how to use latest RIFE in AFI?

[2994dd42d75b0]

I copied all files from RIFE 4.18 and also change paths in .py files Still getting this error

Traceback (most recent call last): File "/AFI-ForwardDeduplicate/interpolate_video_forward.py", line 93, in model.load_state_dict(convert(torch.load('weights/flownet.pkl'))) File "/usr/local/lib/python3.10/dist-packages/torch/nn/modules/module.py", line 2189, in load_state_dict raise RuntimeError('Error(s) in loading state_dict for {}:\n\t{}'.format( RuntimeError: Error(s) in loading state_dict for IFNet: Unexpected key(s) in state_dict: "teacher.conv0.0.0.weight", "teacher.conv0.0.0.bias", "teacher.conv0.1.0.weight", "teacher.conv0.1.0.bias", "teacher.convblock.0.beta", "teacher.convblock.0.conv.weight", "teacher.convblock.0.conv.bias", "teacher.convblock.1.beta", "teacher.convblock.1.conv.weight", "teacher.convblock.1.conv.bias", "teacher.convblock.2.beta", "teacher.convblock.2.conv.weight", "teacher.convblock.2.conv.bias", "teacher.convblock.3.beta", "teacher.convblock.3.conv.weight", "teacher.convblock.3.conv.bias", "teacher.convblock.4.beta", "teacher.convblock.4.conv.weight", "teacher.convblock.4.conv.bias", "teacher.convblock.5.beta", "teacher.convblock.5.conv.weight", "teacher.convblock.5.conv.bias", "teacher.convblock.6.beta", "teacher.convblock.6.conv.weight", "teacher.convblock.6.conv.bias", "teacher.convblock.7.beta", "teacher.convblock.7.conv.weight", "teacher.convblock.7.conv.bias", "teacher.lastconv.0.weight", "teacher.lastconv.0.bias".

I renamed rife48.pkl to flownet.pkl Please tell how to fix this

[routineLife1]

try

model.load_state_dict(torch.load('weights/flownet.pkl'))

[2994dd42d75b0]

Getting This

interpolate_video_forward.py", line 94, in model.load_state_dict(torch.load('weights/flownet.pkl')) File "/usr/local/lib/python3.10/dist-packages/torch/nn/modules/module.py", line 2189, in load_state_dict raise RuntimeError('Error(s) in loading state_dict for {}:\n\t{}'.format( RuntimeError: Error(s) in loading state_dict for IFNet: Missing key(s) in state_dict: "block0.conv0.0.0.weight", "block0.conv0.0.0.bias", "block0.conv0.1.0.weight", "block0.conv0.1.0.bias", "block0.convblock.0.beta", "block0.convblock.0.conv.weight", "block0.convblock.0.conv.bias", "block0.convblock.1.beta", "block0.convblock.1.conv.weight", "block0.convblock.1.conv.bias", "block0.convblock.2.beta", "block0.convblock.2.conv.weight", "block0.convblock.2.conv.bias", "block0.convblock.3.beta", "block0.convblock.3.conv.weight", "block0.convblock.3.conv.bias", "block0.convblock.4.beta", "block0.convblock.4.conv.weight", "block0.convblock.4.conv.bias", "block0.convblock.5.beta", "block0.convblock.5.conv.weight", "block0.convblock.5.conv.bias", "block0.convblock.6.beta", "block0.convblock.6.conv.weight", "block0.convblock.6.conv.bias", "block0.convblock.7.beta", "block0.convblock.7.conv.weight", "block0.convblock.7.conv.bias", "block0.lastconv.0.weight", "block0.lastconv.0.bias", "block1.conv0.0.0.weight", "block1.conv0.0.0.bias", "block1.conv0.1.0.weight", "block1.conv0.1.0.bias", "block1.convblock.0.beta", "block1.convblock.0.conv.weight", "block1.convblock.0.conv.bias", "block1.convblock.1.beta", "block1.convblock.1.conv.weight", "block1.convblock.1.conv.bias", "block1.convblock.2.beta", "block1.convblock.2.conv.weight", "block1.convblock.2.conv.bias", "block1.convblock.3.beta", "block1.convblock.3.conv.weight", "block1.convblock.3.conv.bias", "block1.convblock.4.beta", "block1.convblock.4.conv.weight", "block1.convblock.4.conv.bias", "block1.convblock.5.beta", "block1.convblock.5.conv.weight", "block1.convblock.5.conv.bias", "block1.convblock.6.beta", "block1.convblock.6.conv.weight", "block1.convblock.6.conv.bias", "block1.convblock.7.beta", "block1.convblock.7.conv.weight", "block1.convblock.7.conv.bias", "block1.lastconv.0.weight", "block1.lastconv.0.bias", "block2.conv0.0.0.weight", "block2.conv0.0.0.bias", "block2.conv0.1.0.weight", "block2.conv0.1.0.bias", "block2.convblock.0.beta", "block2.convblock.0.conv.weight", "block2.convblock.0.conv.bias", "block2.convblock.1.beta", "block2.convblock.1.conv.weight", "block2.convblock.1.conv.bias", "block2.convblock.2.beta", "block2.convblock.2.conv.weight", "block2.convblock.2.conv.bias", "block2.convblock.3.beta", "block2.convblock.3.conv.weight", "block2.convblock.3.conv.bias", "block2.convblock.4.beta", "block2.convblock.4.conv.weight", "block2.convblock.4.conv.bias", "block2.convblock.5.beta", "block2.convblock.5.conv.weight", "block2.convblock.5.conv.bias", "block2.convblock.6.beta", "block2.convblock.6.conv.weight", "block2.convblock.6.conv.bias", "block2.convblock.7.beta", "block2.convblock.7.conv.weight", "block2.convblock.7.conv.bias", "block2.lastconv.0.weight", "block2.lastconv.0.bias", "block3.conv0.0.0.weight", "block3.conv0.0.0.bias", "block3.conv0.1.0.weight", "block3.conv0.1.0.bias", "block3.convblock.0.beta", "block3.convblock.0.conv.weight", "block3.convblock.0.conv.bias", "block3.convblock.1.beta", "block3.convblock.1.conv.weight", "block3.convblock.1.conv.bias", "block3.convblock.2.beta", "block3.convblock.2.conv.weight", "block3.convblock.2.conv.bias", "block3.convblock.3.beta", "block3.convblock.3.conv.weight", "block3.convblock.3.conv.bias", "block3.convblock.4.beta", "block3.convblock.4.conv.weight", "block3.convblock.4.conv.bias", "block3.convblock.5.beta", "block3.convblock.5.conv.weight", "block3.convblock.5.conv.bias", "block3.convblock.6.beta", "block3.convblock.6.conv.weight", "block3.convblock.6.conv.bias", "block3.convblock.7.beta", "block3.convblock.7.conv.weight", "block3.convblock.7.conv.bias", "block3.lastconv.0.weight", "block3.lastconv.0.bias", "encode.cnn0.weight", "encode.cnn0.bias", "encode.cnn1.weight", "encode.cnn1.bias", "encode.cnn2.weight", "encode.cnn2.bias", "encode.cnn3.weight", "encode.cnn3.bias". Unexpected key(s) in state_dict: "module.block0.conv0.0.0.weight", "module.block0.conv0.0.0.bias", "module.block0.conv0.1.0.weight", "module.block0.conv0.1.0.bias", "module.block0.convblock.0.beta", "module.block0.convblock.0.conv.weight", "module.block0.convblock.0.conv.bias", "module.block0.convblock.1.beta", "module.block0.convblock.1.conv.weight", "module.block0.convblock.1.conv.bias", "module.block0.convblock.2.beta", "module.block0.convblock.2.conv.weight", "module.block0.convblock.2.conv.bias", "module.block0.convblock.3.beta", "module.block0.convblock.3.conv.weight", "module.block0.convblock.3.conv.bias", "module.block0.convblock.4.beta", "module.block0.convblock.4.conv.weight", "module.block0.convblock.4.conv.bias", "module.block0.convblock.5.beta", "module.block0.convblock.5.conv.weight", "module.block0.convblock.5.conv.bias", "module.block0.convblock.6.beta", "module.block0.convblock.6.conv.weight", "module.block0.convblock.6.conv.bias", "module.block0.convblock.7.beta", "module.block0.convblock.7.conv.weight", "module.block0.convblock.7.conv.bias", "module.block0.lastconv.0.weight", "module.block0.lastconv.0.bias", "module.block1.conv0.0.0.weight", "module.block1.conv0.0.0.bias", "module.block1.conv0.1.0.weight", "module.block1.conv0.1.0.bias", "module.block1.convblock.0.beta", "module.block1.convblock.0.conv.weight", "module.block1.convblock.0.conv.bias", "module.block1.convblock.1.beta", "module.block1.convblock.1.conv.weight", "module.block1.convblock.1.conv.bias", "module.block1.convblock.2.beta", "module.block1.convblock.2.conv.weight", "module.block1.convblock.2.conv.bias", "module.block1.convblock.3.beta", "module.block1.convblock.3.conv.weight", "module.block1.convblock.3.conv.bias", "module.block1.convblock.4.beta", "module.block1.convblock.4.conv.weight", "module.block1.convblock.4.conv.bias", "module.block1.convblock.5.beta", "module.block1.convblock.5.conv.weight", "module.block1.convblock.5.conv.bias", "module.block1.convblock.6.beta", "module.block1.convblock.6.conv.weight", "module.block1.convblock.6.conv.bias", "module.block1.convblock.7.beta", "module.block1.convblock.7.conv.weight", "module.block1.convblock.7.conv.bias", "module.block1.lastconv.0.weight", "module.block1.lastconv.0.bias", "module.block2.conv0.0.0.weight", "module.block2.conv0.0.0.bias", "module.block2.conv0.1.0.weight", "module.block2.conv0.1.0.bias", "module.block2.convblock.0.beta", "module.block2.convblock.0.conv.weight", "module.block2.convblock.0.conv.bias", "module.block2.convblock.1.beta", "module.block2.convblock.1.conv.weight", "module.block2.convblock.1.conv.bias", "module.block2.convblock.2.beta", "module.block2.convblock.2.conv.weight", "module.block2.convblock.2.conv.bias", "module.block2.convblock.3.beta", "module.block2.convblock.3.conv.weight", "module.block2.convblock.3.conv.bias", "module.block2.convblock.4.beta", "module.block2.convblock.4.conv.weight", "module.block2.convblock.4.conv.bias", "module.block2.convblock.5.beta", "module.block2.convblock.5.conv.weight", "module.block2.convblock.5.conv.bias", "module.block2.convblock.6.beta", "module.block2.convblock.6.conv.weight", "module.block2.convblock.6.conv.bias", "module.block2.convblock.7.beta", "module.block2.convblock.7.conv.weight", "module.block2.convblock.7.conv.bias", "module.block2.lastconv.0.weight", "module.block2.lastconv.0.bias", "module.block3.conv0.0.0.weight", "module.block3.conv0.0.0.bias", "module.block3.conv0.1.0.weight", "module.block3.conv0.1.0.bias", "module.block3.convblock.0.beta", "module.block3.convblock.0.conv.weight", "module.block3.convblock.0.conv.bias", "module.block3.convblock.1.beta", "module.block3.convblock.1.conv.weight", "module.block3.convblock.1.conv.bias", "module.block3.convblock.2.beta", "module.block3.convblock.2.conv.weight", "module.block3.convblock.2.conv.bias", "module.block3.convblock.3.beta", "module.block3.convblock.3.conv.weight", "module.block3.convblock.3.conv.bias", "module.block3.convblock.4.beta", "module.block3.convblock.4.conv.weight", "module.block3.convblock.4.conv.bias", "module.block3.convblock.5.beta", "module.block3.convblock.5.conv.weight", "module.block3.convblock.5.conv.bias", "module.block3.convblock.6.beta", "module.block3.convblock.6.conv.weight", "module.block3.convblock.6.conv.bias", "module.block3.convblock.7.beta", "module.block3.convblock.7.conv.weight", "module.block3.convblock.7.conv.bias", "module.block3.lastconv.0.weight", "module.block3.lastconv.0.bias", "module.encode.cnn0.weight", "module.encode.cnn0.bias", "module.encode.cnn1.weight", "module.encode.cnn1.bias", "module.encode.cnn2.weight", "module.encode.cnn2.bias", "module.encode.cnn3.weight", "module.encode.cnn3.bias", "module.teacher.conv0.0.0.weight", "module.teacher.conv0.0.0.bias", "module.teacher.conv0.1.0.weight", "module.teacher.conv0.1.0.bias", "module.teacher.convblock.0.beta", "module.teacher.convblock.0.conv.weight", "module.teacher.convblock.0.conv.bias", "module.teacher.convblock.1.beta", "module.teacher.convblock.1.conv.weight", "module.teacher.convblock.1.conv.bias", "module.teacher.convblock.2.beta", "module.teacher.convblock.2.conv.weight", "module.teacher.convblock.2.conv.bias", "module.teacher.convblock.3.beta", "module.teacher.convblock.3.conv.weight", "module.teacher.convblock.3.conv.bias", "module.teacher.convblock.4.beta", "module.teacher.convblock.4.conv.weight", "module.teacher.convblock.4.conv.bias", "module.teacher.convblock.5.beta", "module.teacher.convblock.5.conv.weight", "module.teacher.convblock.5.conv.bias", "module.teacher.convblock.6.beta", "module.teacher.convblock.6.conv.weight", "module.teacher.convblock.6.conv.bias", "module.teacher.convblock.7.beta", "module.teacher.convblock.7.conv.weight", "module.teacher.convblock.7.conv.bias", "module.teacher.lastconv.0.weight", "module.teacher.lastconv.0.bias".

[routineLife1]

model.load_state_dict(convert(torch.load('weights/flownet.pkl')),strict=false)

[2994dd42d75b0]

Loaded model 0% 0/245.0 [00:00<?, ?it/s]Traceback (most recent call last): File "/AFI-ForwardDeduplicate/interpolate_video_forward.py", line 270, in output0, count = decrease_inference(queue_input.copy()) File "/AFI-ForwardDeduplicate/interpolate_video_forward.py", line 191, in decrease_inference to_numpy(make_inf(inp0, inp1, scale, 0.5)) File "/AFI-ForwardDeduplicate/interpolate_video_forward.py", line 118, in to_numpy return tensor.squeeze(0).permute(1, 2, 0).cpu().numpy() * 255. AttributeError: 'tuple' object has no attribute 'squeeze' FATAL: exception not rethrown

Same error Using RIFE 4.18 to 4.22

[routineLife1]

The return statement in the forward function of IFNet_HDv3.py should be changed to return merged[3]. This means that only the interpolated frame result is needed, not 'tuple'.

[2994dd42d75b0]

Loaded model 0% 0/143.0 [00:00<?, ?it/s]Traceback (most recent call last): File "/interpolate_video_forward_anyfps.py", line 278, in output0, count = decrease_inference(queue_input.copy()) File "/interpolate_video_forward_anyfps.py", line 207, in decrease_inference to_numpy(make_inf(inp0, inp1, scale, 0.5)) File "/interpolate_video_forward_anyfps.py", line 133, in to_numpy return tensor.squeeze(0).permute(1, 2, 0).cpu().numpy() * 255. AttributeError: 'tuple' object has no attribute 'squeeze' terminate called without an active exception

[routineLife1]

IFNet_HDv3.py

    def forward(self, x, timestep=0.5, scale_list=[8, 4, 2, 1], training=False, fastmode=True, ensemble=False):
        if training == False:
            channel = x.shape[1] // 2
            img0 = x[:, :channel]
            img1 = x[:, channel:]
        if not torch.is_tensor(timestep):
            timestep = (x[:, :1].clone() * 0 + 1) * timestep
        else:
            timestep = timestep.repeat(1, 1, img0.shape[2], img0.shape[3])
        f0 = self.encode(img0[:, :3])
        f1 = self.encode(img1[:, :3])
        flow_list = []
        merged = []
        mask_list = []
        warped_img0 = img0
        warped_img1 = img1
        flow = None
        mask = None
        loss_cons = 0
        block = [self.block0, self.block1, self.block2, self.block3]
        for i in range(4):
            if flow is None:
                flow, mask, feat = block[i](torch.cat((img0[:, :3], img1[:, :3], f0, f1, timestep), 1), None,
                                            scale=scale_list[i])
                if ensemble:
                    print("warning: ensemble is not supported since RIFEv4.21")
            else:
                wf0 = warp(f0, flow[:, :2])
                wf1 = warp(f1, flow[:, 2:4])
                fd, m0, feat = block[i](
                    torch.cat((warped_img0[:, :3], warped_img1[:, :3], wf0, wf1, timestep, mask, feat), 1), flow,
                    scale=scale_list[i])
                if ensemble:
                    print("warning: ensemble is not supported since RIFEv4.21")
                else:
                    mask = m0
                flow = flow + fd
            mask_list.append(mask)
            flow_list.append(flow)
            warped_img0 = warp(img0, flow[:, :2])
            warped_img1 = warp(img1, flow[:, 2:4])
            merged.append((warped_img0, warped_img1))
        mask = torch.sigmoid(mask)
        merged[3] = (warped_img0 * mask + warped_img1 * (1 - mask))
        return merged[3]

[2994dd42d75b0]

Working Now Perfectly

Thanks