Op Support: aten.fill.Tensor

ROCm / torch_migraphx

Libraries integrating migraphx with pytorch

BSD 3-Clause "New" or "Revised" License

4 stars 0 forks source link

Op Support: aten.fill.Tensor #116

Open shivadbhavsar opened 1 month ago

shivadbhavsar commented 1 month ago

Add aten converter. Affected models: Torchbench: pyhpc_turbulent_kinetic_energy

CharlieL7 commented 3 weeks ago

MIGraphX does not yet have a GPU fill operator. Would need to implement it to complete this issue.

shivadbhavsar commented 3 weeks ago

We might be able to partially resolve this in cases where the fill input is a constant. There are some passes that do this for torch.zeros, torch.full, torch.arange etc