Nvidia Cuda with 8GB of RAM gives out of memory error

How do I have to modify the settings so that it runs on my RTX3060ti card, because now I get OOM?
It is a 8GB Nvidia Card, capable to run CUDA.

I'm calling Kandinsky with this script: https://gist.github.com/DanPli/0bfcaed38ed678adf3cd3f6aa0420f46

I get the following errors:

(venv) G:\k\Kandinsky>gen.py --prompt "green car"
G:\k\Kandinsky\venv\lib\site-packages\huggingface_hub\file_download.py:637: FutureWarning: `cached_download` is the legacy way to download files from the HF hub, please consider upgrading to `hf_hub_download`
warnings.warn(
making attention of type 'vanilla' with 512 in_channels
making attention of type 'vanilla' with 512 in_channels
making attention of type 'vanilla' with 512 in_channels
Working with z of shape (1, 4, 32, 32) = 4096 dimensions.
0%| | 0/30 [00:01
generate(**vars(parser.parse_args()))
File "G:\k\Kandinsky\gen.py", line 34, in generate
images = model.generate_text2img(
File "G:\k\Kandinsky\venv\lib\site-packages\torch\utils\_contextlib.py", line 115, in decorate_context
return func(*args, **kwargs)
File "G:\k\Kandinsky\venv\lib\site-packages\kandinsky2\kandinsky2_1_model.py", line 341, in generate_text2img return self.generate_img(
File "G:\k\Kandinsky\venv\lib\site-packages\torch\utils\_contextlib.py", line 115, in decorate_context
return func(*args, **kwargs)
File "G:\k\Kandinsky\venv\lib\site-packages\kandinsky2\kandinsky2_1_model.py", line 247, in generate_img
samples = diffusion.p_sample_loop(
File "G:\k\Kandinsky\venv\lib\site-packages\kandinsky2\model\gaussian_diffusion.py", line 413, in p_sample_loop
for sample in self.p_sample_loop_progressive(
File "G:\k\Kandinsky\venv\lib\site-packages\kandinsky2\model\gaussian_diffusion.py", line 466, in p_sample_loop_progressive
out = self.p_sample(
File "G:\k\Kandinsky\venv\lib\site-packages\kandinsky2\model\gaussian_diffusion.py", line 369, in p_sample
out = self.p_mean_variance(
File "G:\k\Kandinsky\venv\lib\site-packages\kandinsky2\model\respace.py", line 102, in p_mean_variance
return super().p_mean_variance(self._wrap_model(model), *args, **kwargs)
File "G:\k\Kandinsky\venv\lib\site-packages\kandinsky2\model\gaussian_diffusion.py", line 251, in p_mean_variance
model_output = model(x, s_t, **model_kwargs)
File "G:\k\Kandinsky\venv\lib\site-packages\kandinsky2\model\respace.py", line 133, in __call__
return self.model(x, new_ts, **kwargs)
File "G:\k\Kandinsky\venv\lib\site-packages\kandinsky2\kandinsky2_1_model.py", line 225, in model_fn
model_out = self.model(combined, ts, **kwargs)
File "G:\k\Kandinsky\venv\lib\site-packages\torch\nn\modules\module.py", line 1501, in _call_impl
return forward_call(*args, **kwargs)
File "G:\k\Kandinsky\venv\lib\site-packages\kandinsky2\model\text2im_model2_1.py", line 95, in forward
h = module(h, emb, xf_out)
File "G:\k\Kandinsky\venv\lib\site-packages\torch\nn\modules\module.py", line 1501, in _call_impl
return forward_call(*args, **kwargs)
File "G:\k\Kandinsky\venv\lib\site-packages\kandinsky2\model\unet.py", line 42, in forward
x = layer(x, encoder_out)
File "G:\k\Kandinsky\venv\lib\site-packages\torch\nn\modules\module.py", line 1501, in _call_impl
return forward_call(*args, **kwargs)
File "G:\k\Kandinsky\venv\lib\site-packages\kandinsky2\model\unet.py", line 265, in forward
h = self.attention(qkv, encoder_out)
File "G:\k\Kandinsky\venv\lib\site-packages\torch\nn\modules\module.py", line 1501, in _call_impl
return forward_call(*args, **kwargs)
File "G:\k\Kandinsky\venv\lib\site-packages\kandinsky2\model\unet.py", line 338, in forward
weight = torch.softmax(weight.float(), dim=-1).type(weight.dtype)
torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 506.00 MiB (GPU 0; 8.00 GiB total capacity; 6.83 GiB already allocated; 0 bytes free; 6.95 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF