dynamic thresholding (maybe?)

ldm/models/diffusion/ddim.py

@@ -4,6 +4,7 @@ import torch
 import numpy as np
 from tqdm import tqdm
 from functools import partial
+from einops import rearrange
 
 from ldm.modules.diffusionmodules.util import make_ddim_sampling_parameters, make_ddim_timesteps, noise_like
 
@@ -73,8 +74,8 @@ class DDIMSampler(object):
                x_T=None,
                log_every_t=100,
                unconditional_guidance_scale=1.,
-               unconditional_conditioning=None,
+               unconditional_conditioning=None, # this has to come in the same format as the conditioning, # e.g. as encoded tokens, ...
-               # this has to come in the same format as the conditioning, # e.g. as encoded tokens, ...
+               dynamic_threshold=None,
                **kwargs
                ):
         if conditioning is not None:
@@ -106,6 +107,7 @@ class DDIMSampler(object):
                                                     log_every_t=log_every_t,
                                                     unconditional_guidance_scale=unconditional_guidance_scale,
                                                     unconditional_conditioning=unconditional_conditioning,
+                                                    dynamic_threshold=dynamic_threshold,
                                                     )
         return samples, intermediates
 
@@ -115,7 +117,7 @@ class DDIMSampler(object):
                       callback=None, timesteps=None, quantize_denoised=False,
                       mask=None, x0=None, img_callback=None, log_every_t=100,
                       temperature=1., noise_dropout=0., score_corrector=None, corrector_kwargs=None,
-                      unconditional_guidance_scale=1., unconditional_conditioning=None,):
+                      unconditional_guidance_scale=1., unconditional_conditioning=None, dynamic_threshold=None):
         device = self.model.betas.device
         b = shape[0]
         if x_T is None:
@@ -150,7 +152,8 @@ class DDIMSampler(object):
                                       noise_dropout=noise_dropout, score_corrector=score_corrector,
                                       corrector_kwargs=corrector_kwargs,
                                       unconditional_guidance_scale=unconditional_guidance_scale,
-                                      unconditional_conditioning=unconditional_conditioning)
+                                      unconditional_conditioning=unconditional_conditioning,
+                                      dynamic_threshold=dynamic_threshold)
             img, pred_x0 = outs
             if callback: callback(i)
             if img_callback: img_callback(pred_x0, i)
@@ -164,7 +167,8 @@ class DDIMSampler(object):
     @torch.no_grad()
     def p_sample_ddim(self, x, c, t, index, repeat_noise=False, use_original_steps=False, quantize_denoised=False,
                       temperature=1., noise_dropout=0., score_corrector=None, corrector_kwargs=None,
-                      unconditional_guidance_scale=1., unconditional_conditioning=None):
+                      unconditional_guidance_scale=1., unconditional_conditioning=None,
+                      dynamic_threshold=None):
         b, *_, device = *x.shape, x.device
 
         if unconditional_conditioning is None or unconditional_guidance_scale == 1.:
@@ -194,6 +198,33 @@ class DDIMSampler(object):
         pred_x0 = (x - sqrt_one_minus_at * e_t) / a_t.sqrt()
         if quantize_denoised:
             pred_x0, _, *_ = self.model.first_stage_model.quantize(pred_x0)
+
+        if dynamic_threshold is not None:
+            # renorm
+            pred_max = pred_x0.max()
+            pred_min = pred_x0.min()
+            pred_x0 = (pred_x0-pred_min)/(pred_max-pred_min)   # 0 ... 1
+            pred_x0 = 2*pred_x0 - 1.   # -1 ... 1
+
+            s = torch.quantile(
+                rearrange(pred_x0, 'b ... -> b (...)').abs(),
+                dynamic_threshold,
+                dim=-1
+            )
+            s.clamp_(min=1.0)
+            s = s.view(-1, *((1,) * (pred_x0.ndim - 1)))
+
+            # clip by threshold
+            #pred_x0 = pred_x0.clamp(-s, s) / s  # needs newer pytorch  # TODO bring back to pure-gpu with min/max
+
+            # temporary hack: numpy on cpu
+            pred_x0 = np.clip(pred_x0.cpu().numpy(), -s.cpu().numpy(), s.cpu().numpy()) / s.cpu().numpy()
+            pred_x0 = torch.tensor(pred_x0).to(self.model.device)
+
+            # re.renorm
+            pred_x0 = (pred_x0 + 1.) / 2.   # 0 ... 1
+            pred_x0 = (pred_max-pred_min)*pred_x0 + pred_min  # orig range
+
         # direction pointing to x_t
         dir_xt = (1. - a_prev - sigma_t**2).sqrt() * e_t
         noise = sigma_t * noise_like(x.shape, device, repeat_noise) * temperature

ldm/modules/encoders/modules.py

@@ -12,6 +12,10 @@ class AbstractEncoder(nn.Module):
     def encode(self, *args, **kwargs):
         raise NotImplementedError
 
+class IdentityEncoder(AbstractEncoder):
+
+    def encode(self, x):
+        return x
 
 
 class ClassEmbedder(nn.Module):