RuntimeError: expected scalar type Half but found Float

[Extocine]

Using the default settings, every time I try to run i keep getting this error. Same issue in Debian 12 & Windows 10

Traceback (most recent call last): File "C:\Users\Extocine\Desktop\ai\LoRA_Easy_Training_Scripts\sd_scripts\train_network.py", line 996, in trainer.train(args) File "C:\Users\Extocine\Desktop\ai\LoRA_Easy_Training_Scripts\sd_scripts\train_network.py", line 748, in train latents = vae.encode(batch["images"].to(dtype=vae_dtype)).latent_dist.sample() File "C:\Users\Extocine\Desktop\ai\LoRA_Easy_Training_Scripts\sd_scripts\venv\lib\site-packages\diffusers\utils\accelerate_utils.py", line 46, in wrapper return method(self, *args, kwargs) File "C:\Users\Extocine\Desktop\ai\LoRA_Easy_Training_Scripts\sd_scripts\venv\lib\site-packages\diffusers\models\autoencoder_kl.py", line 258, in encode h = self.encoder(x) File "C:\Users\Extocine\Desktop\ai\LoRA_Easy_Training_Scripts\sd_scripts\venv\lib\site-packages\torch\nn\modules\module.py", line 1532, in _wrapped_call_impl return self._call_impl(*args, *kwargs) File "C:\Users\Extocine\Desktop\ai\LoRA_Easy_Training_Scripts\sd_scripts\venv\lib\site-packages\torch\nn\modules\module.py", line 1541, in _call_impl return forward_call(args, kwargs) File "C:\Users\Extocine\Desktop\ai\LoRA_Easy_Training_Scripts\sd_scripts\venv\lib\site-packages\diffusers\models\vae.py", line 141, in forward sample = down_block(sample) File "C:\Users\Extocine\Desktop\ai\LoRA_Easy_Training_Scripts\sd_scripts\venv\lib\site-packages\torch\nn\modules\module.py", line 1532, in _wrapped_call_impl return self._call_impl(*args, kwargs) File "C:\Users\Extocine\Desktop\ai\LoRA_Easy_Training_Scripts\sd_scripts\venv\lib\site-packages\torch\nn\modules\module.py", line 1541, in _call_impl return forward_call(*args, *kwargs) File "C:\Users\Extocine\Desktop\ai\LoRA_Easy_Training_Scripts\sd_scripts\venv\lib\site-packages\diffusers\models\unet_2d_blocks.py", line 1247, in forward hidden_states = resnet(hidden_states, temb=None, scale=scale) File "C:\Users\Extocine\Desktop\ai\LoRA_Easy_Training_Scripts\sd_scripts\venv\lib\site-packages\torch\nn\modules\module.py", line 1532, in _wrapped_call_impl return self._call_impl(args, kwargs) File "C:\Users\Extocine\Desktop\ai\LoRA_Easy_Training_Scripts\sd_scripts\venv\lib\site-packages\torch\nn\modules\module.py", line 1541, in _call_impl return forward_call(*args, kwargs) File "C:\Users\Extocine\Desktop\ai\LoRA_Easy_Training_Scripts\sd_scripts\venv\lib\site-packages\diffusers\models\resnet.py", line 659, in forward hidden_states = self.conv2(hidden_states, scale) File "C:\Users\Extocine\Desktop\ai\LoRA_Easy_Training_Scripts\sd_scripts\venv\lib\site-packages\torch\nn\modules\module.py", line 1532, in _wrapped_call_impl return self._call_impl(*args, *kwargs) File "C:\Users\Extocine\Desktop\ai\LoRA_Easy_Training_Scripts\sd_scripts\venv\lib\site-packages\torch\nn\modules\module.py", line 1541, in _call_impl return forward_call(args, kwargs) File "C:\Users\Extocine\Desktop\ai\LoRA_Easy_Training_Scripts\sd_scripts\venv\lib\site-packages\diffusers\models\lora.py", line 163, in forward return F.conv2d( KeyboardInterrupt steps: 0%| | 0/80 [05:20<?, ?it/s] Failed to train because of error: Command '['C:\Users\Extocine\Desktop\ai\LoRA_Easy_Training_Scripts\sd_scripts\venv\Scripts\python.exe', 'sd_scripts\train_network.py', '--config_file=runtime_store\config.toml', '--dataset_config=runtime_store\dataset.toml']' returned non-zero exit status 3221225786.

[Jelosus2]

Hello, could you tell me what's your GPU model and provide the toml file?

[Extocine]

I have an RX 5700 XT

I'm assuming youre talking about the configuration file

[[subsets]]
num_repeats = 1
caption_extension = ".txt"
shuffle_caption = false
flip_aug = false
color_aug = false
random_crop = false
is_reg = false
image_dir = "/home/extocine/Programs/ai/LoRA_Easy_Training_Scripts/training"
keep_tokens = 0

[noise_args]

[sample_args]

[logging_args]

[general_args.args]
pretrained_model_name_or_path = "/home/extocine/Programs/ai/stable-diffusion-webui/models/Stable-diffusion/EasyFluffV11.2.safetensors"
mixed_precision = "fp16"
seed = 23
max_train_epochs = 1
max_data_loader_n_workers = 1
persistent_data_loader_workers = true
max_token_length = 225
prior_loss_weight = 1.0
clip_skip = 2

[general_args.dataset_args]
resolution = 512
batch_size = 1

[network_args.args]
network_dim = 32
network_alpha = 16.0
min_timestep = 0
max_timestep = 1000

[optimizer_args.args]
optimizer_type = "AdamW"
lr_scheduler = "cosine"
learning_rate = 0.0001
max_grad_norm = 1.0

[saving_args.args]
output_dir = "/home/extocine/Programs/ai/trained"
save_precision = "bf16"
save_model_as = "safetensors"
output_name = "extomodel"

[bucket_args.dataset_args]
enable_bucket = true
min_bucket_reso = 256
max_bucket_reso = 1024
bucket_reso_steps = 64

[optimizer_args.args.optimizer_args]
weight_decay = "0.1"
betas = "0.9,0.99"```

[Jelosus2]

I see, AMD cant be used to train in most cases

[Extocine]

So I'm just outta luck?

[Jelosus2]

Sadly Nvidia has the whole monopoly of AI, you can use online alternatives like google colab (free 3h of GPU every 3 days iirc), paperspace (if you are willing to pay for pro subscription) or trainers from AI sites like civitai, tensor.art, etc

[Extocine]

Alright, thanks anyway ^^