Skip to content

Instantly share code, notes, and snippets.

View dsuess's full-sized avatar
🤔
...

Daniel Suess dsuess

🤔
...
View GitHub Profile
@dsuess
dsuess / setup.sh
Last active February 2, 2021 02:46
Lifecycle configuration to setup vim keybindings in SageMaker
source /home/ec2-user/anaconda3/bin/activate /home/ec2-user/anaconda3/envs/JupyterSystemEnv
jupyter labextension install jupyterlab_vim
pip install jupyter_contrib_nbextensions
sudo -H -u ec2-user bash -c "source /home/ec2-user/anaconda3/bin/activate /home/ec2-user/anaconda3/envs/JupyterSystemEnv && jupyter-contrib nbextension install --user && jupyter-nbextension enable select_keymap/main"
@dsuess
dsuess / create_function.py
Last active October 24, 2022 09:35 — forked from dhagrow/create_function.py
Dynamic function creation in Python
"""
Python is a dynamic language, and it is relatively easy to dynamically create
and modify things such as classes and objects. Functions, however, are quite
challenging to create dynamically.
One area where we might want to do this is in an RPC library, where a function
defined on a server needs to be available remotely on a client.
The naive solution is to simply pass arguments to a generic function that
accepts `*args` and `**kwargs`. A lot of information is lost with this approach,
however, in particular the number of arguments taken. Used in an RPC
implementation, this also delays any error feedback until after the arguments
import numpy as np
import torch
from torch import nn
import onnx_tensorrt.backend as backend
import onnx
class Model(nn.Module):
def forward(self, x):
y = (2 * x)[0:1]
@dsuess
dsuess / onnx_tensorrt.py
Last active March 28, 2019 23:55
An example of a failed import of a (trivial) ONNX model in TensorRT
import torch
from torch import nn
import tensorrt as trt
TRT_LOGGER = trt.Logger(trt.Logger.INFO)
class Model(nn.Module):
def forward(self, x):
y = (2 * x)[0:1]
#include <iostream>
#include "xtensor/xio.hpp"
#include "xtensor/xrandom.hpp"
#include "xtensor/xtensor.hpp"
#include "xtensor/xview.hpp"
template <class E>
auto transform(E& x) {
x *= 2;
#include <iostream>
#include "xtensor/xio.hpp"
#include "xtensor/xnpy.hpp"
#include "xtensor/xrandom.hpp"
#include "xtensor/xtensor.hpp"
#include "xtensor/xview.hpp"
using farray = xt::xtensor<float, 2, xt::layout_type::row_major>;
@dsuess
dsuess / Dockerfile
Last active July 31, 2024 10:11
xeus-cling C++ Jupyter kernel inside a docker container
FROM frolvlad/alpine-miniconda3
RUN conda install -y -c conda-forge bash jupyter jupyter_contrib_nbextensions
RUN conda install -y -c conda-forge xeus-cling xtensor
RUN mkdir /notebooks
@dsuess
dsuess / Dockerfile
Created March 7, 2019 05:30
Dockerfile for nnstreamer
FROM ubuntu:18.04
RUN apt-get update && \
apt-get install -y software-properties-common meson libgst-dev \
libgstreamer1.0-dev libgstreamer-plugins-base1.0-dev libglib2.0-dev \
libcairo2-dev locales gstreamer1.0-tools gstreamer1.0-plugins-good \
gstreamer1.0-plugins-bad gstreamer1.0-plugins-ugly git
RUN add-apt-repository ppa:nnstreamer/ppa && \
apt-get update && \
@dsuess
dsuess / Untitled.ipynb
Created January 22, 2019 23:36
VideoReader speed test
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
class VideoWriter:
def __init__(self, output_path, overwrite_output=False):
self.output_path = output_path
self.active = False
self._stream = None
self._shape = None
self.overwrite_output = overwrite_output
def __enter__(self):