[Fix] Fix A14B load quant model bug and tae bug (#277)

* [Fix] Fix A14B load quant model bug and tae bug

* [Fix] Fix A14B load quant model bug and tae bug

lightx2v/models/networks/hunyuan/model.py

@@ -53,7 +53,7 @@ class HunyuanModel:
         return weight_dict
 
     def _load_quant_ckpt(self):
-        ckpt_path = self.config.dit_quantized_ckpt
+        ckpt_path = self.dit_quantized_ckpt
         logger.info(f"Loading quant dit model from {ckpt_path}")
 
         if ckpt_path.endswith(".pth"):

lightx2v/models/networks/wan/model.py

@@ -80,8 +80,6 @@ class WanModel:
             self.dit_quantized_ckpt = None
             assert not self.config.get("lazy_load", False)
 
-        self.config.dit_quantized_ckpt = self.dit_quantized_ckpt
-
         self.weight_auto_quant = self.config.mm_config.get("weight_auto_quant", False)
         if self.dit_quantized:
             assert self.weight_auto_quant or self.dit_quantized_ckpt is not None

lightx2v/models/runners/wan/wan_runner.py

@@ -386,20 +386,22 @@ class MultiModelStruct:
         if self.scheduler.timesteps[self.scheduler.step_index] >= self.boundary_timestep:
             logger.info(f"using - HIGH - noise model at step_index {self.scheduler.step_index + 1}")
             self.scheduler.sample_guide_scale = self.config.sample_guide_scale[0]
-            if self.cur_model_index == -1:
-                self.to_cuda(model_index=0)
-            elif self.cur_model_index == 1:  # 1 -> 0
-                self.offload_cpu(model_index=1)
-                self.to_cuda(model_index=0)
+            if self.config.get("cpu_offload", False) and self.config.get("offload_granularity", "block") == "model":
+                if self.cur_model_index == -1:
+                    self.to_cuda(model_index=0)
+                elif self.cur_model_index == 1:  # 1 -> 0
+                    self.offload_cpu(model_index=1)
+                    self.to_cuda(model_index=0)
             self.cur_model_index = 0
         else:
             logger.info(f"using - LOW - noise model at step_index {self.scheduler.step_index + 1}")
             self.scheduler.sample_guide_scale = self.config.sample_guide_scale[1]
-            if self.cur_model_index == -1:
-                self.to_cuda(model_index=1)
-            elif self.cur_model_index == 0:  # 0 -> 1
-                self.offload_cpu(model_index=0)
-                self.to_cuda(model_index=1)
+            if self.config.get("cpu_offload", False) and self.config.get("offload_granularity", "block") == "model":
+                if self.cur_model_index == -1:
+                    self.to_cuda(model_index=1)
+                elif self.cur_model_index == 0:  # 0 -> 1
+                    self.offload_cpu(model_index=0)
+                    self.to_cuda(model_index=1)
             self.cur_model_index = 1
 
     def offload_cpu(self, model_index):

lightx2v/models/video_encoders/hf/tae.py

@@ -158,7 +158,7 @@ def apply_model_with_memblocks(model, x, parallel, show_progress_bar):
 
 
 class TAEHV(nn.Module):
-    def __init__(self, checkpoint_path="taehv.pth", decoder_time_upscale=(True, True), decoder_space_upscale=(True, True, True), patch_size=1, latent_channels=16):
+    def __init__(self, checkpoint_path="taehv.pth", decoder_time_upscale=(True, True), decoder_space_upscale=(True, True, True), patch_size=1, latent_channels=16, model_type="wan21"):
         """Initialize pretrained TAEHV from the given checkpoint.
 
         Arg:
@@ -173,7 +173,10 @@ class TAEHV(nn.Module):
         self.latent_channels = latent_channels
         self.image_channels = 3
         self.is_cogvideox = checkpoint_path is not None and "taecvx" in checkpoint_path
-        if checkpoint_path is not None and "taew2_2" in checkpoint_path:
+        # if checkpoint_path is not None and "taew2_2" in checkpoint_path:
+        #     self.patch_size, self.latent_channels = 2, 48
+
+        if model_type == "wan22":
             self.patch_size, self.latent_channels = 2, 48
         self.encoder = nn.Sequential(
             conv(self.image_channels * self.patch_size**2, 64),

lightx2v/models/video_encoders/hf/wan/vae_2_2.py

@@ -841,7 +841,9 @@ def _video_vae(pretrained_path=None, z_dim=16, dim=160, device="cpu", cpu_offloa
 
 
 class Wan2_2_VAE:
-    def __init__(self, z_dim=48, c_dim=160, vae_pth=None, dim_mult=[1, 2, 4, 4], temperal_downsample=[False, True, True], dtype=torch.float, device="cuda", cpu_offload=False, offload_cache=False):
+    def __init__(
+        self, z_dim=48, c_dim=160, vae_pth=None, dim_mult=[1, 2, 4, 4], temperal_downsample=[False, True, True], dtype=torch.float, device="cuda", cpu_offload=False, offload_cache=False, **kwargs
+    ):
         self.dtype = dtype
         self.device = device
         self.cpu_offload = cpu_offload

lightx2v/models/video_encoders/hf/wan/vae_tiny.py

@@ -80,7 +80,7 @@ class Wan2_2_VAE_tiny(nn.Module):
         super().__init__()
         self.dtype = dtype
         self.device = torch.device("cuda")
-        self.taehv = TAEHV(vae_pth).to(self.dtype)
+        self.taehv = TAEHV(vae_pth, model_type="wan22").to(self.dtype)
         self.need_scaled = need_scaled
         if self.need_scaled:
             self.latents_mean = [