support s2v framepack

Author: mi804

README.md

@@ -205,7 +205,7 @@ save_video(video, "video1.mp4", fps=15, quality=5)
 
 | Model ID | Extra Parameters | Inference | Full Training | Validate After Full Training | LoRA Training | Validate After LoRA Training |
 |-|-|-|-|-|-|-|
-|[Wan-AI/Wan2.2-S2V-14B](https://www.modelscope.cn/models/Wan-AI/Wan2.2-S2V-14B)|`input_image`, `input_audio`, `audio_sample_rate`, `s2v_pose_video`|[code](./examples/wanvideo/model_inference/Wan2.2-S2V-14B.py)|-|-|-|-|
+|[Wan-AI/Wan2.2-S2V-14B](https://www.modelscope.cn/models/Wan-AI/Wan2.2-S2V-14B)|`input_image`, `input_audio`, `audio_sample_rate`, `s2v_pose_video`|[code](./examples/wanvideo/model_inference/Wan2.2-S2V-14B_multi_clips.py)|-|-|-|-|
 |[Wan-AI/Wan2.2-I2V-A14B](https://modelscope.cn/models/Wan-AI/Wan2.2-I2V-A14B)|`input_image`|[code](./examples/wanvideo/model_inference/Wan2.2-I2V-A14B.py)|[code](./examples/wanvideo/model_training/full/Wan2.2-I2V-A14B.sh)|[code](./examples/wanvideo/model_training/validate_full/Wan2.2-I2V-A14B.py)|[code](./examples/wanvideo/model_training/lora/Wan2.2-I2V-A14B.sh)|[code](./examples/wanvideo/model_training/validate_lora/Wan2.2-I2V-A14B.py)|
 |[Wan-AI/Wan2.2-T2V-A14B](https://modelscope.cn/models/Wan-AI/Wan2.2-T2V-A14B)||[code](./examples/wanvideo/model_inference/Wan2.2-T2V-A14B.py)|[code](./examples/wanvideo/model_training/full/Wan2.2-T2V-A14B.sh)|[code](./examples/wanvideo/model_training/validate_full/Wan2.2-T2V-A14B.py)|[code](./examples/wanvideo/model_training/lora/Wan2.2-T2V-A14B.sh)|[code](./examples/wanvideo/model_training/validate_lora/Wan2.2-T2V-A14B.py)|
 |[Wan-AI/Wan2.2-TI2V-5B](https://modelscope.cn/models/Wan-AI/Wan2.2-TI2V-5B)|`input_image`|[code](./examples/wanvideo/model_inference/Wan2.2-TI2V-5B.py)|[code](./examples/wanvideo/model_training/full/Wan2.2-TI2V-5B.sh)|[code](./examples/wanvideo/model_training/validate_full/Wan2.2-TI2V-5B.py)|[code](./examples/wanvideo/model_training/lora/Wan2.2-TI2V-5B.sh)|[code](./examples/wanvideo/model_training/validate_lora/Wan2.2-TI2V-5B.py)|

README_zh.md

@@ -205,7 +205,7 @@ save_video(video, "video1.mp4", fps=15, quality=5)
 
 |模型 ID|额外参数|推理|全量训练|全量训练后验证|LoRA 训练|LoRA 训练后验证|
 |-|-|-|-|-|-|-|
-|[Wan-AI/Wan2.2-S2V-14B](https://www.modelscope.cn/models/Wan-AI/Wan2.2-S2V-14B)|`input_image`, `input_audio`, `audio_sample_rate`, `s2v_pose_video`|[code](./examples/wanvideo/model_inference/Wan2.2-S2V-14B.py)|-|-|-|-|
+|[Wan-AI/Wan2.2-S2V-14B](https://www.modelscope.cn/models/Wan-AI/Wan2.2-S2V-14B)|`input_image`, `input_audio`, `audio_sample_rate`, `s2v_pose_video`|[code](./examples/wanvideo/model_inference/Wan2.2-S2V-14B_multi_clips.py)|-|-|-|-|
 |[Wan-AI/Wan2.2-I2V-A14B](https://modelscope.cn/models/Wan-AI/Wan2.2-I2V-A14B)|`input_image`|[code](./examples/wanvideo/model_inference/Wan2.2-I2V-A14B.py)|[code](./examples/wanvideo/model_training/full/Wan2.2-I2V-A14B.sh)|[code](./examples/wanvideo/model_training/validate_full/Wan2.2-I2V-A14B.py)|[code](./examples/wanvideo/model_training/lora/Wan2.2-I2V-A14B.sh)|[code](./examples/wanvideo/model_training/validate_lora/Wan2.2-I2V-A14B.py)|
 |[Wan-AI/Wan2.2-T2V-A14B](https://modelscope.cn/models/Wan-AI/Wan2.2-T2V-A14B)||[code](./examples/wanvideo/model_inference/Wan2.2-T2V-A14B.py)|[code](./examples/wanvideo/model_training/full/Wan2.2-T2V-A14B.sh)|[code](./examples/wanvideo/model_training/validate_full/Wan2.2-T2V-A14B.py)|[code](./examples/wanvideo/model_training/lora/Wan2.2-T2V-A14B.sh)|[code](./examples/wanvideo/model_training/validate_lora/Wan2.2-T2V-A14B.py)|
 |[Wan-AI/Wan2.2-TI2V-5B](https://modelscope.cn/models/Wan-AI/Wan2.2-TI2V-5B)|`input_image`|[code](./examples/wanvideo/model_inference/Wan2.2-TI2V-5B.py)|[code](./examples/wanvideo/model_training/full/Wan2.2-TI2V-5B.sh)|[code](./examples/wanvideo/model_training/validate_full/Wan2.2-TI2V-5B.py)|[code](./examples/wanvideo/model_training/lora/Wan2.2-TI2V-5B.sh)|[code](./examples/wanvideo/model_training/validate_lora/Wan2.2-TI2V-5B.py)|

diffsynth/models/wan_video_dit_s2v.py

@@ -410,7 +410,6 @@ def __init__(
         self.cond_encoder = nn.Conv3d(cond_dim, dim, kernel_size=patch_size, stride=patch_size)
         self.casual_audio_encoder = CausalAudioEncoder(dim=audio_dim, out_dim=dim, num_token=num_audio_token, need_global=enable_adain)
         all_modules, all_modules_names = torch_dfs(self.blocks, parent_name="root.transformer_blocks")
-        # TODO: refactor dfs
         self.audio_injector = AudioInjector_WAN(
             all_modules,
             all_modules_names,
@@ -449,7 +448,6 @@ def process_motion_frame_pack(self, motion_latents, drop_motion_frames=False, ad
 
     def inject_motion(self, x, rope_embs, mask_input, motion_latents, drop_motion_frames=True, add_last_motion=2):
         # inject the motion frames token to the hidden states
-        # TODO: check drop_motion_frames = False
         mot, mot_remb = self.process_motion_frame_pack(motion_latents, drop_motion_frames=drop_motion_frames, add_last_motion=add_last_motion)
         if len(mot) > 0:
             x = torch.cat([x, mot[0]], dim=1)

diffsynth/models/wav2vec.py

@@ -183,6 +183,13 @@ def get_audio_embed_bucket_fps(self, audio_embed, fps=16, batch_frames=81, m=0):
 
         return batch_audio_eb, min_batch_num
 
+    def get_audio_feats_per_inference(self, input_audio, sample_rate, processor, fps=16, batch_frames=80, m=0, dtype=torch.float32, device='cpu'):
+        audio_feat = self.extract_audio_feat(input_audio, sample_rate, processor, return_all_layers=True, dtype=dtype, device=device)
+        audio_embed_bucket, min_batch_num = self.get_audio_embed_bucket_fps(audio_feat, fps=fps, batch_frames=batch_frames, m=m)
+        audio_embed_bucket = audio_embed_bucket.unsqueeze(0).permute(0, 2, 3, 1).to(device, dtype)
+        audio_embeds = [audio_embed_bucket[..., i * batch_frames:(i + 1) * batch_frames] for i in range(min_batch_num)]
+        return audio_embeds
+
     @staticmethod
     def state_dict_converter():
         return WanS2VAudioEncoderStateDictConverter()