Add GIMM-VFI support (NeurIPS 2024) with single-pass arbitrary-timestep interpolation

Integrates GIMM-VFI alongside existing BIM/EMA/SGM models. Key feature: generates
all intermediate frames in one forward pass (no recursive 2x passes needed for 4x/8x).

- Vendor gimm_vfi_arch/ from kijai/ComfyUI-GIMM-VFI with device fixes
- Two variants: RAFT-based (~80MB) and FlowFormer-based (~123MB)
- Auto-download checkpoints from HuggingFace (Kijai/GIMM-VFI_safetensors)
- Three new nodes: Load GIMM-VFI Model, GIMM-VFI Interpolate, GIMM-VFI Segment Interpolate
- single_pass toggle: True=arbitrary timestep (default), False=recursive like other models
- ds_factor parameter for high-res input downscaling

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

gimm_vfi_arch/generalizable_INR/modules/layers.py

@@ -0,0 +1,42 @@
+# Copyright (c) Meta Platforms, Inc. and affiliates.
+# All rights reserved.
+
+# This source code is licensed under the license found in the
+# LICENSE file in the root directory of this source tree.
+# --------------------------------------------------------
+
+from torch import nn
+import torch
+
+
+# define siren layer & Siren model
+class Sine(nn.Module):
+    """Sine activation with scaling.
+
+    Args:
+        w0 (float): Omega_0 parameter from SIREN paper.
+    """
+
+    def __init__(self, w0=1.0):
+        super().__init__()
+        self.w0 = w0
+
+    def forward(self, x):
+        return torch.sin(self.w0 * x)
+
+
+# Damping activation from http://arxiv.org/abs/2306.15242
+class Damping(nn.Module):
+    """Sine activation with sublinear factor
+
+    Args:
+        w0 (float): Omega_0 parameter from SIREN paper.
+    """
+
+    def __init__(self, w0=1.0):
+        super().__init__()
+        self.w0 = w0
+
+    def forward(self, x):
+        x = torch.clamp(x, min=1e-30)
+        return torch.sin(self.w0 * x) * torch.sqrt(x.abs())