ERPNext + LLM (ChatGPT, Ollama) | Part -2

[Antony_Praveenkumar]

Here the Part -1

ERPNext + LLM (ChatGPT, Ollama)

Project Description: We’re developing an exciting initiative to enhance ERPNext with AI capabilities, revolutionizing business decision-making. Our goal is to integrate tools like ChatGPT and Ollama for real-time data analysis, enabling features such as AI-driven business suggestions, SQL generation, and automated report creation. Join our open-source collaboration to shape the future of intelligent ERP solutions and empower businesses with data-driven insights. Anyone Intrested Please…

I have done the crawling on these three sites.

1. docs.erpnext.com
2. frappeframework.com
3. frappehr.com

And converted as datasets and stored in the HugginFace. Here is the dataset link.

1. antony-pk/erpnext-docs-ds
2. antony-pk/erpnext-frappeframework-ds
3. antony-pk/erpnext-frappe-hr-ds

From this Custom Dataset I have done the fine tune using the microsoft/Phi-3-mini-4k-instruct pre-trained model. Here the custom model antony-pk/Phi-3-mini-4k-instruct-erpnext.

The model I fine-tuned is really giving very unusual response. Im currently working on that response.

If anyone really interested you can join with me in this research

Please join in the whatsapp group

ERPNext AI research

Contact

Github : Antony-M1 · GitHub
Linkedin : Antony

[max_morais_dmm]

@Antony_Praveenkumar congratulations for starting this!

I see you are using Phi 3, nice choice!