stable-dreambooth
stable-dreambooth copied to clipboard
Victarry
Dreambooth implementation based on Stable Diffusion with minimal code.
File "train.py", line 206, in train_loop(config, model, noise_scheduler, optimizer, train_dataloader) File "train.py", line 131, in train_loop noisy_latents = noise_scheduler.add_noise(latents, noise, timesteps.cpu().numpy()) File "/HPS/EgofaceTrial/work/anaconda3/envs/stable-diffusion/lib/python3.8/site-packages/diffusers/schedulers/scheduling_ddpm.py", line 303, in add_noise timesteps = timesteps.to(original_samples.device)...
for text in datasets: with torch.no_grad(): images = model(text, height=512, width=512, num_inference_steps=50)["sample"] /// sample key work error in conda activate enviroment
I have a similar question AttributeError: 'StableDiffusionPipeline' object has no attribute 'parameters' diffusers 0.15.0 can you help me solve it?
How to get this working on an EC2 instance? any steps/tutorials? Thank you.
Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass RuntimeError: CUDA out of memory. Tried to allocate 4.00 GiB (GPU 0; 47.46 GiB total capacity; 44.29 GiB...
It might be helpful to explain the needed images to train a new model. The partial example with some images in data/dogs/instance is more confusing than it helps. Would it...
I was able to successfully train a model after chaning the diffusers version and changing batch size to 2 but when running inference on the output I only get reconstructions...
Hi, similar to https://github.com/huggingface/diffusers/tree/main/examples/textual_inversion, would you be interested in adding example colab notebook for training and inference users the diffusers pipeline?
