Saved in:
Bibliographic Details
Main Authors: Jindal, Ashvini Kumar, Rajpoot, Pawan Kumar, Parikh, Ankur
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2403.02247
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866911788770000896
author Jindal, Ashvini Kumar
Rajpoot, Pawan Kumar
Parikh, Ankur
author_facet Jindal, Ashvini Kumar
Rajpoot, Pawan Kumar
Parikh, Ankur
contents LLMOps incur significant costs due to hardware requirements, hindering their widespread accessibility. Additionally, a lack of transparency in model training methods and data contributes to the majority of models being non-reproducible. To tackle these challenges, the LLM Efficiency Challenge was introduced at NeurIPS Workshop, aiming to adapt foundation models on a diverse set of tasks via fine-tuning on a single GPU (RTX 4090 or A100 with 40GB) within a 24-hour timeframe. In this system description paper, we introduce Birbal, our Mistral-7B based winning model, fine-tuned on a single RTX 4090 for 16 hours. Birbal's success lies in curating high-quality instructions covering diverse tasks, resulting in a 35% performance improvement over second-best Qwen-14B based submission.
format Preprint
id arxiv_https___arxiv_org_abs_2403_02247
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Birbal: An efficient 7B instruct-model fine-tuned with curated datasets
Jindal, Ashvini Kumar
Rajpoot, Pawan Kumar
Parikh, Ankur
Computation and Language
LLMOps incur significant costs due to hardware requirements, hindering their widespread accessibility. Additionally, a lack of transparency in model training methods and data contributes to the majority of models being non-reproducible. To tackle these challenges, the LLM Efficiency Challenge was introduced at NeurIPS Workshop, aiming to adapt foundation models on a diverse set of tasks via fine-tuning on a single GPU (RTX 4090 or A100 with 40GB) within a 24-hour timeframe. In this system description paper, we introduce Birbal, our Mistral-7B based winning model, fine-tuned on a single RTX 4090 for 16 hours. Birbal's success lies in curating high-quality instructions covering diverse tasks, resulting in a 35% performance improvement over second-best Qwen-14B based submission.
title Birbal: An efficient 7B instruct-model fine-tuned with curated datasets
topic Computation and Language
url https://arxiv.org/abs/2403.02247