Spaces:
Running
on
A10G
Running
on
A10G
Linoy Tsaban
commited on
Commit
·
8cd26eb
1
Parent(s):
406e2d8
Update app.py
Browse files
app.py
CHANGED
@@ -92,77 +92,70 @@ def edit(input_image,
|
|
92 |
# invert and retrieve noise maps and latent
|
93 |
wt, zs, wts = invert(x0 =x0 , prompt_src=src_prompt, num_diffusion_steps=steps, cfg_scale_src=cfg_scale_src)
|
94 |
|
95 |
-
#
|
96 |
-
xT=wts[skip]
|
97 |
-
etas=1.0
|
98 |
-
prompts=[tar_prompt]
|
99 |
-
cfg_scales=[cfg_scale_tar]
|
100 |
-
prog_bar=False
|
101 |
-
zs=zs[skip:]
|
102 |
|
103 |
|
104 |
-
batch_size = len(prompts)
|
105 |
|
106 |
-
cfg_scales_tensor = torch.Tensor(cfg_scales).view(-1,1,1,1).to(sd_pipe.device)
|
107 |
|
108 |
-
text_embeddings = encode_text(sd_pipe, prompts)
|
109 |
-
uncond_embedding = encode_text(sd_pipe, [""] * batch_size)
|
110 |
|
111 |
-
if etas is None: etas = 0
|
112 |
-
if type(etas) in [int, float]: etas = [etas]*sd_pipe.scheduler.num_inference_steps
|
113 |
-
assert len(etas) == sd_pipe.scheduler.num_inference_steps
|
114 |
-
timesteps = sd_pipe.scheduler.timesteps.to(sd_pipe.device)
|
115 |
|
116 |
-
xt = xT.expand(batch_size, -1, -1, -1)
|
117 |
-
op = tqdm(timesteps[-zs.shape[0]:]) if prog_bar else timesteps[-zs.shape[0]:]
|
118 |
|
119 |
-
t_to_idx = {int(v):k for k,v in enumerate(timesteps[-zs.shape[0]:])}
|
120 |
|
121 |
-
for t in op:
|
122 |
-
|
123 |
-
|
124 |
-
|
125 |
-
|
126 |
-
|
127 |
|
128 |
-
|
129 |
-
|
130 |
-
|
131 |
-
|
132 |
-
|
133 |
|
134 |
|
135 |
-
|
136 |
-
|
137 |
-
|
138 |
-
|
139 |
-
|
140 |
-
|
141 |
-
|
142 |
-
|
143 |
-
|
144 |
-
|
145 |
-
|
146 |
-
|
147 |
-
|
148 |
-
|
149 |
-
|
150 |
-
|
151 |
-
|
152 |
|
153 |
-
yield interm_img
|
154 |
|
155 |
-
# # vae decode image
|
156 |
-
# with autocast("cuda"), inference_mode():
|
157 |
-
# x0_dec = sd_pipe.vae.decode(1 / 0.18215 * w0).sample
|
158 |
-
# if x0_dec.dim()<4:
|
159 |
-
# x0_dec = x0_dec[None,:,:,:]
|
160 |
-
# img = image_grid(x0_dec)
|
161 |
-
# return img
|
162 |
|
163 |
-
|
164 |
|
165 |
-
|
166 |
|
167 |
|
168 |
|
|
|
92 |
# invert and retrieve noise maps and latent
|
93 |
wt, zs, wts = invert(x0 =x0 , prompt_src=src_prompt, num_diffusion_steps=steps, cfg_scale_src=cfg_scale_src)
|
94 |
|
95 |
+
# #
|
96 |
+
# xT=wts[skip]
|
97 |
+
# etas=1.0
|
98 |
+
# prompts=[tar_prompt]
|
99 |
+
# cfg_scales=[cfg_scale_tar]
|
100 |
+
# prog_bar=False
|
101 |
+
# zs=zs[skip:]
|
102 |
|
103 |
|
104 |
+
# batch_size = len(prompts)
|
105 |
|
106 |
+
# cfg_scales_tensor = torch.Tensor(cfg_scales).view(-1,1,1,1).to(sd_pipe.device)
|
107 |
|
108 |
+
# text_embeddings = encode_text(sd_pipe, prompts)
|
109 |
+
# uncond_embedding = encode_text(sd_pipe, [""] * batch_size)
|
110 |
|
111 |
+
# if etas is None: etas = 0
|
112 |
+
# if type(etas) in [int, float]: etas = [etas]*sd_pipe.scheduler.num_inference_steps
|
113 |
+
# assert len(etas) == sd_pipe.scheduler.num_inference_steps
|
114 |
+
# timesteps = sd_pipe.scheduler.timesteps.to(sd_pipe.device)
|
115 |
|
116 |
+
# xt = xT.expand(batch_size, -1, -1, -1)
|
117 |
+
# op = tqdm(timesteps[-zs.shape[0]:]) if prog_bar else timesteps[-zs.shape[0]:]
|
118 |
|
119 |
+
# t_to_idx = {int(v):k for k,v in enumerate(timesteps[-zs.shape[0]:])}
|
120 |
|
121 |
+
# for t in op:
|
122 |
+
# idx = t_to_idx[int(t)]
|
123 |
+
# ## Unconditional embedding
|
124 |
+
# with torch.no_grad():
|
125 |
+
# uncond_out = sd_pipe.unet.forward(xt, timestep = t,
|
126 |
+
# encoder_hidden_states = uncond_embedding)
|
127 |
|
128 |
+
# ## Conditional embedding
|
129 |
+
# if prompts:
|
130 |
+
# with torch.no_grad():
|
131 |
+
# cond_out = sd_pipe.unet.forward(xt, timestep = t,
|
132 |
+
# encoder_hidden_states = text_embeddings)
|
133 |
|
134 |
|
135 |
+
# z = zs[idx] if not zs is None else None
|
136 |
+
# z = z.expand(batch_size, -1, -1, -1)
|
137 |
+
# if prompts:
|
138 |
+
# ## classifier free guidance
|
139 |
+
# noise_pred = uncond_out.sample + cfg_scales_tensor * (cond_out.sample - uncond_out.sample)
|
140 |
+
# else:
|
141 |
+
# noise_pred = uncond_out.sample
|
142 |
+
# # 2. compute less noisy image and set x_t -> x_t-1
|
143 |
+
# xt = reverse_step(sd_pipe, noise_pred, t, xt, eta = etas[idx], variance_noise = z)
|
144 |
+
|
145 |
+
# # interm denoised img
|
146 |
+
# with autocast("cuda"), inference_mode():
|
147 |
+
# x0_dec = sd_pipe.vae.decode(1 / 0.18215 * xt).sample
|
148 |
+
# if x0_dec.dim()<4:
|
149 |
+
# x0_dec = x0_dec[None,:,:,:]
|
150 |
+
# interm_img = image_grid(x0_dec)
|
151 |
+
# yield interm_img
|
152 |
|
153 |
+
# yield interm_img
|
154 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
155 |
|
156 |
+
output = sample(wt, zs, wts, prompt_tar=tar_prompt, cfg_scale_tar=cfg_scale_tar, skip=skip)
|
157 |
|
158 |
+
return output
|
159 |
|
160 |
|
161 |
|