Hello everyone, in the open source community, AI-related projects are the fastest updated. If you don’t look at them for a few days, there will be some excellent projects.
1. FinGPT
I have posted large language models in various fields before, such as those in the medical field:
Also published a large language model in the legal field:
Now, in the financial field, FinGPT is here!
However, unlike the previous ones, FinGPT is not a well-trained large language model, but a complete framework that provides a complete pipeline for LLM training and fine-tuning in the financial field.
AIGC Technical Exchange
Technology must learn to share and communicate, and it is not recommended to work behind closed doors. One person can go fast, and a group of people can go farther.
Good articles are inseparable from fans’ sharing, communication, and recommendations. Dry data, data sharing, data, and technical exchange improvements can all be obtained by adding an exchange group. The group has more than 2,000 members. The best way to add notes is: source + Interest direction, easy to find like-minded friends.
Method ①, add WeChat account: mlc2060, remarks: from CSDN + large model
Method ②, WeChat search official account: machine learning community, background reply: large model
The whole consists of four parts:
-
Data Source: It consists of news, social media, company announcements, search trends, open source datasets, etc.
-
Data Engineering: This part cleans the data, calculates token features, and prompt tags, etc. It belongs to the preprocessing stage of data.
-
LLMs (Large Language Model): It supports various APIs and various open source LLMs, which can be used for training and fine-tuning.
-
Applications: The last part is the upper layer application, which provides many application demos.
In short, using the FinGPT framework, you can use existing APIs to get some interesting financial applications, and you can also use your own data, use open source models, and train your own private models.
The data sources are very rich. For example, news sources include:
Social Media Sources:
Combined with company announcements, search trends, public data sets, and multi-party data, an LLM for the financial field can be trained or fine-tuned.
Let FinGPT provide you with some consulting, opinions and other services in the financial field.
project address:
https://github.com/ai4finance-foundation/fingpt
Personally, I am quite curious about such large models.
You said that if everyone uses the same trained model to do financial investment-related consulting, the answers they get should be similar.
People believe in it again, and they all make similar operations. Could it be that AI directly controls the direction of the market.
2. Chat2DB
The "gospel" of crud boy, friends who often write sql, you can experience Chat2DB.
By handing over the task of writing sql and optimizing commands to AI, we will have more time to fish 
.
We describe in natural language, and Chat2DB helps us write SQL language well.
At present, the bottom layer uses the chatgpt api, I tested it, and the writing effect is barely good.
However, considering that there is no training optimization for related large models for sql data, this effect is also acceptable.
I believe that if someone really wants to do it, a large language model for sql scenarios can be trained, which will definitely be a labor-saving artifact.
Interested friends, you can experience it.
project address:
https://github.com/alibaba/Chat2DB