Видео

I think the demo needs to be updated to handle "granted access" to the Google gemma models. You also need an HF token in your colab secrets to access the models: change the checkpoint you've been granted access to in the calls to create the tokenizer and the model and add the token=your_hf_token to each of the calls.

Outstanding video. Subscribed.

As the context windows of llm increase will it affect rag badly cause after some time the tokens will cost less than now

Thank you!

Notebook: github.com/mongodb-developer/GenAI-Showcase/blob/main/notebooks/agents/how_to_build_ai_agent_claude_3_5_sonnet_llamaindex_mongodb.ipynb Best Repo for you: github.com/mongodb-developer/GenAI-Showcase Article Version: www.mongodb.com/developer/products/atlas/claude_3_5_sonnet_rag/

Try out OpenSpeech and let me know what you think: bit.ly/3VYVfc8

Richmond, sir, I need your help. I want to enter AI Agent Development, so how can I start? I just completed high school, and in mid-August, my college will start. And zero coding exp. so please help

Richmond, sir, I need your help. I want to enter AI Agent Development, so how can I start? I just completed high school, and in mid-August, my college will start. And zero coding exp. so please help

great

Glad I was recommended this! Brilliant video

Thank you for sharing, very interesting, have a great day. :-)

Thank you, Rich for the real world use case. Other youtuber are just building snake games.

Thanks for Watching Links Notebook: bit.ly/4eyKPbs Best Repo for you: bit.ly/4bWraQA AI Agents With Memory Video: ruclips.net/user/liveL49AxDcOURU?si=uBj8SuvCIJdVKUoz

Your channel is a gem. Love this content 👏

Are you running it locally or in a cloud environment?

Thats nice. You run that locally or set up in cloud environment?

Please I how no one from the stackup bounty challenge is here, because we are going to have a big problem😂😂

Sir, Im trying to implement your code, but I found an other error, this part does not work, if I find a solution by myself I paste here: I stucked here: from llama_index.core.node_parser import SentenceSplitter parser = SentenceSplitter() nodes = parser.get_nodes_from_documents(llama_documents) for node in nodes: node_embedding = embed_model.get_text_embedding( node.get_content(metadata_mode="all") ) node.embedding = node_embedding this line just loads and loads and does not go forward: node.get_content(metadata_mode="all") Maybe the problem is this: I'll check later: github com/run-llama/llama_index/issues/12200

At the second step of the article has problems

Where are you from Sir? I like your accent and the vid is great too

Thanks for your clear explanations and moderate pace. I appreciate that you're not rushing in your videos and taking the time to explain each step. Cheers mate!

the google collab does not work after this step: collection.delete_many({}) "SSL handshake failed:..."

Thank you so much for this brother <3

Great content 🔥. subscribed 💯

Wonderful conversation - nice questions and amazing answers. Thank you planning and organizing this video. +++++++

how to reuse these embeddings so that we don't have to recreate them?

hey, iam doing the same thing. can i ask you specific questions ?

ayooooo appreciate how thorough you are salute!

Great video! So if I understood this correctly, RAG basically uses an external vector database to retrieve first the most relevant information performing a similarity search, then grabs this information and it "appends" it to the user prompt resulting in a larger prompt with better contextualization, am I right ?

quick and clean, top marks

Hello, thanks for the video! I get an error ServerSelectionTimeoutError when I execute collection.delete_many({}) in spite of having a successful connection to MongoDB in the previous step, do you know what could be the reason? Thanks!

hi brother , where are you located , I am in Dubai . How can we connect . Linkedin , IG will be fine

I like video but you have to explain each and everything about code.then beginner will be understand why that line of code is written .

Thanks for watching 🧾 Article: mdb.link/rag_claude_mongodb 💻 Code: bit.ly/3TqQcB1 📈 Hugging Face Dataset: huggingface.co/datasets/MongoDB/tech-news-embeddings

Hi there, this has been really helpful! I have a question though. What if instead of the "plot" you have a huge file. And let's assume we created the embedding for it at OpenAI and now we have it in the search results. Would adding it to the completion query at OpenAI cause a problem? Since the context is gonna be huge. Or am I missing something? This is what I mean. Imagine the completion api is being called with a user query as such: "Answer this user query: " + query + " with the following context: " + search_result What if the search_result is the contents of that huge file? You see what I mean? Thank you!

Great article :) awesome.

new to channel, good stuff

Im building an MVP for my startup and was on my way to building this as a way to search a database and this is a great start.

4:13 PALM stack? Ok I can dig that terminology

@2:30 I vote AI Engineers. A nice simple title (not to be confused with ML engineers)

A MongoDB vector database isnt as widely covered with RAG. Well done

🧾 Article: mdblink.com/rag_with_gemma_hg... 💻 Code: bit.ly/42MMjJS 📈 Hugging Face Dataset: huggingface.co/datasets/MongoDB/embedded_movies Thanks for Watching.

THis is great thank you for sharing! Very clear and concise instructions I think. I'm trying to wrap my head around the following use-case for this, where the I was wondering if you (or any other) would know how to transfer this very clear instruction to two similar-but-different use-case, I find this helps me understand better : SO in this example, we have rows with a string column that we convert into embeddings and use with the mongoDB. So far so good. These two examples vary from your example. the one closet to it is also most similar to your example: 1) a bunch of PDfs with academic articles about the same research topic. I'd like to ask questions about the topic. 2) a whole book from the public domain? It's 1.000 pages, so plenty of text. I would like to 'ask questions' to the book. So in the first case, one could make a dataframe where the rows would be articles (one per article), and the string could be either the abstract or idealy the entire article. Would this be sufficient for somewhat complex inquires like 'give me the different definitions of worksite inequality from the articles', or 'list the datasources from the articles'? In the second example, I'd like to ask questions about the text - such as 'how do the protagonist evolve his convictions about Man as the ruler of Nature'? here, I would be in doubt about how fitting this example is to this use case, since there is no obvious way of making rows of data - perhaps by chapter, or even by 5-page intervals? Seems like it could get very expensive very quickly. I know this is a lot to ask - just putting it out there, hoping someone has the time to help me understand the different utility of this. Anyway really good video, it's already gold! :)

Very concise and informative. Good stuff. Link to github repo is broken !! Please fix it. Thanks and keep up the good work.

🧾 Article: mdblink.com/polm_ai_stack 💻 Code: bit.ly/3UJVbOc 📈 Hugging Face Dataset: huggingface.co/datasets/AIatM...

Thanks for watching 🧾 Article: bit.ly/47YqOGQ 💻 Code: bit.ly/3u7duSV 📈 Hugging Face Dataset: huggingface.co/datasets/MongoDB/embedded_movies Notebook: bit.ly/486IYpW

I understand writing is learning, but I imagine Bex could also have a very successful youtube channel teaching! Alternatively, would love to see Bex do some cross promotion collaboration videos with other programming/teaching legends.. some of my favorite are: sentdex , George Hotz, Corridor Crew, Keith Galli , Olivia Sarikas, SECourses, and bycloud!

This podcast with Shashank Kalanithi is a goldmine of insights for businesses! From data analytics to finance and productivity, it's a must-listen for anyone looking to thrive in today's data-driven world.

Had such a great time on the show!! Thanks again for having me Richmond :)

TIMESTAMPS: 00:01:18 Introduction to Eva 00:02:18 Eva's educational background and internship 00:03:54 Attending a conference as an intern 00:05:18 Electroencephalogram (EEG) 00:05:33 Exploring educational opportunities outside of Spain 00:07:12 Mind Controlled Wheelchair 00:08:19 Trash Picking Robot 00:10:33 Brain Computer Interfaces 00:13:38 OpenBCI 00:15:20 Galea and the synergy between VR and BCI 00:18:06 Invasive vs Non-Invasive approach to BCI 00:19:56 Time estimation on the Commercial introduction of BCI 00:26:27 Programming Languages used to build and use BCI products 00:27:41 Galea price points 00:29:54 Day to Day of A BCI Software Engineer 00:31:49 User Feedback after using BCI hardwares 00:32:46 Meditation and Brain Computer Interfaces 00:33:49 Practicality of BCI headgears 00:35:51 Standing out as an Intern 00:38:06 Values of OpenBCI 00:42:10 Crazy and Interesting BCI ideas 00:44:58 Handling Large Quantities of Data from Sensors 00:47:55 Movies that bring BCI to life 00:53:30 Operating Systems of the Mind 00:59:05 Mental Health and Dealing With stress 01:01:30 Surviving as a Hardware Companies In The Remote Era 01:04:31 Diversity In Tech 01:08:02 UniPeers 01:11:16 What degrees gets you into BCI Roles