Ben Harper

CUDA memory leak with PyTorch 0.4.0

4

I've just upgraded to PyTorch 0.4.0 (built from source, on v0.4.0 branch), and rebuilt warp-ctc. Now I'm trying to train some CTC based models, and I get steady memory leaks...

Add an article link to the `GORM Community` page describing how to use strongly typed JSON fields with GORM

k8s StatefulSet needs service name override

3

### What happened? When creating a `StatefulSet`, `pulumi up` hangs with "Finding Pods to direct traffic to". If I override the pulumi-generated name of the Headless Service, then everything works....

Add support for prepared statements to "transaction" mode

4

Before doing any more work on this PR, I'd like to just sense-check if there's some glaring issue that I'm missing here. We've been running this in beta production for...

Any stable build combination?

12

Hi, I've tried many variations today, trying to get this to build, but all of them fail, and all with different errors. I've tried several mapnik branches, including [master, v3.0.12,...

Reshape operator expects Shape(tuple)

5

`Version: onnx-mxnet (0.4.2)` Breaking file: [sequence-tiny.zip](https://github.com/onnx/onnx-mxnet/files/1958549/sequence-tiny.zip) I'm getting the following error when trying to import my ONNX model: > Invalid Parameter format for shape expect Shape(tuple) but value='', in operator...

Motivation to automatically remove annex-b?

1

Here's a reference to the code I'm referring to: https://github.com/bluenviron/gortsplib/blob/a99cc55082b05bf50ddd7a2c68b0ba070b9764b5/pkg/format/rtph264/decoder.go#L233 I have two questions: 1. Firstly, on line 233, why `bytes.Contains(nalu, []byte{0x00, 0x00, 0x00, 0x01})` instead of `bytes.HasPrefix`? The `Contains`...