Guide to Fine-Tuning Llama 3 on TIR for Financial Queries

June 6, 2024

A Note on Open-Source Models

Open-source LLMs have gained massive recognition for their accessibility and flexibility. Their ability to provide complete control over the source code, data, and deployment as well as offer an active community for support and development have given them an advantage over closed-source LLMs. This makes them ideal for a wide range of users, from researchers and educators to developers and businesses as they offer a range of customizations targeted towards each domain.

Meta has been a significant contributor to the Large Language Model (LLM) space, starting with the launch of Llama. This model marked Meta's entry into the competition with key players by demonstrating the potential of open-source pre-trained models. The subsequent release of Llama-2 further solidified Meta's position, triggering an explosion of open-source LLMs that were more efficient in written tasks, code, and problem-solving. The latest in this series is Llama-3, which continues Meta's commitment to democratizing powerful AI models. 

LLaMA 3: An Overview

Llama 3 is the latest contribution of Meta to the open-source LLM space. It shows considerable improvements in benchmarks vis a vis Llama 2 and other open-source LLMs like Gemma. Trained on a 7 times larger dataset than Llama 2, with 4 times more code, it shows considerably better performance in tasks like question-answering, coding, creative writing, extraction, inhabiting a character/persona, open question-answering, reasoning, rewriting, and summarization.

Image from Meta AI

Llama 3 is available in 2 versions: Llama3-8B and Llama3-70B. The model was trained on a whopping 15 trillion tokens and the context length was doubled from 4K to 8K for both the 8B and 70B models. Meta has also used grouped query attention for both models. Llama 3 maintains the decoder-only transformer design that was successful in its earlier versions. It includes a tokenizer that can manage 128,256 tokens.

In this tutorial, we will explore fine-tuning, deployment, and inference with Llama 3 on the TIR AI Platform by E2E Networks.

Fine-Tuning, Deployment, and Inference with Llama 3 on the TIR AI Platform


The tutorial will be covered using the model and functions from the Hugging Face ecosystem. The model will be fetched from Hugging Face models. Launch a Jupyter Notebook on the E2E TIR AI Platform or connect a node via SSH, and login using your Hugging Face credentials.

!huggingface-cli login

Install the required libraries for fine-tuning.

%pip install -U bitsandbytes
%pip install -U transformers
%pip install -U peft
%pip install -U accelerate
%pip install -U trl
%pip install -U wandb
%pip install torch

import os,torch
if torch.cuda.get_device_capability()[0] >= 8:
    !pip install -qqq flash-attn
    attn_implementation = "flash_attention_2"
    torch_dtype = torch.bfloat16
    attn_implementation = "eager"
    torch_dtype = torch.float16

Set up wandb and login with API key.

import wandb
wandb.login(key = API_KEY)

Llama 3 is a gated model in Hugging Face. Make sure you have requested and received access to the model from the repo page before proceeding to use it.


Let’s start fine-tuning the model for finance queries. Import the required libraries.

from transformers import AutoModelForCausalLM, AutoTokenizer, BitsAndBytesConfig,HfArgumentParser,TrainingArguments,pipeline, logging
from peft import LoraConfig, PeftModel, prepare_model_for_kbit_training, get_peft_model
from datasets import load_dataset
from trl import SFTTrainer


For this tutorial, I will be using the finance alpaca dataset.  Load the dataset from HuggingFace.

dataset_name = "gbharti/finance-alpaca"
dataset = load_dataset(dataset_name, split = "train")

Import the tokenizer from the model repo

base_model = "meta-llama/Meta-Llama-3-8B"
tokenizer = AutoTokenizer.from_pretrained(base_model, trust_remote_code=True)
tokenizer.padding_side = 'right'
tokenizer.pad_token = tokenizer.eos_token
tokenizer.add_eos_token = True

Adding a prompt template helps with consistency, clarity, and more control over the model. Here is one that we can use. Load the dataset and make necessary changes in the training part.

prompt = """Below is an instruction that describes a finance-related task, paired with an input that provides further context. Write a response that appropriately completes the request.

### Instruction:

### Input:

### Response:

EOS_TOKEN = tokenizer.eos_token # Must add EOS_TOKEN
def formatting_prompts_func(examples):
    instructions = examples["instruction"]
    inputs       = examples["input"]
    outputs      = examples["output"]
    texts = []
    for instruction, input, output in zip(instructions, inputs, outputs):
        # Must add EOS_TOKEN, otherwise your generation will go on forever!
        text = prompt.format(instruction, input, output) + EOS_TOKEN
    return { "text" : texts, }

dataset =, batched = True,)

EOS_TOKEN is the end-of-string token used by the tokenizer. This token is appended to each formatted string to indicate the end of the text.

Preparing Llama 3

I will be using PEFT (PyTorch Efficient Fine-Tuning) and LoRA (Low-Rank Adapters) to fine-tune Llama 3 for efficient and effective model training. PEFT provides efficient fine-tuning capabilities for large language models like Llama 3. LoRA reduces computational requirements and memory footprint, making the fine-tuning process more manageable.

Set up a BitsAndBytesConfig for 4-bit quantization and load the pre-trained model with this configuration.

bnb_config = BitsAndBytesConfig(
model = AutoModelForCausalLM.from_pretrained(
model.config.use_cache = True 
model.config.pretraining_tp = 1
model = prepare_model_for_kbit_training(model)
peft_config = LoraConfig(
    target_modules=['up_proj', 'down_proj', 'gate_proj', 'k_proj', 'q_proj', 'v_proj', 'o_proj']
model = get_peft_model(model, peft_config)
new_model = "Llama-3-Finance"

Finally, define LoraConfig for Low-Rank Adapters (LoRA) and apply this configuration to the model using the get_peft_model.

Model Training

Train the model and save the weights. You can access the weights later for deployment.

training_arguments = TrainingArguments(
trainer = SFTTrainer(
    max_seq_length = 2048,
    packing= False,
    dataset_num_proc = 2,

model = trainer.model.merge_and_un

Creating Model Repository

To create the inference endpoint, we will need to move the model to a model repository on TIR.

  1. Go to TIR Dashboard.
  2. Choose a project.
  3. Go to Model Repository section.
  1. Click on Create Model.
  1. Enter a model name. 
  2. Select Model Type as Custom for the custom container as we are deploying a fine-tuned model. Click on CREATE.
  1. You will now see details of the EOS (E2E Object Storage) bucket created for this model.
  2. EOS offers an API that is compatible with S3 for uploading and downloading data. In this tutorial, we'll be utilizing the MinIO CLI.
  3. Set up MinIO CLI using the commands on the following page.
  1. Use the commands to go to the folder containing the model snapshots. Then upload the model files that were saved after the training phase to the model repository.

If you cannot find the model files in the cache, try saving the fine-tuned LoRA adapter, reloading the base model, and merging the weights again.

Note: Use the correct model repository and folder names. The above shown is just a demo.

You can see and confirm the file contents in the model repository after uploading.

Deployment & Inference

Kudos! You are almost there. Model weights are uploaded to the TIR Model Repository. Now it is ready to use in an endpoint and serve API requests.

  • From the API tokens section, create an API token if you don't already have one. This will be used in your requests.
  • Go to Model Endpoints section.
  • Create a new Endpoint.
  • Choose LLAMA-3-8B-IT model card.
  • Choose Link with Model Repository and select the Model Repo from the dropdown.
  • Set the required settings.
  • Pick a suitable GPU plan of your choice. It is recommended to select GPUs V100 and above for Llama 3 inference.
  • Add a custom endpoint name and set environment variables.

The endpoint will take a few minutes to be ready.

For testing the inference, find the Python and cURL requests in the API requests section of your endpoint. Replace the API_KEY.


curl --location '' --header 'Content-Type: application/json' --header 'Authorization: Bearer $API_KEY -X POST --data '{
  "text_input": "What's the difference between Term and Whole Life insurance?",
  "max_tokens": 1000


import requests
import json
url = ""

payload = json.dumps({
  "text_input": "What's the difference between Term and Whole Life insurance?",
  "max_tokens": 1000

headers = {
    'Content-Type': 'application/json',
    'Authorization': 'Bearer $API_KEY'

response = requests.request("POST", url, headers=headers, data=payload)

Here is a sample response:

["<|begin_of_text|>Below is an instruction that describes a finance related task, paired with an input that provides further context. Write a response that appropriately completes the request.\n\n### Instruction:\nWhat's the difference between Term and Whole Life insurance?\n\n### Input:\n\n\n### Response:\nTerm life insurance is a type of life insurance that provides coverage for the life of the insured. Whole life insurance is a type of life insurance that provides coverage for the life of the insured. Term life insurance is a type of life insurance that provides coverage for the life of the insured. Whole life insurance is a type of"]

Wrapping Up

You have learned how to fine-tune and deploy Llama 3 seamlessly on the TIR AI platform of E2E Networks. I recommend exploring other model containers, such as vision models, to further understand the simplicity of this platform for model deployment and inference. By exploring the pipelines and foundation studio sections, you’ll gain more insights into no-code training and fine-tuning generative AI models. Hope this tutorial helped you to understand better the steps required in AI model deployment and inference. 


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