3jO+SSKJrJrJrJrJr SSKrSSKJr SSKJ r SSK J r SSK J r Jr SSKJr SSKJr SS KJrJr SS KJr SS KJr /S QrS SSSS.r"SS\\5r"SS5r"SS5rg))AnyListOptionalTupleUnionN)nn) kornia_config)PILImage)basicsronnx)ONNXExportMixin) ModelBase)OutputRangePostProcessorResizePreProcessor)SmallSRNetWrapper)CachedDownloader)RRDBNetBuilderSmallSRBuilderSuperResolutionzUhttps://github.com/xinntao/Real-ESRGAN/releases/download/v0.1.0/RealESRGAN_x4plus.pthzUhttps://github.com/xinntao/Real-ESRGAN/releases/download/v0.1.1/RealESRNet_x4plus.pthz`https://github.com/xinntao/Real-ESRGAN/releases/download/v0.2.2.4/RealESRGAN_x4plus_anime_6B.pthzUhttps://github.com/xinntao/Real-ESRGAN/releases/download/v0.2.1/RealESRGAN_x2plus.pth)RealESRGAN_x4plusRealESRNet_x4plusRealESRGAN_x4plus_anime_6BRealESRGAN_x2plusc \rSrSr%SrSr\\S'\\ \S'\\ \S'\\ \S'\ R"5S\ \ R\\ R4S \ \ R\\ R44S j5rSS\ \ R\\ R4S \\ \ R\\ R4S \S \ \ R\\ R\S44SjjrSS\ \ R\\ R4S \\ \ R\\ R4S\\S \S S 4 SjjrSS\\S\S\S\\\\\4S\S S4 SjjrSrg )r+zASuperResolution is a module that wraps an super resolution model.super_resolutionnameinput_image_sizeoutput_image_sizepseudo_image_sizeimagesreturnc2URU5n[U[[-5(aUSnOUn[U[5(a'UVs/sHo0R US5SPM nnOUR U5nUR U5$s snf)zForward pass of the super resolution model. Args: images: If list of RGB images. Each image is a torch.Tensor with shape :math:`(3, H, W)`. If torch.Tensor, a torch.Tensor with shape :math:`(B, 3, H, W)`. Returns: output torch.Tensor. rN) pre_processor isinstancelisttuplemodelpost_processor)selfr!outputimage out_imagess Y/home/wildlama/miniconda3/lib/python3.13/site-packages/kornia/contrib/super_resolution.pyforwardSuperResolution.forward3s##F+ fdUl + +AYFF fd # #BHI&**U4[1!4&JIJF+J"":..Js BN edge_maps output_typez Image.ImagecUcURU5n/nUHnURU5 M URXC[U[R 5S9$)zDraw the super resolution results. Args: images: input torch.Tensor. edge_maps: detected edges. output_type: type of the output. Returns: output torch.Tensor. )is_batch)r/append_tensor_to_typer%torchTensor)r*r!r1r2r+edge_maps r. visualizeSuperResolution.visualizeJsY"   V,I!H MM( #"##F*VUZUaUaBb#cc directorychURXU5nURXSS9 URXSSS9 g)a Save the super resolution results. Args: images: input torch.Tensor. edge_maps: detected edges. output_type: type of the output. directory: where to save outputs. output_type: backend used to generate outputs. Returns: output torch.Tensor. _src)suffix_srN)r: _save_outputs)r*r!r1r=r2outputss r.saveSuperResolution.savecs:(..K@ 6V< 7e.onnx" will be used. image_size: The size to which input images will be resized during preprocessing. If None, image_size will be dynamic. For DexiNed, recommended scale is 352. include_pre_and_post_processor: Whether to include the pre-processor and post-processor in the exported model. save: If to save the model or load it. additional_metadata: Additional metadata to add to the ONNX model. kwargs: Additional arguments for converting to onnx. kornia_z.onnxi`) input_shape output_shape pseudo_shaper(rDrH)rr to_onnxrrr r()r*rFrGrDrHrIs r.rRSuperResolution.to_onnx{s4  !$))E2I&&    Q 5 5 ;T=R=R=XVXYa!7!7!=2t?U?U?[Y[\Q 6 6 =#t?U?U?\Y\]8$djj 3    r<)Nr7)NNr7)NTTN)__name__ __module__ __qualname____firstlineno____doc__rstr__annotations__rintr7inference_moderr8rr/r:rDboolrrrR__static_attributes__rTr<r.rr+sK"D#"sm#}$}$ /eELL$u||2D$DE/%PUP\P\^bchcoco^pPpJq//2HL" dellD$667dE%,,U\\0B"BCDd d u||T%,,/m1DD E d8HL#'" =ellD$667=E%,,U\\0B"BCD=C= =  =  =4$(/3?C ' C=' )-'  ' &d5c?&;< '  '  ' ' r<rc<\rSrSrSr\S S\S\S\4Sjj5r Sr g) rzSProvide a static factory to build Residual-in-Residual Dense Block (RRDB) networks. model_name pretrainedr"c US:Xa-[RRRSSSSSSS9nOUS:Xa-[RRRSSSSSSS9nOuUS :Xa-[RRRSSSS SSS9nOBUS :Xa-[RRRSSSSSS S9nO[ S US35eSnU(ah[ Un[ R"X@SS[RS9nUR[R"U[R"S5S9SSS9 UR5 [U[ R""5[%SSS9US9$)aBuild a preconfigured RRDB (Residual-in-Residual Dense Block) model. The returned object is a :class:`SuperResolution` wrapper configured for inference. Internally, this factory selects architecture hyperparameters based on ``model_name``, and optionally loads pretrained weights. Args: model_name: Name of the RRDB variant to construct. Supported options are: ``"RealESRGAN_x4plus"``, ``"RealESRNet_x4plus"``, ``"RealESRGAN_x4plus_anime_6B"``, and ``"RealESRGAN_x2plus"``. pretrained: If ``True``, download and load pretrained weights for the selected model variant. Returns: A :class:`SuperResolution` wrapper ready for inference. Raises: ValueError: If ``model_name`` is not one of the supported variants. rrM@ ) num_in_ch num_out_chnum_feat num_block num_grow_chscalerrrModel z{ not found. Please choose from 'RealESRGAN_x4plus', 'RealESRNet_x4plus', 'RealESRGAN_x4plus_anime_6B', 'RealESRGAN_x2plus'.NTz.pth)downloadr@ cache_dircpu) map_location params_ema)strictgg?)min_valmax_valr$r)r)r archs rrdbnet_archRRDBNet ValueErrorURLsrdownload_to_cacher hub_onnx_dirload_state_dictr7loaddeviceevalrrIdentityr)rbrcr( model_pathurls r.buildRRDBNetBuilder.builds* , ,MM..66B"RT\]7E. .MM..66B"RT\]7E7 7MM..66B!QS[\7E. .MM..66B"RT\]7E %oo   z"C);;$vIcIcJ  ! !%**ZellSXFY"Z[g"hqu ! v  ++-3CM   r<rTN)rT) rUrVrWrXrY staticmethodrZr^rrr_rTr<r.rrs-]8 #8 8 Q`8 8 r<rc L\rSrSrSr\S S\S\S\S\ \S\ 4 S jj5r S r g) rzFProvide a static factory to build lightweight Super-Resolution models.Nrbrcupscale_factor image_sizer"cUR5S:Xa [X!S9nO[SUS35e[U[ SS5[ R "5US9nUc SUlU$X5lUS-Ul X5lU$)aaBuild a lightweight super-resolution (SR) model wrapper. SR stands for *super-resolution*, which means generating a higher resolution output from a lower resolution input. Args: model_name: Name of the model variant to build. Currently only ``"small_sr"`` is supported. pretrained: If ``True``, load pretrained weights for the selected model. upscale_factor: Integer scale used by ``SmallSRNetWrapper``. For example, ``3`` maps an input image of size ``H x W`` to an output image of approximately ``3H x 3W``. image_size: Optional fixed input size. When provided, output metadata is configured as ``image_size * upscale_factor``. Returns: A :class:`SuperResolution` instance configured with resizing pre-processing and identity post-processing. Raises: ValueError: If ``model_name`` is not supported. small_sr)rcrqz* not found. Please choose from 'small_sr'.rzrM) lowerrr~rrrrr rr)rbrcrrr(srs r.rSmallSRBuilder.builds4     +%nLEvj\1[\] ]  ,S#6;;=    #&B  #- #->B #-  r<rT)rTrMN) rUrVrWrXrYrrZr^r\rrrr_rTr<r.rrsHPtx**26*OR*dlmpdq* **r<r)typingrrrrrr7r kornia.configr kornia.core.externalr Imager r kornia.core.mixin.onnxr kornia.models.baserkornia.models.processorsrrkornia.models.small_srrkornia.onnx.downloadr__all__rrrrrTr<r.rso$54 '2.2(Q41 Aqp#Ep w iw t< < ~..r<