rebeccali

rebeccali

from tensorflow.python.client import device_lib
print(device_lib.list_local_devices())

rebeccali

How to get linux to let me flash Sparkfun Pro Micro

In trying to get Redox Wireless flashed, you have to upload the firmware to your arduino pro micro with QMK firmware.

1. Plug in the micro to your computer, check that it is actually connected with a datacable by running ls /dev/tty* and seeing a new device appear.
2. Run make redox_W:default:avrdude
3. Quickly reset the pro micro by shorting GND and RST twice quickly
4. When it says waiting for /dev/whatever to be writable, in another terminal run sudo chmod 666 /dev/whatever so you have permission to write
5. it should flash now haha
6. You might have to add your user to the dialout group. Not sure if that mattered.

rebeccali

# Add some executable to the dmenu. Basically just adds a symlink so PATH can see it.
ln -s /path/to/my/script /usr/bin/

rebeccali

class DoubleDeepQNetwork(nn.Module):
    def __init__(self, xdim, udim, hdim=16):
        super().__init__()
        if torch.cuda.is_available() == True:
            self.dtype_double = torch.cuda.FloatTensor
            self.dtype_int = torch.cuda.LongTensor
        else:
            raise Exception('NO CUDA')
            self.dtype_double = torch.FloatTensor
            self.dtype_int = torch.LongTensor

rebeccali

class DoubleDeepQNetwork(nn.Module):
    def __init__(self, xdim, udim, hdim=16):
        super().__init__()
        if torch.cuda.is_available() == True:
            self.dtype_double = torch.cuda.FloatTensor
            self.dtype_int = torch.cuda.LongTensor
        else:
            raise Exception('NO CUDA')
            self.dtype_double = torch.FloatTensor
            self.dtype_int = torch.LongTensor

rebeccali

 
import datetime 

timestamp = datetime.datetime.now().strftime("%d-%m-%Y_%H-%M-%S")

rebeccali

import torch.autograd
import torch.nn as nn
from au_functional import jacobian
import torch
from torchdiffeq import odeint

true_y0 = torch.tensor([[2., 0.]], requires_grad=True)
data_size=5
t = torch.linspace(0., 25., data_size)
true_A = torch.tensor([[-0.1, 2.0], [-2.0, -0.1]])

rebeccali

 import rosbag
 
with rosbag.Bag(bagfile) as bag:
    start_time = bag.get_start_time()
    topic_name = "/mytopic/ddk"
    for (i, (topic, msg, t)) in enumerate(bag.read_messages(topics=[topic_name])):
            # Do some stuff with the bag 
            pose = msg.pose.pose.position
            positionx = pose.x 
            vel = msg.twist.twist.linear

rebeccali

How to install external ROS packages

1. First, you will need to source your overall ros installation. This enables the basic ROS tools (rospack, rosrun, etc.).

source /opt/ros/kinetic/setup.bash 

2. Source your catkin workspace. See http://wiki.ros.org/ROS/Tutorials/InstallingandConfiguringROSEnvironment if you do not know what this is. To my knowledge, this sets a bunch of environment variables.