I Made AI to Auto Categorise 10000 Comments in Google Sheets with 0$

Просмотров 4,7 тыс.

% 220

🌈 Welcome to our latest tutorial! 🌈
In today's video, I'm thrilled to guide you through the process of automatically categorising comments in Google Sheets using an open-source language model, all at ZERO COST! Whether you're overwhelmed by the volume of feedback on your RU-vid channel or looking to streamline customer insights for your business, this tool is a game-changer. 🚀✨
*If you like this video:*
Tweet something positive or what you like about these tutorials in @MervinPraison
"@MervinPraison ......................."
What You'll Learn:
Setting up your project in Google Cloud Console and enabling APIs. 🛠️
Creating and managing service account credentials for Google Sheets access. 🔐
Integrating an open-source language model to read and categorise each row based on content. 🤖
Running the entire process locally on your computer, ensuring privacy and saving costs. 💻💰
Timestamps:
0:00 Introduction to Automated Comment Categorisation
0:37 Setting Up Google Cloud Console & APIs
1:16 Creating Service Account & Managing Credentials
2:50 Accessing Google Sheets & Setting Permissions
3:09 Installing Necessary Python Packages
3:28 Coding the Categorisation Script
5:48 Running the Classification Process & Final Thoughts
👉 Subscribe and hit the bell icon to not miss out on future insights into Artificial Intelligence and tech tutorials. Your likes and shares help us reach more people and encourage us to produce more content like this. 💖
🔗 Resources:
Ollama Tutorials: ru-vid.com/group/PLYQsp-tXX9w5A0gsQ0iaJead_HvMzyl2M
Patreon: patreon.com/MervinPraison
Ko-fi: ko-fi.com/mervinpraison
Discord: discord.gg/nNZu5gGT59
Twitter / X : mervinpraison
Code: mer.vin/2024/02/ollama-auto-categorise/
Google Cloud Console: console.cloud.google.com/
OpenAI Documentation: openai.com/api/
#AI #GoogleSheets #Ollama #ChatGPTGoogleSheets #GoogleSheetsAI #GoogleSheetsTutorial #GPTForSheets #HowToInstallGPTForGoogleSheets #HowToUseChatGPTInGoogleSheets #GoogleSheetsAndChatGPT #HowToInstallChatGPTForGoogleSheets #AIGoogleSheets #OpenAI #GPT #ChatGPT #GoogleSheetsAIData #GoogleSheetsAIFree #GoogleSheetsAIPlugin #GoogleSheetsAIWorkspace #GPTInGoogleSheets #ChatGPTInGoogleSheets #AIForGoogleSheets #OllamaGoogleSheets #OllamaSheets #OllamaCSV

Хобби

Опубликовано:

20 фев 2024

Ссылка:

Скачать:

Готовим ссылку...

Добавить в:

Мой плейлист

Посмотреть позже

Комментарии : 43

@moviesnowhere9660 4 месяца назад

I appreciate your time and your interest in this . Thank you for sharing such interesting and amazing videos ❤

@MervinPraison 4 месяца назад

My Pleasure :) Thank you for your comment.

@IdPreferNot1 4 месяца назад

Great example and setup. Once we get a get a stronger local model (come on Llama 3!), i think the winning low cost personal AI solution will be using ollama, google cloud and its APIs, and the OpenAI API style and python library.

@mountee 4 месяца назад

would be interesting to see the autogen version of the same program - great job as always

@MervinPraison 4 месяца назад

Good suggestion! Thank you, will try to incorporate similar approach.

@MrCharkteeth 4 месяца назад

Thank you for making videos. I'm not a computer scientist or a programmer, but I've found LLMs interesting, and your videos are helping me learn where to start.

@nachoeigu 4 месяца назад

Why did you import openai if llama2 is outside Open Ai ecosystem?

@svenandreas5947 4 месяца назад

Really great stuff, is there any way to have langgraph not just running through a predefined way, but checking if he knows the username and ask if not having, continuing after this?

@ScottzPlaylists 4 месяца назад

👋 This could be turned into an even greater tool: an AI File Organizer. 👍❗ It could move downloaded files (or any folder specified) Into a folder hierarchy already created where you place organized files. 🎯 To modify your tool a little, Filenames could replace comments, and Folder Names could replace Catagories. (but only use the existing folder names with full path) 🎯 Once the new spreadsheet exists where the files are organized into folders, and optionally everything is approved, then run another little code section to move the Files into the correct folders. ❗

@MervinPraison 4 месяца назад

Great Idea! Thanks for sharing :)

@historybufs 4 месяца назад

Awesome dude. You are doing great

@MervinPraison 4 месяца назад

Thank you

@basilbrush7878 4 месяца назад

very useful, thanks

@MeinDeutschkurs 4 месяца назад

This is cool. Currently I work on something similar. JavaScript should be rated: save, potentially unsafe, definitely unsafe (XSS)

@MervinPraison 4 месяца назад

Nice. Please share you final product if you think it would be useful for the community.

@MeinDeutschkurs 4 месяца назад

@@MervinPraison as soon as I discover the final prompt, I‘ll publish it on GitHub. I plan to use two prompts. An outer prompt, telling the AI the base instructions about its verbosity (to avoid to be tricked by the content: avoid jailbreak), and an inner prompt to ask for the output based on the user‘s input. I hope this model is able to work with capsuled information. GPT-3.5-turbo can do it partially, GPT-4 is doing well, but certainly on a cost. Can‘t wait to get my fingers on mistral-next (or its official - hopefully downloadable - release).

@MeinDeutschkurs 4 месяца назад

I will. Btw, my long comment was deleted, because it was mixed up with an URL-comment. Model name with a dot. I don‘t want to write everything again.

@adventurelens001 4 месяца назад

Love this, thanks Mervin

@MervinPraison 4 месяца назад

Thank you

@abdelkaioumbouaicha 4 месяца назад

📝 Summary of Key Points: 📌 The video demonstrates how to automatically categorize text in Google Sheets using AI to efficiently analyze and understand comments or feedback. 🧐 The process involves setting up Google Cloud Console, enabling APIs, creating service accounts, installing necessary Python packages, defining functions for classification, and running the categorization process. 🚀 By utilizing a language model like OpenAI's GPT-3 locally on your computer, you can categorize comments based on examples provided, improving efficiency and understanding of feedback. 💡 Additional Insights and Observations: 💬 "This is amazing now you are able to run this completely locally on your computer with zero cost." 📊 The video emphasizes the importance of categorizing comments to understand sentiments, questions, requests, errors, and more effectively. 📣 Concluding Remarks: The video showcases a practical and cost-effective way to automatically categorize text in Google Sheets using AI, enabling efficient analysis of comments or feedback. By following the outlined steps, users can enhance their understanding of audience sentiments and improve content based on valuable insights. Generated using TalkBud

@MervinPraison 4 месяца назад

Thank you

@xorq Месяц назад

so instead of the openai trick you can import ollama and then (it didnt work for me to put a fake key as it stopped bercause invalid) def process_comment(comment): prompt = create_prompt(comment) response = ollama.chat(model='llama3:instruct', messages=[{'role': 'user', 'content': prompt}]) result = response['message']['content'].strip() return result

@rahuldinesh2840 Месяц назад

This code is not running. Can you check again?

@alfonsor.6722 2 месяца назад

Amazing video, and very useful, thanks for it!! Do you know about similar implementations (Google Sheets using AI) using local Open Source LLMs and tools (like Mistral and Llama2 with Ollama)? really appreciate any information.

@MervinPraison 2 месяца назад

In this video I showed how to use ollama and llama2

@hectorprx 4 месяца назад

Thanks, How would you automate populating sheets with comments? All the best

@MervinPraison 4 месяца назад

I used RU-vid API , but haven't automated that yet.

@GreenmeResearch 4 месяца назад

Would be interesting how to do this for all video's in my whole channel. First collect all containing comments per video in the channel. Then process all

@MervinPraison 4 месяца назад

First fetch comments using RU-vid API, then do same as this video suggests.

@ollimacp 4 месяца назад

Thanks, next time without notification sounds, so I not might search for minutes, which device send me what not realizing its at your side? :D

@MervinPraison 4 месяца назад

haha. Just to clarify, notification sound at which point of time in the video?

@ollimacp 4 месяца назад

4:39 @@MervinPraison

@yoninazarathy7570 4 месяца назад

Nice video!

@MervinPraison 4 месяца назад

Thank you

@biskero 4 месяца назад

very cool !

@MervinPraison 4 месяца назад

Thank you

@vaisakhkm783 4 месяца назад

wait, was it that easy.. now i have to try it.. thanks...

@MervinPraison 4 месяца назад

Sure, Give it a try

@globallama8094 4 месяца назад

Well its not really 0$, you need to invest into hardware to spin up LLMs locally.

@MervinPraison 4 месяца назад

Based on the understanding that most people will have their own personal computer to run these Models.

@globallama8094 4 месяца назад

@@MervinPraison Since you're required to put at least 2k$ to have a reasonable token/s speed ratio, but appreciate you showing capabilities, it's definitely game changer.

@IdPreferNot1 4 месяца назад

@lama8094 Nope :) $1300 if that. -----> Used i7 ($500). GTX 1660 ti ($300-7.5 cuda score). (32 gb DDR $80), Monitor ($300). Cheap SSDs $100. $20 Windows 11 license so you can do both windows and linux. This is PLENTY for fast local generation. You can always build faster, including an awesome Grock PC for $50K, but this is all anyone needs to deep dive into AI, including locally. The GTX 1660 is a great start for GPU acceleration. Get a i7 chip with AVX2... less than 10 years old.