# FinText

> A financial NLP repository


## Posts

- [Citation](https://fintext.ai/?p=747)
- [Download](https://fintext.ai/?p=746)
- [Hello world!](https://fintext.ai/?p=1): Welcome to WordPress. This is your first post. Edit or delete it, then start writing!

## Pages

- [Citation](https://fintext.ai/?page_id=770): Rahimikia, Eghbal, and Felix Drinkall. “Re (visiting) large language models in finance.” Available at SSRN 4963618 (2024).
- [Download](https://fintext.ai/?page_id=755): The look-ahead bias benchmark contains a structured set of evaluation probes organised into three columns: the first column provides the prompt identifier, the second column contains the prompt with a masked element, and the third column specifies the corresponding correct (golden) year. The benchmark is accessible through this link: Download Look-ahead Benchmark
- [Look-head bias Benchmark](https://fintext.ai/?page_id=731): The look-ahead bias benchmark introduced in Re(Visiting) Large Language Models in Finance is a set of structured event–year probes that test whether models correctly link financial events to the time period in which they actually occurred using only information available at that point in time. By evaluating how models perform on these probes, the benchmark detects whether predictions are influenced by future information, providing a clear and controlled way to identify information leakage and assess temporal alignment in financial machine learning models.
- [Download (HParam)](https://fintext.ai/?page_id=566): The models are available on Hugging Face. You can access them using the page linked below. Additional TSFMs not available for direct download on Hugging Face can be downloaded here. This page provides Chronos models using dynamic quantization range settings, as well as models using a tighter fixed range of [-2, 2]. The first group of models is labeled ‘Dynamic,’ and the second group is labeled ‘TR.’ Hugging Face Page Chronos (Tiny) – Dynamic Chronos (Mini) – Dynamic Chronos (Small) – Dynamic Chronos (Tiny) – TR Chronos (Mini) – TR Chronos (Small) – TR 2000 2001 2002 2003 2004 2005 […]
- [Citation](https://fintext.ai/?page_id=338): Rahimikia, Eghbal and Ni, Hao and Wang, Weiguan, Re(Visiting) Time Series Foundation Models in Finance (November 18, 2025). Available at SSRN: https://ssrn.com/abstract=5770562 or http://dx.doi.org/10.2139/ssrn.5770562
- [Download (Synthetic)](https://fintext.ai/?page_id=337): Hugging Face Page The models are available on Hugging Face. You can access them using the page linked below. Additional TSFMs not available for direct download on Hugging Face can be downloaded here. This page covers the TSFMs labeled ‘Augmented’ in the paper, which combine global data with synthetic data for pre-training. Chronos (Tiny) – Synthetic Chronos (Mini) – Synthetic Chronos (Small) – Synthetic TimesFM (8M) – Synthetic TimesFM (20M) – Synthetic 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2000 2001 2002 2003 2004 […]
- [Review](https://fintext.ai/?page_id=335): Built on the study “Re(Visiting) Time Series Foundation Models in Finance” (Rahimikia, Ni, and Wang, 2025), FinText advances the frontiers of AI-driven financial forecasting through large-scale, domain-specific Time Series Foundation Models (TSFMs). These large scale mdoels models are pre-trained from scratch on massive datasets—spanning over two billion daily excess return observations across 89 markets—to capture the complex temporal dynamics of global finance. This study shows that generic foundation models fail to generalize effectively to financial data, whereas finance-native pre-training leads to significant improvements in predictive accuracy and portfolio performance. Each model is chronologically pre-trained to prevent look-ahead bias, ensuring realistic, […]
- [Time Series Foundation Models for Finance](https://fintext.ai/?page_id=331): FinText is an open research initiative dedicated to advancing Artificial Intelligence and Machine Learning in finance. Our flagship release, FinText-TSFM, introduces a comprehensive suite of over 600 time series foundation models (TSFMs) pre-trained on large-scale financial data across 94 global markets.  
- [Gold-Standard Financial Benchmarks](https://fintext.ai/?page_id=294): We introduce the first gold-standard financial benchmark for systematically comparing word embeddings using a financial language framework. This benchmark covers seven groups of financial analogies. Each group contains 80 analogies reaching 2660 unique analogies in total for all groups. All financial analogies are developed using the Bureau van Dijk’s Orbis database and are available for download. In the table below, the first five groups cover publicly listed US companies, the sixth group mixes US and UK publicly listed companies, and the last group mixes US, UK, China, and Japan publicly listed companies. ‘Ticker’ is the security ticker identifier, ‘Name’ is […]
- [Citation](https://fintext.ai/?page_id=193): Rahimikia, Eghbal and Zohren, Stefan and Poon, Ser-Huang, Realised Volatility Forecasting: Machine Learning via Financial Word Embedding (July 28, 2021). Available at SSRN 3895272.
- [Contact Us](https://fintext.ai/?page_id=174)
- [Examples](https://fintext.ai/?page_id=117): The figure below presents the 2D visualisation of the principal component analysis (PCA) of the word embedding 300-dimensional vectors. Dimension 1 (x-axis) and Dimension 2 (y-axis) show the first and second obtained dimensions. The tokens are chosen from groups of technology companies (‘microsoft’, ‘ibm’, ‘google’, and ‘adobe’), financial services and investment banks (‘barclays’, ‘citi’, ‘ubs’, and ‘hsbc’), and retail businesses (‘tesco’ and ‘walmart’). Word2Vec is shown in the top row, and FastText is shown in the bottom row. Google is a publicly available word embedding trained on a part of the Google News dataset, and WikiNews is another publicly available […]
- [Download](https://fintext.ai/?page_id=44): Gold-Standard Financial Benchmarks(18.00KB) This package contains the gold-standard financial benchmarks and the code for reading and applying them to the desired word embedding.   Word2Vec/CBOW(5.86GB) This word embedding is developed based on Word2Vec algorithm and CBOW model. FastText/CBOW(10.80GB) This word embedding is developed based on FastText algorithm and CBOW model. Word2Vec/Skip-gram(5.85GB) This word embedding is developed based on Word2Vec algorithm and Skip-gram model. FastText/Skip-gram(10.80GB) This word embedding is developed based on FastText algorithm and Skip-gram model. Python code(3.54KB) This python code can be used for reading and representing the downloaded word embeddings.
- [Word Embeddings](https://fintext.ai/?page_id=35): This page contains Word2Vec and FastText models, the purpose-built financial word embeddings for financial textual analysis. Dow Jones Newswires Text News Feed from January 1, 2000, to September 14, 2015, is used for developing these financial word embeddings. This contains millions of news stories (2,733,035 unique tokens) covering finance, economics, politics, etc., from various news agencies worldwide. Also, extensive text pre-processing is applied to ensure this big textual data is empty of redundant characters, sentences, and structures. Four variations are available to download, containing Word2Vec and FastText algorithms via CBOW and Skip-gram models.
- [FinText | A Financial AI Repository](https://fintext.ai/): FinText is a repository of specialised AI models for accounting, finance, and related fields. It offers a variety of models designed to support research in these domains. News Update (April 2026): Our project has been featured by S&P Global. The video is available to watch below. You can also watch it HERE. News Update (April 2026): Our project has been featured by Hewlett Packard Enterprise (HPE). Read the full article HERE. News Update (March 2026): Our project has been featured in additional media coverage. Read the full article HERE. Update (Feburary 2026): A new video discussing the project is available […]


[comment]: # (Generated by Hostinger Tools Plugin)