thipokKub

As of pytorch-nightly build `1.13.0.dev20220522` - CPU `18.835263124999983` s - MPS `18.363079624999955` s

I haven't tested my code but this should do the trick ```python class PlanarFlow(nn.Module): def __init__(self, dim): super().__init__() self.weight = nn.Parameter(torch.Tensor(1, dim)) self.scale = nn.Parameter(torch.Tensor(1, dim)) self.bias = nn.Parameter(torch.Tensor(1)) self.tanh...

Not supported - `aten::l1_loss_backward.grad_input` - `aten::kl_div_backward` Code ```python X, y = torch.rand(16, 10).to("mps"), torch.rand(16, 1).to("mps") model = nn.Linear(10, 1).to("mps") criterion = nn.L1Loss() # nn.KLDivLoss() loss = criterion(model(X), y) loss.backward() ```...

I'm not sure, I don't know how exactly on a couple things - How to add MMM model to [CausalPy](https://github.com/pymc-labs/CausalPy), but likely to do with [this script template](https://github.com/pymc-labs/CausalPy/blob/main/causalpy/pymc_models.py) - Not...

Sorry, I've got a bit off-track. What I was trying to do is finding a repeatable testing, and calibration pipeline for MMM model, which is not exactly what I asked....

Hello! Thanks for the reply To use `df['trend']` as a control feature, we have to transform it beforehand (not learnable). It might be more desirable to learn it from data...

You are right, the saturation effect is a bit messy. Considering the adstock effect is the only effect in the model that carry past state information into the future, and...

Okay, I think I've got what lightweight MMM do for its response curve prediction. From what I've found, they **did not** account for adstock lag effect, and only calculate its...

Another thing I did is to inspect response/incremental response curve relationship to the offset (previous state). As shown in the graph below, we can see that the response curve change...

I've ported [uber/orbit](https://github.com/uber/orbit) like model in for time-varying coefficient in pymc. Here are MVP snippets ```python # From: https://github.com/uber/orbit/blob/dev/orbit/utils/kernels.py def gaussian_kernel(t, t_refs, rho: float = 0.1, alpha: float = 1,...