You respect people's time while also giving quality content. Just what I was looking for!
@smellslikeupdog80
Жыл бұрын
I saw this video and immediately subscribed. I want to see what I can learn from you for building chatGPT applications. this is incredibly useful already
@vrsen
Жыл бұрын
Thank you! A lot more to come!
@dripdrops3310
Жыл бұрын
Thank you for taking into account attention spans as short as mine. That was really dense in information.
@StephenRayner
Жыл бұрын
Feedback: reduce volume by 10-20% Increase lighting, contrast of background to face is not cool. Remove sound effect of images sliding in or reduce by 10-20%. Same with clicking sounds Video editing, transitions are on point. Assets shown are relevant and support narrative. Nice work. Scroll speed on web pages shown are good. You talk non stop, which I like. But top left show a heading for each paragraph you are discussing. So we can stay focused.
@StephenRayner
Жыл бұрын
I need a moment or two to take in. You say main goal is to… around 3:10 ish. Display the text “main goal” as you transition and say this.
@vrsen
Жыл бұрын
Got it! Thanks!
@JoeRepka
Жыл бұрын
Best video I've seen. Eliminates most of the reasons I don't usually watch videos.
@siddhantjagdish2554
Жыл бұрын
Thanks mate. Crisp and clear to the point, Keep on making such videos. Kudos!
@juliovac2113
Жыл бұрын
Man you are so much better than most channels honestly! I learn so quick vs these content creators that spend 45m to tell you what you know and teach you to use an API key for a python based chatbot only haha. Anyways came to ask if I can integrate this pinecone memory with the other script you have for searching with google api, for a ongoing conversation with access to google queries?
@vrsen
Жыл бұрын
This is the exact reason I started this channel!
@vrsen
Жыл бұрын
Yes, you can certainly add memory to almost any use case, although it might require some tedious prompt engineering sometimes
@ashishjoshi357
6 ай бұрын
Bro you saved my ass ,it was just what i needed!
@samario_torres
Жыл бұрын
Thanks for content. I think turning the music down just a little bit would be a good idea.
@vrsen
Жыл бұрын
👍
@gelbuilding
Жыл бұрын
Great videos, Very detailed and concise. Thank you.
@SF-gn6uw
Жыл бұрын
Clear, excellent
@kenw9722
Жыл бұрын
Your videos are great, I hope you continue to create this type of content. Thank you!
@vrsen
Жыл бұрын
Thank you! I will!
@mahandolatabadi2600
Жыл бұрын
so great, keep making videos.
@kunalsoni7681
10 ай бұрын
undoubtedly :) great demonstration
@jojomondag
Жыл бұрын
Straight too the point im impressed!
@devcore1
9 ай бұрын
What service do you use for the stock videos?
@Yogic-ignition
10 ай бұрын
How can i add sources link, like from where this answers were generated from my knowledge base?
@831digital
Жыл бұрын
Wow... good info, no BS. Subbed.
@atylerblack164
Жыл бұрын
You have shared very valuable information, I hope you have great success in your channel!
@vrsen
Жыл бұрын
Thanks man! I’m trying. Please share if you’d like to help
@navinmahendran
Жыл бұрын
Love your style of videos!
@maximebeaupre8396
Жыл бұрын
Very well explained.
@baileyfaulk991
Жыл бұрын
I am having an error when I run the query. In the line that says: prompt += "Context: " + " ".join(chosen_text) chosen_text is a list and does not want to join to the string. You may want to fix this on the attached code.
@vrsen
Жыл бұрын
This line is correct, maybe not all elements in the list are strings. Check how you parse the document or convert them to strings: chosen_text = [str(text) for text in chosen_text]
@appsdoolot9602
Жыл бұрын
❤love it. nice explain. I hope you have great success in your channel!
@vrsen
Жыл бұрын
Thanks, let me know what you want to see next
@robertputneydrake
Жыл бұрын
sposibo bolshoi!!!!!!!!!!!!!
@sethhavens1574
Жыл бұрын
great vid dude - what did you use for the front ends? would be good to see some tutorials on that but too 👍
@vrsen
Жыл бұрын
Next JS🖤 Yes, in the future
@devcore1
9 ай бұрын
Is your client's work public online somewhwere?
@filmikathe1491
Жыл бұрын
Really great! Just one help, can you please say where can I find the "Pinecone Endpoint" code? I required it for another project. Love you content❤.
@vrsen
Жыл бұрын
Just wrap the last part of the notebook (where you get vectors and ask gpt) using your backend framework. Make sure to copy all the packages as well.
@vrsen
Жыл бұрын
You are trying to create a backend endpoint with this code right?
@filmikathe1491
Жыл бұрын
@@vrsen I was actually trying to clone a repository and it needs my OpenAI key, pinecone api and a pinecone endpoint. I just couldn't find it on the dashboard. Thanks for replying!
@quantumbyte-studios
Жыл бұрын
I keep getting problems with the dimensions not matching between my data and pinecone, any suggestions how to conceptually understand this?
@vrsen
Жыл бұрын
Probably your embedding model and your index have different numbers of dimensions. Openai, for example uses 1536.
@mpaforoufakis
Жыл бұрын
well done! subscribed
@Govi-fu1ti
Жыл бұрын
Amazing, Keep the Good work
@vrsen
Жыл бұрын
Thank you, I will
@rissalatahmed6719
Жыл бұрын
Banger video! ❤️
@google-mudak
Жыл бұрын
Спасибо! По твоему коду удалось провести индексирование и векторизацию
@clayallen_art
Жыл бұрын
This could have been a great tutorial, but the background music is way to distracting, and its far too up front in the mix, its in the same spot as your voice. I had to mute and read subtitles, but then I had to constantly pause to see the infographics.
@vrsen
Жыл бұрын
Thank you for feedback. Is there anything else?
@Maisonier
Жыл бұрын
Amazing video! like and suscribe.
@pplytas
Жыл бұрын
Nice one!
@vrsen
Жыл бұрын
Thank you🙏
@andrecassal
Жыл бұрын
thanks for this video, ... one question that I cannot seem to have a good answer to... I understand we need to upload the data to pinecone... but when we connect pinecone with openai, does any information gets sent to openai API? or is the data processed on-premisse?
@vrsen
Жыл бұрын
No, pinecone is not connected with openai api in any way, but since we are creating embeddings using openai, you do have to call openai api both for upload and for querying your pinecone index.
@vrsen
Жыл бұрын
You are free to use any other LLM, however, instead of openai
@arexperiences1169
Жыл бұрын
amigo una pregunta cuando dices que vas a insertar el chatbot en el sitio web de tu cliente a traves de un como se hace eso?
@vrsen
Жыл бұрын
Sorry don’t speak Spanish
@rezabaram400
Жыл бұрын
Why so much sounds and effects? It's distracting and hard to follow. You're welcome.
@chanduramayanam236
Жыл бұрын
document can be how many pages maximum?
@vrsen
Жыл бұрын
As many as you want, but for super large loads, you have to increase the number of pods on pinecone
@gurgelihans
Жыл бұрын
Does memory mean that I only have to "teach" the model once with the content of my word document? And for all future prompts (to e.g. OpenAI) to retrieve information from the word doc, I don't have to send the document content and basically reduce the number of token to a minimum?
@vrsen
Жыл бұрын
Not exactly. The model will pick out only the most relevant pieces from your documents, but you still need to pass them in the prompt when you want to generate a response.
@vrsen
Жыл бұрын
You do have to upload only once, though
@tidharcrimow6534
Жыл бұрын
can you please provide the word document needed for running the program?
@vrsen
Жыл бұрын
I’d be happy to, but this data does not belong to me. Try searching the web or using any word documents of your own. It should be more fun this way!
@Evox402
Жыл бұрын
Very good tutorial, I finally understood how pincone and GPT work together, thanks a lot :) I left a subscription 😁 Maybe ideas for the future or at least stuff I have problems understanding: - Conversation Memory - Agents and their usage - Langchain, LlamaIndex and possible other tools - Maybe how context handled in chatGPT and should be handled ones own applications
@vrsen
Жыл бұрын
Thank you for advice!
@appaesthetics1277
Жыл бұрын
Pinecone reminds me of glands
@skirankumar000
Жыл бұрын
How to contact you there is no email in the about section of your KZitem channel
@vrsen
Жыл бұрын
Click “view email address” please
@hansofmadata3565
Жыл бұрын
Have you tried creating sparse dense hybrid vectors?
@vrsen
Жыл бұрын
No, not yet. Is the performance boost significant?
@hansofmadata3565
Жыл бұрын
@@vrsen I started working in this space about a month ago. I’d watched some of the Pinecone videos to get up to speed and with everyone at Pinecone praising the sparse dense indexes I just dove right in with the hybrid index solutions. Compared to some Langchain tutorials I did using dense only I’m very pleased with the results. I’d be very interested to hear your thoughts if you do try them out since you seem to have a lot of great knowledge in this field. Thanks for the video and the chat!
@josephshaff5194
Жыл бұрын
🤣
@vrsen
Жыл бұрын
✌️
@AndrewOrekhoff
Жыл бұрын
Amazing tools and review! Please tell me how to contact you?
@vrsen
Жыл бұрын
Use my email on about page of my channel
@GayathriG-h5h
Жыл бұрын
chunks_with_embeddigns = [] for chunk in tqdm(text_chunks): embedding = get_embedding(chunk, engine='text-embedding-ada-002') chunks_with_embeddigns.append({"text": chunk, "embedding": embedding}) I did whatever you have mentioned in the video but for this above code i got this error , RetryError: RetryError[]
@vrsen
Жыл бұрын
Make sure your chunk is a string and embedding is a list
Пікірлер: 80