Browse Source

Merge branch 'master' into pythongosssss-socketrework

pull/27/head
comfyanonymous 2 years ago
parent
commit
8074a58a1a
  1. 2
      README.md
  2. 2
      comfy/samplers.py
  3. 30
      comfy/sd.py
  4. 3
      notebooks/comfyui_colab.ipynb

2
README.md

@ -50,6 +50,8 @@ Nvidia users should install torch using this command:
Nvidia users should also install Xformers for a speed boost but can still run the software without it.
```pip install xformers```
#### Troubleshooting
If you get the "Torch not compiled with CUDA enabled" error, uninstall torch with:

2
comfy/samplers.py

@ -167,7 +167,7 @@ def sampling_function(model_function, x, timestep, uncond, cond, cond_scale, con
timestep_ = torch.cat([timestep] * batch_chunks)
if control is not None:
c['control'] = control.get_control(input_x, timestep_, c['c_crossattn'])
c['control'] = control.get_control(input_x, timestep_, c['c_crossattn'], len(cond_or_uncond))
output = model_function(input_x, timestep_, cond=c).chunk(batch_chunks)
del input_x

30
comfy/sd.py

@ -359,6 +359,28 @@ class VAE:
samples = samples.cpu()
return samples
def resize_image_to(tensor, target_latent_tensor, batched_number):
tensor = utils.common_upscale(tensor, target_latent_tensor.shape[3] * 8, target_latent_tensor.shape[2] * 8, 'nearest-exact', "center")
target_batch_size = target_latent_tensor.shape[0]
current_batch_size = tensor.shape[0]
print(current_batch_size, target_batch_size)
if current_batch_size == 1:
return tensor
per_batch = target_batch_size // batched_number
tensor = tensor[:per_batch]
if per_batch > tensor.shape[0]:
tensor = torch.cat([tensor] * (per_batch // tensor.shape[0]) + [tensor[:(per_batch % tensor.shape[0])]], dim=0)
current_batch_size = tensor.shape[0]
if current_batch_size == target_batch_size:
return tensor
else:
return torch.cat([tensor] * batched_number, dim=0)
class ControlNet:
def __init__(self, control_model, device="cuda"):
self.control_model = control_model
@ -368,7 +390,7 @@ class ControlNet:
self.device = device
self.previous_controlnet = None
def get_control(self, x_noisy, t, cond_txt):
def get_control(self, x_noisy, t, cond_txt, batched_number):
control_prev = None
if self.previous_controlnet is not None:
control_prev = self.previous_controlnet.get_control(x_noisy, t, cond_txt)
@ -378,7 +400,7 @@ class ControlNet:
if self.cond_hint is not None:
del self.cond_hint
self.cond_hint = None
self.cond_hint = utils.common_upscale(self.cond_hint_original, x_noisy.shape[3] * 8, x_noisy.shape[2] * 8, 'nearest-exact', "center").to(self.control_model.dtype).to(self.device)
self.cond_hint = resize_image_to(self.cond_hint_original, x_noisy, batched_number).to(self.control_model.dtype).to(self.device)
if self.control_model.dtype == torch.float16:
precision_scope = torch.autocast
@ -516,7 +538,7 @@ class T2IAdapter:
self.cond_hint_original = None
self.cond_hint = None
def get_control(self, x_noisy, t, cond_txt):
def get_control(self, x_noisy, t, cond_txt, batched_number):
control_prev = None
if self.previous_controlnet is not None:
control_prev = self.previous_controlnet.get_control(x_noisy, t, cond_txt)
@ -525,7 +547,7 @@ class T2IAdapter:
if self.cond_hint is not None:
del self.cond_hint
self.cond_hint = None
self.cond_hint = utils.common_upscale(self.cond_hint_original, x_noisy.shape[3] * 8, x_noisy.shape[2] * 8, 'nearest-exact', "center").float().to(self.device)
self.cond_hint = resize_image_to(self.cond_hint_original, x_noisy, batched_number).float().to(self.device)
if self.channels_in == 1 and self.cond_hint.shape[1] > 1:
self.cond_hint = torch.mean(self.cond_hint, 1, keepdim=True)
self.t2i_model.to(self.device)

3
notebooks/comfyui_colab.ipynb

@ -35,7 +35,8 @@
"source": [
"!git clone https://github.com/comfyanonymous/ComfyUI\n",
"%cd ComfyUI\n",
"!pip install -r requirements.txt"
"!pip install -r requirements.txt\n",
"!pip install xformers"
]
},
{

Loading…
Cancel
Save