refactor: integrate sharpfin for high-quality image resize

Vendor sharpfin library (Apache 2.0) and add centralized wrapper
module (images_sharpfin.py) replacing torchvision tensor/PIL
conversion and resize operations throughout the codebase.

- Add modules/sharpfin/ vendored library with MKS2021, Lanczos3,
  Mitchell, Catmull-Rom kernels and optional Triton sparse acceleration
- Add modules/images_sharpfin.py wrapper with to_tensor(), to_pil(),
  pil_to_tensor(), normalize(), resize(), resize_tensor()
- Add resize_quality and resize_linearize_srgb settings
- Add MKS2021 and Lanczos3 upscaler entries
- Replace torchvision.transforms.functional imports across 18 files
- to_pil() auto-detects HWC/BHWC layout, adds .round() before uint8
- Sparse Triton path falls back to dense GPU on compilation failure
- Mixed-axis resize splits into two single-axis scale() calls
- Masks and non-sRGB data always use linearize=False

modules/processing_vae.py

@@ -334,7 +334,7 @@ def vae_decode(latents, model, output_type='np', vae_type='Full', width=None, he
 
 def vae_encode(image, model, vae_type='Full'): # pylint: disable=unused-variable
     jobid = shared.state.begin('VAE Encode')
-    import torchvision.transforms.functional as f
+    from modules import images_sharpfin
     if shared.state.interrupted or shared.state.skipped:
         return []
     if not hasattr(model, 'vae') and hasattr(model, 'pipe'):
@@ -342,7 +342,7 @@ def vae_encode(image, model, vae_type='Full'): # pylint: disable=unused-variable
     if not hasattr(model, 'vae'):
         shared.log.error('VAE not found in model')
         return []
-    tensor = f.to_tensor(image.convert("RGB")).unsqueeze(0).to(devices.device, devices.dtype_vae)
+    tensor = images_sharpfin.to_tensor(image.convert("RGB")).unsqueeze(0).to(devices.device, devices.dtype_vae)
     if vae_type == 'Tiny':
         latents = taesd_vae_encode(image=tensor)
     elif vae_type == 'Full' and hasattr(model, 'vae'):