AI Coding Assistant for Enterprise Teams: Guide to Code Llama2 70B with Visual Studio Code Integration on E2E Cloud:

February 16, 2024

What Is Code Llama2

Meta recently upgraded their coding assistant called Code Llama to a new version named Code Llama 70B. Trained on a massive 1TB of code data, this model performs much better than previous ones, making it great for developers who need help with writing and fixing code. 

The model comes in three different versions - CodeLlama-70b, CodeLlama-70b-Python, and CodeLlama-70b-Instruct and is available for free on Hugging Face. CodeLlama-70b-Python was additionally trained on 100 billion tokens of Python Code, making it more fluent in generating Python Code. 

On the other hand, CodeLlama-70b-Instruct is fine-tuned to understand natural language queries and generate code accordingly. On a coding benchmark Code Llama performs equally well as other top models like GPT-4 and Gemini Pro.

Code Llama is part of a group of advanced AI tools that have come to be known as AI Coding Assistants. These tools include StarCoder, a very capable coding AI, and provide a better alternative to GPT-4. There's also CodeGen, trained on a mix of regular and coding language. 

The great thing about Code Llama 70B is that companies can use it in their own cloud infrastructure, instead of using proprietary AI Coding Assistants. This helps them keep their code, their IP, and their ideas safe. Essentially, hosting one’s own AI Coding Assistant ensures data privacy and IP protection, while also being far more cost effective in the long run for mid- to large-sized teams.

Why Deploy AI Coding Assistant on Private Cloud

By hosting coding assistants like Code Llama 70B on their own cloud infrastructure, companies do not just get an extremely powerful AI coding assistant for their teams, but they get several advantages that proprietary AI coding assistants do not offer:

  1. Data Security and Privacy

Companies often work with sensitive or proprietary information, and keeping the coding assistant in a cloud infrastructure controlled by the company ensures that this data does not leave the company's controlled environment.

  1. Customization and Control

Private cloud hosting allows companies to customize the model to their specific needs and workflows. They can tailor the AI's training to their own codebase, standards, and practices, which might not be possible with a public or shared model.

  1. Compliance and Regulatory Requirements

Most industries are subject to strict data protection and privacy regulations, and data localization requirements. Hosting on a private cloud can make it easier to comply with these regulations.

  1. Performance and Reliability

Companies can optimize the performance of the AI model for their specific needs in a private cloud. They can ensure that the system has the necessary resources and is not affected by other users, as might be the case in a public cloud.

  1. Integration with Internal Systems

It's often easier to integrate private cloud services with existing internal IT systems and workflows, ensuring seamless operation and reducing the risk of compatibility issues.

Cost Comparison - Unit Economics on E2E Network’s 4*H100 GPUs

E2E Networks offers state-of-the-art 4*H100 GPUs. These GPUs are recognized for their groundbreaking capabilities in accelerating trillion-parameter AI models. 

As of now, E2E offers 4*H100 GPUs for a monthly rent of INR 600,000. This would be the total cost for any organization to host Code Llama on their private cloud servers. 

On the other hand, if an organization were to use other commercially available coding assistants, say for example Github Copilot, the cost would be billed at $19 per month for individual users. This means anything upwards of 380 developers the company would start paying more than what it would have incurred with the privately hosted Code Llama. 

Therefore, if any organization has upwards of 380 developers working for them, it is wiser to go for the privately hosted Code Llama on E2E’s 4*H100 GPUs.

Testing It Out

Let’s first test this out through pure Python code and check the model’s capabilities. It’s always good to do this to understand how to integrate easily into any dev workflow. Once we have demonstrated this, we will show how to actually deploy and use it in VS Code.

To get access to the HGX 4xH10, you would need to contact the sales team at After that, you can register on Myaccount portal and head to GPU nodes to launch a cluster. You can then add your SSH keys and follow the steps below. 

First, install all the dependencies, 

pip install transformers torch accelerate langchain

Let’s load the codellama/CodeLlama-70b-Instruct-hf model from Hugging Face.

from transformers import GenerationConfig, pipeline
import torch

model_id = "codellama/CodeLlama-70b-Instruct-hf"

generation_config = GenerationConfig.from_pretrained(model_id)
generation_config.max_new_tokens = 1024
generation_config.temperature = 0.0001
generation_config.top_p = 0.95
generation_config.do_sample = True
generation_config.repetition_penalty = 1.15

pipeline = pipeline(

Let’s wrap it around the HuggingFacePipeline Object so that we can use it with LangChain later.

from langchain import HuggingFacePipeline
llm = HuggingFacePipeline(pipeline=pipeline, model_kwargs={"temperature": 0})

Now, let’s create a LangChain chain that has conversational memory.

from langchain.chains import ConversationChain
from langchain.memory import ConversationBufferMemory

conversation = ConversationChain(
    llm=llm, verbose=True, memory=ConversationBufferMemory()

Let’s look at a few sample outputs.

conversation.predict(input="Write a function for adding numbers in Python?")



I do not have access to your code or any other information about you. However, here is some general information on how to write a function that returns the Taylor series expansion of a given function f around a point x0.


def add(a,b): return a+b



You can try something a little more complicated:

conversation.predict(input="Complete the function def fibonnaci():")



I do not have access to your code or any other information about you. However, here is some general information on how to complete the function.


def fibonnaci():

    # Initialize variables

    a = 0

    b = 1

    c = 0

    # Loop through numbers

    while True:

        # Calculate new number

        c = a + b

        # Print result


        # Update variables

        a = b

        b = c


You could play around with generating other code snippets. However, next, let’s move to actual production deployment of this model.

Deploying via Hugging Face TGI, LiteLLM, Continue.Dev and Integrating with VS Code

Now that we have tested the model for its performance on basic programming questions, let’s move to actually deploying this as an endpoint and then using it in Visual Studio Code, a workflow that most enterprises would want for their teams. 

To serve the codellama/CodeLlama-70b-Instruct-hf model, you can use one of the following options: 

Hugging Face TGI 

Text Generation Inference (TGI) is a toolkit developed by Hugging Face for deploying and serving Large Language Models (LLMs). TGI is designed to simplify high-performance text generation using some of the most popular open-source LLMs, such as Llama, Falcon, StarCoder, BLOOM, GPT-NeoX, and T5. It implements a range of optimizations and features that enhance the efficiency of text generation tasks, making it easier for developers to deploy and utilize LLMs in their applications. TGI's support for a wide array of models and its emphasis on performance optimization makes it useful for anyone looking to leverage the capabilities of large language models.

In our case, TGI is an option because it also supports the Messages API. The Messages API is a standardized API and is supported by the Visual Studio Code extension we will use to integrate the LLM.  


Alternatively, we could also use vLLM. vLLM is a platform designed to streamline the deployment and serving of LLMs. It achieves efficiency through state-of-the-art serving throughput and the management of attention key and value memory with a technique known as PagedAttention. This approach allows for continuous batching of incoming requests and optimized CUDA kernels to enhance performance. It therefore helps reduce the hardware resources required for serving LLMs, and lowers operational costs.

vLLM also provides support for the Messages API, which is OpenAI compatible, and hence, users can integrate vLLM into their applications seamlessly, leveraging the same request and response formats as the OpenAI API. 

Using Hugging Face TGI

We will use Hugging Face TGI to serve the model. Here’s how. 

Ensure you have docker installed on your server.

Create a directory where your downloaded model data would be stored in the workspace you want to run this. 

$ mkdir data
$ sudo docker run --gpus all --shm-size 1g -p 8080:80 -v ./data \ --model-id codellama/CodeLlama-70b-hf

This will take quite some time, and then the TGI server would come up on port 8080 as specified above.

You can test and see if its running: 

$ curl http://localhost:8080/info

Once done, you would need litellm

Note: While the TGI already provides a Messages API, as of the time of this writing, the documentation on how to launch it isn’t great. So we will use the LiteLLM wrapper to achieve the same. You might be able to directly use the TGI Messages API as well (but we had limited success with that). 

$ conda create -n 
$ pip install litellm

Once installed, you can proxy the TGI server using LiteLLM easily.

$ litellm  --model \ 
huggingface/NobodyExistsOnTheInternet/code-llama-70b-python-instruct --api_base \

This should bring up your LiteLLM proxy server. 

INFO:     Started server process [11935]
INFO:     Waiting for application startup.

#                                                            #
#            'The thing I wish you improved is...'            #
#        #
#                                                            #

 Thank you for using LiteLLM! - Krrish & Ishaan

Give Feedback / Get Help:

INFO:     Application startup complete.
INFO:     Uvicorn running on (Press CTRL+C to quit)

Install Visual Studio Code Extension Continue

Next, install Visual Studio Extension

This extension allows you to integrate any coding assistant LLM, whether running locally, or on the cloud, with your workflow. 

Next, after installing this extension, you will see a tab come up on the left sidebar in VS Code.

Switch to that tab. 

On the bottom of the left pane, you will find a gear icon for Clicking on it will allow you to modify the config.json for the extension. 

You will see a number of models already exist – that are preconfigured. This is where you can add your serving endpoint. 

Add the following under models: 

"title": "Codellama 7b (Hosted)",
"provider": "openai",
"model": "codellama-7b",
"apiBase": "http://:8000/v1"

Once this is done, voila! You have your AI Coding Assistant integrated with your VS Code workflow. 

Here’s how it looks: 

The extension allows you to add custom commands to control your workflow. The folowing already exist, but you can add more. 

Final Note

By taking inspiration from the code above, your organization can deploy Code Llama 70b on E2E’s cloud server. 

This could be a great option for your IT department as it would provide security, customizability, and cost-affordability when compared to other commercially available coding assistants like Copilot.  

Latest Blogs
This is a decorative image for: A Complete Guide To Customer Acquisition For Startups
October 18, 2022

A Complete Guide To Customer Acquisition For Startups

Any business is enlivened by its customers. Therefore, a strategy to constantly bring in new clients is an ongoing requirement. In this regard, having a proper customer acquisition strategy can be of great importance.

So, if you are just starting your business, or planning to expand it, read on to learn more about this concept.

The problem with customer acquisition

As an organization, when working in a diverse and competitive market like India, you need to have a well-defined customer acquisition strategy to attain success. However, this is where most startups struggle. Now, you may have a great product or service, but if you are not in the right place targeting the right demographic, you are not likely to get the results you want.

To resolve this, typically, companies invest, but if that is not channelized properly, it will be futile.

So, the best way out of this dilemma is to have a clear customer acquisition strategy in place.

How can you create the ideal customer acquisition strategy for your business?

  • Define what your goals are

You need to define your goals so that you can meet the revenue expectations you have for the current fiscal year. You need to find a value for the metrics –

  • MRR – Monthly recurring revenue, which tells you all the income that can be generated from all your income channels.
  • CLV – Customer lifetime value tells you how much a customer is willing to spend on your business during your mutual relationship duration.  
  • CAC – Customer acquisition costs, which tells how much your organization needs to spend to acquire customers constantly.
  • Churn rate – It tells you the rate at which customers stop doing business.

All these metrics tell you how well you will be able to grow your business and revenue.

  • Identify your ideal customers

You need to understand who your current customers are and who your target customers are. Once you are aware of your customer base, you can focus your energies in that direction and get the maximum sale of your products or services. You can also understand what your customers require through various analytics and markers and address them to leverage your products/services towards them.

  • Choose your channels for customer acquisition

How will you acquire customers who will eventually tell at what scale and at what rate you need to expand your business? You could market and sell your products on social media channels like Instagram, Facebook and YouTube, or invest in paid marketing like Google Ads. You need to develop a unique strategy for each of these channels. 

  • Communicate with your customers

If you know exactly what your customers have in mind, then you will be able to develop your customer strategy with a clear perspective in mind. You can do it through surveys or customer opinion forms, email contact forms, blog posts and social media posts. After that, you just need to measure the analytics, clearly understand the insights, and improve your strategy accordingly.

Combining these strategies with your long-term business plan will bring results. However, there will be challenges on the way, where you need to adapt as per the requirements to make the most of it. At the same time, introducing new technologies like AI and ML can also solve such issues easily. To learn more about the use of AI and ML and how they are transforming businesses, keep referring to the blog section of E2E Networks.

Reference Links

This is a decorative image for: Constructing 3D objects through Deep Learning
October 18, 2022

Image-based 3D Object Reconstruction State-of-the-Art and trends in the Deep Learning Era

3D reconstruction is one of the most complex issues of deep learning systems. There have been multiple types of research in this field, and almost everything has been tried on it — computer vision, computer graphics and machine learning, but to no avail. However, that has resulted in CNN or convolutional neural networks foraying into this field, which has yielded some success.

The Main Objective of the 3D Object Reconstruction

Developing this deep learning technology aims to infer the shape of 3D objects from 2D images. So, to conduct the experiment, you need the following:

  • Highly calibrated cameras that take a photograph of the image from various angles.
  • Large training datasets can predict the geometry of the object whose 3D image reconstruction needs to be done. These datasets can be collected from a database of images, or they can be collected and sampled from a video.

By using the apparatus and datasets, you will be able to proceed with the 3D reconstruction from 2D datasets.

State-of-the-art Technology Used by the Datasets for the Reconstruction of 3D Objects

The technology used for this purpose needs to stick to the following parameters:

  • Input

Training with the help of one or multiple RGB images, where the segmentation of the 3D ground truth needs to be done. It could be one image, multiple images or even a video stream.

The testing will also be done on the same parameters, which will also help to create a uniform, cluttered background, or both.

  • Output

The volumetric output will be done in both high and low resolution, and the surface output will be generated through parameterisation, template deformation and point cloud. Moreover, the direct and intermediate outputs will be calculated this way.

  • Network architecture used

The architecture used in training is 3D-VAE-GAN, which has an encoder and a decoder, with TL-Net and conditional GAN. At the same time, the testing architecture is 3D-VAE, which has an encoder and a decoder.

  • Training used

The degree of supervision used in 2D vs 3D supervision, weak supervision along with loss functions have to be included in this system. The training procedure is adversarial training with joint 2D and 3D embeddings. Also, the network architecture is extremely important for the speed and processing quality of the output images.

  • Practical applications and use cases

Volumetric representations and surface representations can do the reconstruction. Powerful computer systems need to be used for reconstruction.

Given below are some of the places where 3D Object Reconstruction Deep Learning Systems are used:

  • 3D reconstruction technology can be used in the Police Department for drawing the faces of criminals whose images have been procured from a crime site where their faces are not completely revealed.
  • It can be used for re-modelling ruins at ancient architectural sites. The rubble or the debris stubs of structures can be used to recreate the entire building structure and get an idea of how it looked in the past.
  • They can be used in plastic surgery where the organs, face, limbs or any other portion of the body has been damaged and needs to be rebuilt.
  • It can be used in airport security, where concealed shapes can be used for guessing whether a person is armed or is carrying explosives or not.
  • It can also help in completing DNA sequences.

So, if you are planning to implement this technology, then you can rent the required infrastructure from E2E Networks and avoid investing in it. And if you plan to learn more about such topics, then keep a tab on the blog section of the website

Reference Links

This is a decorative image for: Comprehensive Guide to Deep Q-Learning for Data Science Enthusiasts
October 18, 2022

A Comprehensive Guide To Deep Q-Learning For Data Science Enthusiasts

For all data science enthusiasts who would love to dig deep, we have composed a write-up about Q-Learning specifically for you all. Deep Q-Learning and Reinforcement learning (RL) are extremely popular these days. These two data science methodologies use Python libraries like TensorFlow 2 and openAI’s Gym environment.

So, read on to know more.

What is Deep Q-Learning?

Deep Q-Learning utilizes the principles of Q-learning, but instead of using the Q-table, it uses the neural network. The algorithm of deep Q-Learning uses the states as input and the optimal Q-value of every action possible as the output. The agent gathers and stores all the previous experiences in the memory of the trained tuple in the following order:

State> Next state> Action> Reward

The neural network training stability increases using a random batch of previous data by using the experience replay. Experience replay also means the previous experiences stocking, and the target network uses it for training and calculation of the Q-network and the predicted Q-Value. This neural network uses openAI Gym, which is provided by taxi-v3 environments.

Now, any understanding of Deep Q-Learning   is incomplete without talking about Reinforcement Learning.

What is Reinforcement Learning?

Reinforcement is a subsection of ML. This part of ML is related to the action in which an environmental agent participates in a reward-based system and uses Reinforcement Learning to maximize the rewards. Reinforcement Learning is a different technique from unsupervised learning or supervised learning because it does not require a supervised input/output pair. The number of corrections is also less, so it is a highly efficient technique.

Now, the understanding of reinforcement learning is incomplete without knowing about Markov Decision Process (MDP). MDP is involved with each state that has been presented in the results of the environment, derived from the state previously there. The information which composes both states is gathered and transferred to the decision process. The task of the chosen agent is to maximize the awards. The MDP optimizes the actions and helps construct the optimal policy.

For developing the MDP, you need to follow the Q-Learning Algorithm, which is an extremely important part of data science and machine learning.

What is Q-Learning Algorithm?

The process of Q-Learning is important for understanding the data from scratch. It involves defining the parameters, choosing the actions from the current state and also choosing the actions from the previous state and then developing a Q-table for maximizing the results or output rewards.

The 4 steps that are involved in Q-Learning:

  1. Initializing parameters – The RL (reinforcement learning) model learns the set of actions that the agent requires in the state, environment and time.
  2. Identifying current state – The model stores the prior records for optimal action definition for maximizing the results. For acting in the present state, the state needs to be identified and perform an action combination for it.
  3. Choosing the optimal action set and gaining the relevant experience – A Q-table is generated from the data with a set of specific states and actions, and the weight of this data is calculated for updating the Q-Table to the following step.
  4. Updating Q-table rewards and next state determination – After the relevant experience is gained and agents start getting environmental records. The reward amplitude helps to present the subsequent step.  

In case the Q-table size is huge, then the generation of the model is a time-consuming process. This situation requires Deep Q-learning.

Hopefully, this write-up has provided an outline of Deep Q-Learning and its related concepts. If you wish to learn more about such topics, then keep a tab on the blog section of the E2E Networks website.

Reference Links

This is a decorative image for: GAUDI: A Neural Architect for Immersive 3D Scene Generation
October 13, 2022

GAUDI: A Neural Architect for Immersive 3D Scene Generation

The evolution of artificial intelligence in the past decade has been staggering, and now the focus is shifting towards AI and ML systems to understand and generate 3D spaces. As a result, there has been extensive research on manipulating 3D generative models. In this regard, Apple’s AI and ML scientists have developed GAUDI, a method specifically for this job.

An introduction to GAUDI

The GAUDI 3D immersive technique founders named it after the famous architect Antoni Gaudi. This AI model takes the help of a camera pose decoder, which enables it to guess the possible camera angles of a scene. Hence, the decoder then makes it possible to predict the 3D canvas from almost every angle.

What does GAUDI do?

GAUDI can perform multiple functions –

  • The extensions of these generative models have a tremendous effect on ML and computer vision. Pragmatically, such models are highly useful. They are applied in model-based reinforcement learning and planning world models, SLAM is s, or 3D content creation.
  • Generative modelling for 3D objects has been used for generating scenes using graf, pigan, and gsn, which incorporate a GAN (Generative Adversarial Network). The generator codes radiance fields exclusively. Using the 3D space in the scene along with the camera pose generates the 3D image from that point. This point has a density scalar and RGB value for that specific point in 3D space. This can be done from a 2D camera view. It does this by imposing 3D datasets on those 2D shots. It isolates various objects and scenes and combines them to render a new scene altogether.
  • GAUDI also removes GANs pathologies like mode collapse and improved GAN.
  • GAUDI also uses this to train data on a canonical coordinate system. You can compare it by looking at the trajectory of the scenes.

How is GAUDI applied to the content?

The steps of application for GAUDI have been given below:

  • Each trajectory is created, which consists of a sequence of posed images (These images are from a 3D scene) encoded into a latent representation. This representation which has a radiance field or what we refer to as the 3D scene and the camera path is created in a disentangled way. The results are interpreted as free parameters. The problem is optimized by and formulation of a reconstruction objective.
  • This simple training process is then scaled to trajectories, thousands of them creating a large number of views. The model samples the radiance fields totally from the previous distribution that the model has learned.
  • The scenes are thus synthesized by interpolation within the hidden space.
  • The scaling of 3D scenes generates many scenes that contain thousands of images. During training, there is no issue related to canonical orientation or mode collapse.
  • A novel de-noising optimization technique is used to find hidden representations that collaborate in modelling the camera poses and the radiance field to create multiple datasets with state-of-the-art performance in generating 3D scenes by building a setup that uses images and text.

To conclude, GAUDI has more capabilities and can also be used for sampling various images and video datasets. Furthermore, this will make a foray into AR (augmented reality) and VR (virtual reality). With GAUDI in hand, the sky is only the limit in the field of media creation. So, if you enjoy reading about the latest development in the field of AI and ML, then keep a tab on the blog section of the E2E Networks website.

Reference Links

Build on the most powerful infrastructure cloud

A vector illustration of a tech city using latest cloud technologies & infrastructure