| import scipy |
| import PIL |
| import numpy as np |
| import torch |
| import torch.nn.functional as F |
|
|
| @torch.no_grad() |
| def project_face_embs(pipeline, face_embs): |
|
|
| ''' |
| face_embs: (N, 512) normalized ArcFace embeddings |
| ''' |
|
|
| arcface_token_id = pipeline.tokenizer.encode("id", add_special_tokens=False)[0] |
|
|
| input_ids = pipeline.tokenizer( |
| "photo of a id person", |
| truncation=True, |
| padding="max_length", |
| max_length=pipeline.tokenizer.model_max_length, |
| return_tensors="pt", |
| ).input_ids.to(pipeline.device) |
|
|
| face_embs_padded = F.pad(face_embs, (0, pipeline.text_encoder.config.hidden_size-512), "constant", 0) |
| token_embs = pipeline.text_encoder(input_ids=input_ids.repeat(len(face_embs), 1), return_token_embs=True) |
| token_embs[input_ids==arcface_token_id] = face_embs_padded |
|
|
| prompt_embeds = pipeline.text_encoder( |
| input_ids=input_ids, |
| input_token_embs=token_embs |
| )[0] |
|
|
| return prompt_embeds |
|
|
|
|
| def project_face_embs_with_grad(encoder, tokenizer, face_embs): |
| """ |
| Same as project_face_embs but allows gradients for training. |
| """ |
| arcface_token_id = tokenizer.encode("id", add_special_tokens=False)[0] |
|
|
| input_ids = tokenizer( |
| "photo of a id person", |
| truncation=True, |
| padding="max_length", |
| max_length=tokenizer.model_max_length, |
| return_tensors="pt", |
| ).input_ids.to(encoder.device) |
|
|
| face_embs_padded = F.pad(face_embs, (0, encoder.config.hidden_size - 512), "constant", 0) |
|
|
| input_ids_batch = input_ids.repeat(len(face_embs), 1) |
| token_embs = encoder(input_ids=input_ids_batch, return_token_embs=True) |
|
|
| face_embs_padded = face_embs_padded.to(token_embs.dtype) |
| token_embs[input_ids_batch == arcface_token_id] = face_embs_padded |
|
|
| prompt_embeds = encoder( |
| input_ids=input_ids_batch, |
| input_token_embs=token_embs |
| )[0] |
|
|
| return prompt_embeds |
|
|
|
|
| def image_align(img, |
| face_landmarks, |
| output_size=1024, |
| transform_size=4096, |
| enable_padding=True): |
| |
| |
|
|
| lm = face_landmarks |
| lm_eye_left = lm[36:42] |
| lm_eye_right = lm[42:48] |
| lm_mouth_outer = lm[48:60] |
|
|
| eye_left = np.mean(lm_eye_left, axis=0) |
| eye_right = np.mean(lm_eye_right, axis=0) |
| eye_avg = (eye_left + eye_right) * 0.5 |
| eye_to_eye = eye_right - eye_left |
| mouth_left = lm_mouth_outer[0] |
| mouth_right = lm_mouth_outer[6] |
| mouth_avg = (mouth_left + mouth_right) * 0.5 |
| eye_to_mouth = mouth_avg - eye_avg |
|
|
| x = eye_to_eye - np.flipud(eye_to_mouth) * [-1, 1] |
| x /= np.hypot(*x) |
| x *= max(np.hypot(*eye_to_eye) * 2.0, np.hypot(*eye_to_mouth) * 1.8) |
| y = np.flipud(x) * [-1, 1] |
| c = eye_avg + eye_to_mouth * 0.1 |
| quad = np.stack([c - x - y, c - x + y, c + x + y, c + x - y]) |
| qsize = np.hypot(*x) * 2 |
|
|
| img = img.convert('RGB') |
|
|
| shrink = int(np.floor(qsize / output_size * 0.5)) |
| if shrink > 1: |
| rsize = (int(np.rint(float(img.size[0]) / shrink)), |
| int(np.rint(float(img.size[1]) / shrink))) |
| img = img.resize(rsize, PIL.Image.LANCZOS) |
| quad /= shrink |
| qsize /= shrink |
|
|
| border = max(int(np.rint(qsize * 0.1)), 3) |
| crop = (int(np.floor(min(quad[:, 0]))), int(np.floor(min(quad[:, 1]))), |
| int(np.ceil(max(quad[:, 0]))), int(np.ceil(max(quad[:, 1])))) |
| crop = (max(crop[0] - border, 0), max(crop[1] - border, 0), |
| min(crop[2] + border, img.size[0]), min(crop[3] + border, img.size[1])) |
| if crop[2] - crop[0] < img.size[0] or crop[3] - crop[1] < img.size[1]: |
| img = img.crop(crop) |
| quad -= crop[0:2] |
|
|
| pad = (int(np.floor(min(quad[:, 0]))), int(np.floor(min(quad[:, 1]))), |
| int(np.ceil(max(quad[:, 0]))), int(np.ceil(max(quad[:, 1])))) |
| pad = (max(-pad[0] + border, 0), max(-pad[1] + border, 0), |
| max(pad[2] - img.size[0] + border, 0), max(pad[3] - img.size[1] + border, 0)) |
| if enable_padding and max(pad) > border - 4: |
| pad = np.maximum(pad, int(np.rint(qsize * 0.3))) |
| img = np.pad(np.float32(img), |
| ((pad[1], pad[3]), (pad[0], pad[2]), (0, 0)), 'reflect') |
| h, w, _ = img.shape |
| y, x, _ = np.ogrid[:h, :w, :1] |
| mask = np.maximum( |
| 1.0 - np.minimum(np.float32(x) / pad[0], np.float32(w - 1 - x) / pad[2]), |
| 1.0 - np.minimum(np.float32(y) / pad[1], np.float32(h - 1 - y) / pad[3])) |
| blur = qsize * 0.02 |
| img += (scipy.ndimage.gaussian_filter(img, [blur, blur, 0]) - img) * np.clip(mask * 3.0 + 1.0, 0.0, 1.0) |
| img += (np.median(img, axis=(0, 1)) - img) * np.clip(mask, 0.0, 1.0) |
| img = PIL.Image.fromarray(np.uint8(np.clip(np.rint(img), 0, 255)), 'RGB') |
| quad += pad[:2] |
|
|
| img = img.transform((transform_size, transform_size), PIL.Image.QUAD, |
| (quad + 0.5).flatten(), PIL.Image.BILINEAR) |
| if output_size < transform_size: |
| img = img.resize((output_size, output_size), PIL.Image.LANCZOS) |
|
|
| return img |
|
|
