Abhinav Moudgil amoudgl

## hello-gpt.py
"""
Adapted from modula's Hello GPT tutorial:
https://github.com/modula-systems/modula/blob/ede2ba72a1b9de3e1f44156db058b5c32c682941/examples/hello-gpt.ipynb

This script simply exposes dataloader seed as command line argument to test
training sensitivity to seed.

Usage:

python hello-gpt.py --seed 0

## output.log
Input shape: (12, 64)
Target shape: (12, 64)
First input sequence: [41 53 50 42  1 40 50 53 53 42] ...
First target sequence: [53 50 42  1 40 50 53 53 42  1] ...

Decoded input: cold blood no spark of honour bides.

NORTHUMBERLAND:
Be thou a

## hello-gpt-nan.ipynb

      
              1 file
            
          
              0 forks
            
          
                0 comments
              
            
              0 stars
            
          
                amoudgl
                / hello-gpt-nan.ipynb
            
            
              Last active
              May 30, 2025 17:56
            
              
                NaNs in modula hello GPT tutorial 
              
          
      Loading

      Sorry, something went wrong. Reload?
      Sorry, we cannot display this file.
      Sorry, this file is invalid so it cannot be displayed.
      
          Viewer requires iframe.
      
    
## tx_lopt.py
# Copyright 2021 DeepMind Technologies Limited. All Rights Reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#     http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,

## dummyfig.tex
%% This part goes in preamble
\newcommand{\dummyfig}[1]{
  \centering
  \fbox{
    \begin{minipage}[c][0.33\textheight][c]{0.5\textwidth}
      \centering{#1}
    \end{minipage}
  }
}

## trajectory_visualization_v2.ipynb

      
              1 file
            
          
              0 forks
            
          
                0 comments
              
            
              0 stars
            
          
                amoudgl
                / trajectory_visualization_v2.ipynb
            
            
              Created
              May 9, 2021 18:24
                — forked from ronghanghu/trajectory_visualization_v2.ipynb
            
              
                Speaker-Follower visualization
              
          
      Loading

      Sorry, something went wrong. Reload?
      Sorry, we cannot display this file.
      Sorry, this file is invalid so it cannot be displayed.
      
          Viewer requires iframe.
      
    
## setup_visdom.md

      
              1 file
            
          
              1 fork
            
          
                8 comments
              
            
              12 stars
            
          
                amoudgl
                / setup_visdom.md
            
            
              Last active
              July 4, 2023 11:27
            
              
                Setup Visdom on Remote Server
              
          
    Setup Visdom on Remote Server

Install visdom on your local system and remote server.
pip3 install visdom

On remote server, do:

  
## README.md

      
              1 file
            
          
              3 forks
            
          
                0 comments
              
            
              6 stars
            
          
                amoudgl
                / README.md
            
            
              Last active
              April 17, 2024 06:25
            
              
                TLP benchmark
              
          
    Checkout the lastest versions of our datasets!

Following are the higher quality versions of TLP and TinyTLP datasets. The resolution is still the same but images are much sharper and have higher quality.
TLP V2 (87.2 GB)

https://drive.google.com/open?id=1mv0ULctCzGn4gzum_Sb6kwXwB2QjZwu1
TLP V2 Sequences


## deep-qlearning.py
# simple neural network implementation of qlearning
import gym
from gym import wrappers
import numpy as np
import tensorflow as tf

# build environment
env = gym.make("FrozenLake-v0")
env = wrappers.Monitor(env, '/tmp/frozenlake-qlearning', force=True)
n_obv = env.observation_space.n

## cartpolev1_pg.py
# monte carlo policy gradient algorithm
# use neural network to decide the policy
# from observations and rewards, update the parameters of the neural networks to optimize the policy

import numpy as np
import tensorflow as tf
import gym
from gym import wrappers

# initialize constants
	"""
	Adapted from modula's Hello GPT tutorial:
	https://github.com/modula-systems/modula/blob/ede2ba72a1b9de3e1f44156db058b5c32c682941/examples/hello-gpt.ipynb

	This script simply exposes dataloader seed as command line argument to test
	training sensitivity to seed.

	Usage:

	python hello-gpt.py --seed 0
	Input shape: (12, 64)
	Target shape: (12, 64)
	First input sequence: [41 53 50 42 1 40 50 53 53 42] ...
	First target sequence: [53 50 42 1 40 50 53 53 42 1] ...

	Decoded input: cold blood no spark of honour bides.

	NORTHUMBERLAND:
	Be thou a
	# Copyright 2021 DeepMind Technologies Limited. All Rights Reserved.
	#
	# Licensed under the Apache License, Version 2.0 (the "License");
	# you may not use this file except in compliance with the License.
	# You may obtain a copy of the License at
	#
	# http://www.apache.org/licenses/LICENSE-2.0
	#
	# Unless required by applicable law or agreed to in writing, software
	# distributed under the License is distributed on an "AS IS" BASIS,
	%% This part goes in preamble
	\newcommand{\dummyfig}[1]{
	\centering
	\fbox{
	\begin{minipage}[c][0.33\textheight][c]{0.5\textwidth}
	\centering{#1}
	\end{minipage}
	}
	}
	# simple neural network implementation of qlearning
	import gym
	from gym import wrappers
	import numpy as np
	import tensorflow as tf

	# build environment
	env = gym.make("FrozenLake-v0")
	env = wrappers.Monitor(env, '/tmp/frozenlake-qlearning', force=True)
	n_obv = env.observation_space.n
	# monte carlo policy gradient algorithm
	# use neural network to decide the policy
	# from observations and rewards, update the parameters of the neural networks to optimize the policy

	import numpy as np
	import tensorflow as tf
	import gym
	from gym import wrappers

	# initialize constants