MCPcopy Create free account
hub / github.com/OpenImagingLab/FlashVSR / forward

Method forward

diffsynth/models/wan_video_image_encoder.py:571–580  ·  view source on GitHub ↗

imgs: [B, 3, H, W] of torch.float32. - mean: [0.48145466, 0.4578275, 0.40821073] - std: [0.26862954, 0.26130258, 0.27577711] txt_ids: [B, L] of torch.long. Encoded by data.CLIPTokenizer.

(self, imgs, txt_ids)

Source from the content-addressed store, hash-verified

569 self.init_weights()
570
571 def forward(self, imgs, txt_ids):
572 """
573 imgs: [B, 3, H, W] of torch.float32.
574 - mean: [0.48145466, 0.4578275, 0.40821073]
575 - std: [0.26862954, 0.26130258, 0.27577711]
576 txt_ids: [B, L] of torch.long. Encoded by data.CLIPTokenizer.
577 """
578 xi = self.visual(imgs)
579 xt = self.textual(txt_ids)
580 return xi, xt
581
582 def init_weights(self):
583 # embeddings

Callers 2

forwardMethod · 0.45
forwardMethod · 0.45

Calls

no outgoing calls

Tested by

no test coverage detected