  state,
         if v != v_old:
             changes.append(k)
 
-    state["config_changes"] = changes
-    state["config_new"] = server_config
-    state["config_old"] = old_server_config
-
-    global attention_mode, profile, compile, transformer_filename_t2v, transformer_filename_i2v, text_encoder_filename, vae_config, boost, lora_dir
+    global attention_mode, profile, compile, transformer_filename_t2v, transformer_filename_i2v, text_encoder_filename, vae_config, boost, lora_dir, reload_needed
     attention_mode = server_config["attention_mode"]
     profile = server_config["profile"]
     compile = server_config["compile"]
@@ -781,7 +795,7 @@ def apply_changes(  state,
     if  all(change in ["attention_mode", "vae_config", "default_ui", "boost", "save_path", "metadata_choice"] for change in changes ):
         pass
     else:
-        state["_reload_needed"] = True
+        reload_needed = True
 
 
     yield "<DIV ALIGN=CENTER>The new configuration has been succesfully applied</DIV>"
@@ -931,8 +945,7 @@ def generate_video(
 
 ):
 
-    global wan_model, offloadobj
-    reload_needed = state.get("_reload_needed", False)
+    global wan_model, offloadobj, reload_needed
     file_model_needed = model_needed(image2video)
     if file_model_needed !=  model_filename or reload_needed:
         del wan_model
@@ -943,7 +956,7 @@ def generate_video(
         yield f"Loading model {get_model_name(file_model_needed)}..."
         wan_model, offloadobj, trans = load_models(image2video)
         yield f"Model loaded"
-        state["_reload_needed"] =  False
+        reload_needed=  False
 
     from PIL import Image
     import numpy as np
@@ -956,11 +969,12 @@ def generate_video(
     elif attention_mode in attention_modes_supported:
         attn = attention_mode
     else:
-        gr.Info(f"You have selected attention mode '{attention_mode}'. However it is not installed on your system. You should either install it or switch to the default 'sdpa' attention.")
+        gr.Info(f"You have selected attention mode '{attention_mode}'. However it is not installed or supported on your system. You should either install it or switch to the default 'sdpa' attention.")
         return
 
     if state.get("validate_success",0) != 1:
         return
+    raw_resolution = resolution
     width, height = resolution.split("x")
     width, height = int(width), int(height)
 
@@ -1310,15 +1324,9 @@ def generate_video(
                     nrow=1,
                     normalize=True,
                     value_range=(-1, 1))
-                
-                configs = {
-                    'prompt': prompt,
-                    'negative_prompt': negative_prompt,
-                    'resolution': resolution,
-                    'video_length': video_length,
-                    'seed': seed,
-                    'num_inference_steps': num_inference_steps,
-                }
+
+                configs = get_settings_dict(state, use_image2video, prompt, 0 if image_to_end == None else 1 , video_length, raw_resolution, num_inference_steps, seed, repeat_generation, multi_images_gen_type, guidance_scale, flow_shift, negative_prompt, loras_choices, 
+                      loras_mult_choices, tea_cache , tea_cache_start_step_perc, RIFLEx_setting, slg_switch, slg_layers, slg_start, slg_end)
 
                 metadata_choice = server_config.get("metadata_choice","metadata")
                 if metadata_choice == "json":
@@ -1632,19 +1640,15 @@ def switch_advanced(state, new_advanced, lset_name):
     else:
         return  gr.Row(visible=new_advanced), gr.Row(visible=True), gr.Button(visible=True), gr.Row(visible= False), gr.Dropdown(choices=lset_choices, value= lset_name)
 
-def save_settings(state, prompt, image_prompt_type, video_length, resolution, num_inference_steps, seed, repeat_generation, multi_images_gen_type, guidance_scale, flow_shift, negative_prompt, loras_choices, 
+
+def get_settings_dict(state, i2v, prompt, image_prompt_type, video_length, resolution, num_inference_steps, seed, repeat_generation, multi_images_gen_type, guidance_scale, flow_shift, negative_prompt, loras_choices, 
                       loras_mult_choices, tea_cache_setting, tea_cache_start_step_perc, RIFLEx_setting, slg_switch, slg_layers, slg_start_perc, slg_end_perc):
 
-    if state.get("validate_success",0) != 1:
-        return
-
-    loras_choices
     loras = state["loras"]
     activated_loras = [Path( loras[int(no)]).parts[-1]  for no in loras_choices ]
 
-    ui_defaults = {
+    ui_settings = {
         "prompts": prompt,
-        "image_prompt_type": image_prompt_type,
         "resolution": resolution,
         "video_length": video_length,
         "num_inference_steps": num_inference_steps,
@@ -1664,6 +1668,24 @@ def save_settings(state, prompt, image_prompt_type, video_length, resolution, nu
         "slg_start_perc": slg_start_perc,
         "slg_end_perc": slg_end_perc
     }
+
+    if i2v:
+        ui_settings["type"] = "Wan2.1GP by DeepBeepMeep - image2video"
+        ui_settings["image_prompt_type"] = image_prompt_type,
+    else: 
+        ui_settings["type"] = "Wan2.1GP by DeepBeepMeep - text2video"
+
+    return ui_settings
+
+def save_settings(state, prompt, image_prompt_type, video_length, resolution, num_inference_steps, seed, repeat_generation, multi_images_gen_type, guidance_scale, flow_shift, negative_prompt, loras_choices, 
+                      loras_mult_choices, tea_cache_setting, tea_cache_start_step_perc, RIFLEx_setting, slg_switch, slg_layers, slg_start_perc, slg_end_perc):
+
+    if state.get("validate_success",0) != 1:
+        return
+
+    ui_defaults = get_settings_dict(state, use_image2video, prompt, image_prompt_type, video_length, resolution, num_inference_steps, seed, repeat_generation, multi_images_gen_type, guidance_scale, flow_shift, negative_prompt, loras_choices, 
+                      loras_mult_choices, tea_cache_setting, tea_cache_start_step_perc, RIFLEx_setting, slg_switch, slg_layers, slg_start_perc, slg_end_perc)
+
     defaults_filename = get_settings_file_name(use_image2video)
 
     with open(defaults_filename, "w", encoding="utf-8") as f:
@@ -1780,8 +1802,8 @@ def generate_video_tab(image2video=False):
                         delete_lset_btn = gr.Button("Delete", size="sm", min_width= 1)
                         cancel_lset_btn = gr.Button("Don't do it !", size="sm", min_width= 1 , visible=False)  
             video_to_continue = gr.Video(label= "Video to continue", visible= image2video and False) #######
-
-            image_prompt_type = gr.Radio( [("Use only a Start Image", 0),("Use both a Start and an End Image", 1)], value =ui_defaults["image_prompt_type"], label="Location", show_label= False, scale= 3, visible=image2video)
+            image_prompt_type= ui_defaults.get("image_prompt_type",0)
+            image_prompt_type_radio = gr.Radio( [("Use only a Start Image", 0),("Use both a Start and an End Image", 1)], value =image_prompt_type, label="Location", show_label= False, scale= 3, visible=image2video)
 
             if args.multiple_images:  
                 image_to_continue = gr.Gallery(
@@ -1793,9 +1815,9 @@ def generate_video_tab(image2video=False):
             if args.multiple_images:  
                 image_to_end  = gr.Gallery(
                         label="Images as ending points for new videos", type ="pil", #file_types= "image", 
-                        columns=[3], rows=[1], object_fit="contain", height="auto", selected_index=0, interactive= True, visible=False)
+                        columns=[3], rows=[1], object_fit="contain", height="auto", selected_index=0, interactive= True, visible=image_prompt_type==1)
             else:
-                image_to_end = gr.Image(label= "Last Image for a new video", type ="pil", visible= False)
+                image_to_end = gr.Image(label= "Last Image for a new video", type ="pil", visible=image_prompt_type==1)
 
             def switch_image_prompt_type_radio(image_prompt_type_radio):
                 if args.multiple_images:
@@ -1803,7 +1825,7 @@ def generate_video_tab(image2video=False):
                 else:
                     return gr.Image(visible = (image_prompt_type_radio == 1)  )
 
-            image_prompt_type.change(fn=switch_image_prompt_type_radio, inputs=[image_prompt_type], outputs=[image_to_end]) 
+            image_prompt_type_radio.change(fn=switch_image_prompt_type_radio, inputs=[image_prompt_type_radio], outputs=[image_to_end]) 
 
 
             advanced_prompt = advanced
@@ -1973,7 +1995,7 @@ def generate_video_tab(image2video=False):
 
 
         save_settings_btn.click( fn=validate_wizard_prompt, inputs =[state, wizard_prompt_activated_var, wizard_variables_var,  prompt, wizard_prompt, *prompt_vars] , outputs= [prompt]).then(
-            save_settings, inputs = [state, prompt, image_prompt_type, video_length, resolution, num_inference_steps, seed, repeat_generation, multi_images_gen_type, guidance_scale, flow_shift, negative_prompt, 
+            save_settings, inputs = [state, prompt, image_prompt_type_radio, video_length, resolution, num_inference_steps, seed, repeat_generation, multi_images_gen_type, guidance_scale, flow_shift, negative_prompt, 
                                                          loras_choices, loras_mult_choices, tea_cache_setting, tea_cache_start_step_perc, RIFLEx_setting, slg_switch, slg_layers,
                                                          slg_start_perc, slg_end_perc  ], outputs = [])
         save_lset_btn.click(validate_save_lset, inputs=[lset_name], outputs=[apply_lset_btn, refresh_lora_btn, delete_lset_btn, save_lset_btn,confirm_save_lset_btn, cancel_lset_btn, save_lset_prompt_drop])
@@ -2080,8 +2102,10 @@ def generate_configuration_tab():
             value=server_config.get("save_path", save_path)
         )
         def check(mode): 
-            if not mode in attention_modes_supported:
+            if not mode in attention_modes_installed:
                 return " (NOT INSTALLED)"
+            elif not mode in attention_modes_supported:
+                return " (NOT SUPPORTED)"
             else:
                 return ""
         attention_choice = gr.Dropdown(
@@ -2238,7 +2262,7 @@ def create_demo():
         }
     """
     with gr.Blocks(css=css, theme=gr.themes.Soft(primary_hue="sky", neutral_hue="slate", text_size="md")) as demo:
-        gr.Markdown("<div align=center><H1>Wan 2.1<SUP>GP</SUP> v3.0 <FONT SIZE=4>by <I>DeepBeepMeep</I></FONT> <FONT SIZE=3> (<A HREF='https://github.com/deepbeepmeep/Wan2GP'>Updates</A>)</FONT SIZE=3></H1></div>")
+        gr.Markdown("<div align=center><H1>Wan 2.1<SUP>GP</SUP> v3.1 <FONT SIZE=4>by <I>DeepBeepMeep</I></FONT> <FONT SIZE=3> (<A HREF='https://github.com/deepbeepmeep/Wan2GP'>Updates</A>)</FONT SIZE=3></H1></div>")
         gr.Markdown("<FONT SIZE=3>Welcome to Wan 2.1GP a super fast and low VRAM AI Video Generator !</FONT>")
         
         with gr.Accordion("Click here for some Info on how to use Wan2GP", open = False):

requirements.txt

@@ -16,6 +16,6 @@ gradio>=5.0.0
 numpy>=1.23.5,<2
 einops
 moviepy==1.0.3
-mmgp==3.3.1
+mmgp==3.3.2
 peft==0.14.0
 mutagen

wan/image2video.py

@@ -177,7 +177,7 @@ class WanI2V:
         logging.info(f"Creating WanModel from {model_filename}")
         from mmgp import offload
 
-        self.model = offload.fast_load_transformers_model(model_filename, modelClass=WanModel)
+        self.model = offload.fast_load_transformers_model(model_filename, modelClass=WanModel, writable_tensors= False)
         self.model.eval().requires_grad_(False)
 
         if t5_fsdp or dit_fsdp or use_usp:

wan/modules/attention.py

@@ -30,6 +30,7 @@ try:
             max_seqlen_kv,
         ):
         return sageattn_varlen(q, k, v, cu_seqlens_q, cu_seqlens_kv, max_seqlen_q, max_seqlen_kv)
+    
 except ImportError:
     sageattn_varlen_wrapper = None
 
@@ -38,11 +39,12 @@ import warnings
 
 try:
     from sageattention import sageattn
-    from .sage2_core import sageattn as alt_sageattn
+    from .sage2_core import sageattn as alt_sageattn, is_sage_supported
+    sage_supported =  is_sage_supported()
 except ImportError:
     sageattn = None
     alt_sageattn = None
-
+    sage_supported = False
 # @torch.compiler.disable()
 def sageattn_wrapper(
         qkv_list,
@@ -129,6 +131,14 @@ def get_attention_modes():
 
     return ret
 
+def get_supported_attention_modes():
+    ret = get_attention_modes()
+    if not sage_supported:
+        if "sage" in ret:
+            del ret["sage"]
+        if "sage2" in ret:
+            del ret["sage2"]
+    return ret
 
 __all__ = [
     'pay_attention',

wan/modules/clip.py

@@ -519,8 +519,11 @@ class CLIPModel:
                 device=device)
         self.model = self.model.eval().requires_grad_(False)
         logging.info(f'loading {checkpoint_path}')
-        self.model.load_state_dict(
-            torch.load(checkpoint_path, map_location='cpu'), assign= True)
+        from mmgp import offload
+        # self.model.load_state_dict(
+        #     torch.load(checkpoint_path, map_location='cpu'), assign= True)
+
+        offload.load_model_data(self.model, checkpoint_path.replace(".pth", "-bf16.safetensors"), writable_tensors= False)
 
         # init tokenizer
         self.tokenizer = HuggingfaceTokenizer(

wan/modules/sage2_core.py

@@ -51,6 +51,15 @@ from sageattention.quant import per_channel_fp8
 
 from typing import Any, List, Literal, Optional, Tuple, Union
 import warnings
+import os
+
+def is_sage_supported():
+    device_count = torch.cuda.device_count()
+    for i in range(device_count):
+        major, minor = torch.cuda.get_device_capability(i)
+        if major < 8:
+            return False
+    return True
 
 def get_cuda_arch_versions():
     cuda_archs = []

wan/modules/t5.py

@@ -496,7 +496,7 @@ class T5EncoderModel:
                 device=device).eval().requires_grad_(False)
         logging.info(f'loading {checkpoint_path}')
         from mmgp import offload
-        offload.load_model_data(model,checkpoint_path )
+        offload.load_model_data(model,checkpoint_path, writable_tensors= False )
 
         self.model = model
         if shard_fn is not None:

wan/modules/vae.py

@@ -744,11 +744,12 @@ def _video_vae(pretrained_path=None, z_dim=None, device='cpu', **kwargs):
     with torch.device('meta'):
         model = WanVAE_(**cfg)
 
+    from mmgp import offload
     # load checkpoint
     logging.info(f'loading {pretrained_path}')
-    model.load_state_dict(
-        torch.load(pretrained_path, map_location=device), assign=True)
-
+    # model.load_state_dict(
+    #     torch.load(pretrained_path, map_location=device), assign=True)
+    offload.load_model_data(model, pretrained_path.replace(".pth", "_bf16.safetensors"), writable_tensors= False)    
     return model
 
 
@@ -778,7 +779,7 @@ class WanVAE:
         self.model = _video_vae(
             pretrained_path=vae_pth,
             z_dim=z_dim,
-        ).eval().requires_grad_(False).to(device)
+        ).eval() #.requires_grad_(False).to(device)
 
     def encode(self, videos, tile_size = 256, any_end_frame = False):
         """

wan/text2video.py

@@ -89,7 +89,7 @@ class WanT2V:
         from mmgp import offload
 
 
-        self.model = offload.fast_load_transformers_model(model_filename, modelClass=WanModel)
+        self.model = offload.fast_load_transformers_model(model_filename, modelClass=WanModel, writable_tensors= False)