My Deep Learning Laptop (in progress)

Hardware

I purchased a new laptop. (My eight year old Dell Windoze 10 (originally Windoze 8) laptop would not load Deep Learning Applications)

• CUK HP Omen 17t Gaming Notebook (Intel i7-10750H Six Core Processor 12MB Cache, 2.6GHz-5.0GHz, 45W), 32GB RAM, 512GB NVMe SSD + 1TB HDD, NVIDIA GeForce RTX 2080 Super 8GB, 17.3" FHD 300Hz IPS, Windows 10 Home) Gamer Laptop Computer.
• note: I did not spend extra for 4k display laptop screen display will use 4k HDMI out

I purchased a monitor. (My ten year old Dell monitor was getting blurry)

• Sceptre 35 Inch Curved UltraWide 21: 9 LED Creative Monitor QHD 3440x1440 Frameless AMD Freesync HDMI DisplayPort Up to 100Hz, Machine Black 2020 (C355W-3440UN)

           (note: Hoping Laptop screen HDMI1 4k 100Hz output supports 3440x1440)

Software

Windoze 10 Home

Steps

1. Booted up Laptop - answered questions - setup Chrome as default browser - setup Google account - update and shutdown <done>

2. Attach Monitor, Keyboard/Mouse to Laptop and Power On. <done>

3. Investigate Deep Learning Software to load. (Anaconda, CUDA, TensorFlow, PyTorch, Keras,...) <done>

4. Download Anaconda3 Individual Edition Windows 64 via link https://anaconda.com/download <done>

    (includes Numpy, Scipy, MatPlotLib, Pandas, NLTK, SKlearn)

    (can still be updated i.e. conda install numpy,...)

    Open Anaconda3 Command Prompt type ipython at prompt (If you want to test)
    import numpy as np
    import matplotlib.pyplot as plt
    a = np.random.randn(1000)
    plt.plot(a)
    plt.show()
    plt.hist(a, bins = 15)
    plt.show()

5. Install Deep Learning Libraries: Tensorflow, Keras, CNTL, Pytorch, Theano 

   Open Anaconda3 Command Prompt type ipython at prompt (If you haven't already)

   pip install tensorflow <done>

   pip install Keras (but first: conda install pip) <done> 

   pip install nltk <done>

   pip install theano <done>

   pip install http://cntk.ai/...whl (I need to google Microsoft CNTK python)

   pip install https://cntk.ai/PythonWheel/GPU/cntk_gpu-2.7.post1-cp36-cp36m-win_amd64.whl (chose for NVIDEA GPU)

   In [3]: pip install https://cntk.ai/PythonWheel/GPU/cntk_gpu-2.7.post1-cp36-cp36m-win_amd64.whl
   ERROR: cntk_gpu-2.7.post1-cp36-cp36m-win_amd64.whl is not a supported wheel on this platform.
   Note: you may need to restart the kernel to use updated packages.

   pip install -c peterjc123 pytorch <did not work>

   Error:

   In [17]: pip install -c peterjc123 pytorch
   ERROR: Could not open requirements file: [Errno 2] No such file or directory: 'peterjc123'
   Note: you may need to restart the kernel to use updated packages.

   Suggested Fix:   

   Anaconda Navigator - environments → channels → add "conda-forge" update channels (did not work but should be done anyway) 

   <Break>

6. NVIDEA GPU and Theano

    Installed Conda Toolkit already

    conda install theano <left bracket> pygpu ]

7. Open AI Gym

   Open Anaconda3 Command Prompt type ipython at prompt (If you haven't already)

   pip install gym

   conda install -c menpo ffmpeg (if you want to play and save videos with AI gym)

    pip