torch.cuda.OutOfMemoryError: CUDA out of memory. / 6G card

Hello I'm using 'Model Inference' in RVC
In FAQ the out of memory error is mentioned, it's recommended to "for inference, adjust the x_pad, x_query, x_center, and x_max settings in the config.py file as needed."
Which values I should put in config file, what would be the correct adjustments?
This is the error I get after paths to the audio file, .pth and .index files in RVC are set and I click on 'Convert':

Traceback (most recent call last):
File "F:\RVC\infer-web.py", line 203, in vc_single
audio_opt = vc.pipeline(
File "F:\RVC\vc_infer_pipeline.py", line 361, in pipeline
self.vc(
File "F:\RVC\vc_infer_pipeline.py", line 249, in vc
(net_g.infer(feats, p_len, pitch, pitchf, sid)[0][0, 0])
File "F:\RVC\lib\infer_pack\models.py", line 752, in infer
m_p, logs_p, x_mask = self.enc_p(phone, pitch, phone_lengths)
File "F:\RVC\runtime\lib\site-packages\torch\nn\modules\module.py", line 1501, in _call_impl
return forward_call(*args, **kwargs)
File "F:\RVC\lib\infer_pack\models.py", line 104, in forward
x = self.encoder(x * x_mask, x_mask)
File "F:\RVC\runtime\lib\site-packages\torch\nn\modules\module.py", line 1501, in _call_impl
return forward_call(*args, **kwargs)
File "F:\RVC\lib\infer_pack\attentions.py", line 65, in forward
y = self.attn_layers[i](x, x, attn_mask)
File "F:\RVC\runtime\lib\site-packages\torch\nn\modules\module.py", line 1501, in _call_impl
return forward_call(*args, **kwargs)
File "F:\RVC\lib\infer_pack\attentions.py", line 221, in forward
x, self.attn = self.attention(q, k, v, mask=attn_mask)
File "F:\RVC\lib\infer_pack\attentions.py", line 265, in attention
relative_weights = self._absolute_position_to_relative_position(p_attn)
File "F:\RVC\lib\infer_pack\attentions.py", line 346, in _absolute_position_to_relative_position
x_flat = F.pad(x_flat, commons.convert_pad_shape([[0, 0], [0, 0], [length, 0]]))
torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 428.00 MiB (GPU 0; 6.00 GiB total capacity; 3.02 GiB already allocated; 373.12 MiB free; 3.81 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

Traceback (most recent call last):
File "F:\RVC\runtime\lib\site-packages\gradio\routes.py", line 321, in run_predict
output = await app.blocks.process_api(
File "F:\RVC\runtime\lib\site-packages\gradio\blocks.py", line 1007, in process_api
data = self.postprocess_data(fn_index, result["prediction"], state)
File "F:\RVC\runtime\lib\site-packages\gradio\blocks.py", line 953, in postprocess_data
prediction_value = block.postprocess(prediction_value)
File "F:\RVC\runtime\lib\site-packages\gradio\components.py", line 2076, in postprocess
processing_utils.audio_to_file(sample_rate, data, file.name)
File "F:\RVC\runtime\lib\site-packages\gradio\processing_utils.py", line 206, in audio_to_file
data = convert_to_16_bit_wav(data)
File "F:\RVC\runtime\lib\site-packages\gradio\processing_utils.py", line 219, in convert_to_16_bit_wav
if data.dtype in [np.float64, np.float32, np.float16]:
AttributeError: 'NoneType' object has no attribute 'dtype'

Start with config.py and reproduce the Model Inference path from infer-web.py, then inspect the pipeline calls in vc_infer_pipeline.py and the failing inference code in lib/infer_pack/models.py. Determine which documented settings affect this 6G-card failure and verify the result by running the same conversion without the CUDA out-of-memory error.