How run 2 instance of llama3.1:8b at the same time? Thank you so much.

This no longer work

Great Video!!!

hi, can you update this so we can use LLama 3.1 instead, please provide a version that works with Apple silicon as this one fails on my M3 Mac

excellent video 'but' you go to the install of a LLama3.1 405B, excellent. Install this into VS Code, really great. 'but' I am not sure if putting up OpenVoice.git runs agains / using the llama3.1? please verify, sorry i am a newbie at python also. thanks.

and ubuntu?

Thank you so much, this is insane!!

I'm just trying to understand here. How is this different from let's say put all that tool information in the system property of /generate? In the end the LLM decides to use it or not, there's no turning back but to adjust the prompt. It also does not take the json result and generate a coherent sentence after like a normal chat would (ex: the trip will take X amount), unless you would run that after once more with the json result, just for that.

nice

Please create a video on Using Open Source Models in production to create Multimodal RAG Chatbot using private data

Very Well explained ! Would love to see more videos of this series. Also when we can expect the Open Source RAG ChatBot for private data. Please try to make it asap we all are waiting for your amazing videos with great explanation

can you please add a video on how to change ui of ollama web ui i want to do some minor changes please add a video i see there are not more videos in youtube for this content so it will be best for you to add this video ;)

What a giant pile of bullshit. You can't just "generate" research with a machine learning algorithm. It can only generate remixes of things it has been trained on. What a stupid idea. The last thing science needs is more AI bullshit.

Good stuff brother

can you make a video on integrating ollama (local llama 3.1) with MS-teams to do note taking and summarizing the meetings afterwards ? Thanks

Great video. PLEASE drop the background music. Higher speed review of your videos is ruined with it.

When I run python3 ingest.py, I am getting an error ModuleNotFoundError: No module named 'chromadb'

It absolutely refuses to use the tools. It keeps going to the basic llama3.1 llm to answer all my queries. When given two airport codes that llama3.1 could not resolve, but were in the database, it just gave search engine recommendations.

Help me understand - in the antonyms example, you have a get_antonyms() function that can optionally be used if the solution is found within get_antonyms() function. This would just be a classic expert system, 'software 1.0' use case. If antonym is not found in get_antonym(), the LLM can just return its LLM result instead. This would be a 'software 3.0' use case. But does the LLM use the contents of get_antonym() as an example too, so its context is extended or prompted by the antonym function's example contents? Thank you for the example!

I would really appreciate it if you focussed on explaining concepts and code rather than going through the installation process of something that you have already created and written.

Thanks!

how can I load an already existing pyhton folder in my local drive? thank you very much for the video. I really apreciate it!

when I deploy the gpu, do I start spending immediately? or do I spend only when I execute code?

Humans are never been as close to ww3 as now. Humans are the cause for most dangers and pollution and bad influences for nature. Not AI

It is irrelevant. Powerful intellect without consciousness is much more dangerous for humans.

Microsoft has a standardized AI Chat Protocol API. That's what I use. Makes it really easy, especially when you want to make different LLMs chat with eachother.

Like these ‘into the details’ videos, thx

There will never be AI with consciousness. Fullstop.

It's becoming clear that with all the brain and consciousness theories out there, the proof will be in the pudding. By this I mean, can any particular theory be used to create a human adult level conscious machine. My bet is on the late Gerald Edelman's Extended Theory of Neuronal Group Selection. The lead group in robotics based on this theory is the Neurorobotics Lab at UC at Irvine. Dr. Edelman distinguished between primary consciousness, which came first in evolution, and that humans share with other conscious animals, and higher order consciousness, which came to only humans with the acquisition of language. A machine with only primary consciousness will probably have to come first. What I find special about the TNGS is the Darwin series of automata created at the Neurosciences Institute by Dr. Edelman and his colleagues in the 1990's and 2000's. These machines perform in the real world, not in a restricted simulated world, and display convincing physical behavior indicative of higher psychological functions necessary for consciousness, such as perceptual categorization, memory, and learning. They are based on realistic models of the parts of the biological brain that the theory claims subserve these functions. The extended TNGS allows for the emergence of consciousness based only on further evolutionary development of the brain areas responsible for these functions, in a parsimonious way. No other research I've encountered is anywhere near as convincing. I post because on almost every video and article about the brain and consciousness that I encounter, the attitude seems to be that we still know next to nothing about how the brain and consciousness work; that there's lots of data but no unifying theory. I believe the extended TNGS is that theory. My motivation is to keep that theory in front of the public. And obviously, I consider it the route to a truly conscious machine, primary and higher-order. My advice to people who want to create a conscious machine is to seriously ground themselves in the extended TNGS and the Darwin automata first, and proceed from there, by applying to Jeff Krichmar's lab at UC Irvine, possibly. Dr. Edelman's roadmap to a conscious machine is at arxiv.org/abs/2105.10461

Thank you for the video. I'm new to LLMs, could you explain what's the role of an LLM in the process of the function calling? Is it a flexible user query? Or we just added more capabilities to existing LLM?

This seems like a fine-tuned routing llm. Function calling is a bad but acceptable term for JSON output. The llm is not calling any function. Just venting. Lol

Will it recalculate embeddings everytime I add more pdf documents?

🫰🫰🫰

I cant find the Deployment URL as illustrated. Where do I check them ?

please tell me how can run local-server in my server and use web gui chatbox (or somthing have like copilot or...) in my cleint (or any computer in my network) tnx

Thank you for the great video! I'm planning to create a chatbot using LM Studio for personal purposes in the same way as the image you uploaded. In the image above, it seems that the chatbot is implemented without inputting or learning a separate dataset, but how can I input the dataset I prepared based on this image and implement it? And when implementing a chatbot like the method shown in this video, in what format should I learn or input the dataset so that it can be implemented smoothly?

I get the error running when python app.py: pydantic.v1.error_wrappers.ValidationError: 2 validation errors for NVEModel base_url field required (type=value_error.missing) infer_path field required (type=value_error.missing) What is the problem please?

eh, your RAG system you developed is still the best one I found and use..........

Hello! I have followed this, but I have a problem at the end. On visual code in "docker exec ollama_cat ollama pull mistral:7b-...", there is "error response from daemon: no such container: ollama_cat. How can I resolve? Thanks

It does not beat OpenAI, Gemini, Claude... it's cool but not that smart at all

BG MUsic too loud in thr start

Why does it start answering before the person is done talking? It makes it look fake. It seems like the AI is a recording and the people asking questions aren't taking fast enough to match the recording.

I have developed a similar software. The LLM changes according to the question you ask. For instance, when a mathematical question is asked during a conversation, it automatically connects to the GPT API for mathematical calculations. When an informatics question is asked, it connects to the Claude model. And it can write code on its own. For example, when you say "Get the first 100 products on my website," it writes the necessary Python code, runs it, and retrieves the code. One of its most important features is that it has memory. Regardless of which language model it connects to, it can remember everything with the database architecture I have set up. It constantly monitors you with a camera and performs sentiment analysis. These are recorded in the memory, and its behavior towards you changes because I have set it up so that the prompt changes automatically. I am in the final stage now. I have integrated uncensored LLMs. By uncensored LLMs, I mean that it directly answers questions like "How can I easily defraud someone" without requiring prompt engineering. And of course, you are talking to an avatar on the screen. I can't wait to publish my project. Kyutai's post inspired me about my project. I will develop them and write them here as a comment. Maybe I will even send it to you to try out.

Great content. Thank you very much.

Really bro?!

Quite good.

how you can limit users to not overload my telegram bot with queries when they decide to click around too much?

how to handle rate limiting for users who gonna overload my service with queries? is there a way to build inside delay for every query so even for users with a unlimited tear he cant overload the system with queries since he can only do 3 concurrent jobs and every job needs a set 4 minutes (but technically it only needs 2.5 minutes)

what kind of vector database is used for the RAG in this extension?