Відео

Thank you for the comment. I have done RAG video before. Gonna make some more in the future.

I have discord link in the main page. But now the activity level is low. You may also find my twitter and other social media account in the main page.

This is a good question. The model output gives only the depth map. Evaluating the model or further analyzing the result requires additional steps. I have not discovered these yet.

All videos manually edited, reference text selected from reference documents, and let notebooklm generate podcast.

感谢建议

Thank you. But your question I suppose is about cryptocurrency, which is an area I will explore but currently I can not offer you any advice.

github.com/LearnByDoing2024

不客气

I never investigate the right pronounciation before, and I am used to read words in italian style. If you can share the right pronounciation I would appreciate.

Claudia

@@DigitalInsipire creative

github.com/LearnByDoing2024/UA-cam/tree/main/20240912_iphone_news_sentiments

Thank you for the comment. Yes the models could be biased. So far I can not find source to evidence, for example, for a specific topic which training data was used, to train any models. We need to be cautious to verify facts.

Thank you very much.

Hi, I should have mentioned that, the nested chat has two examples in the autogen site. At first I followed microsoft.github.io/autogen/docs/notebooks/agentchat_nested_sequential_chats/. I think in this way it is easier to manipulate. Or, following microsoft.github.io/autogen/docs/notebooks/agentchat_nestedchat/ could require more control in the process(terminate conditions microsoft.github.io/autogen/docs/tutorial/chat-termination/). If you ever tried the code critic agent flow, you might realize that if the grading of code quality increases at small margin each time, the talk can be very long, which could further evidence that talk control is important. Another way is adding user input, to decide termination or not. In all, I assume that model capability changes the conversation length(if max round not controled). Then, you may set terminate conditions(I call it somewhat soft controls). Finally, human input/intervention.

I've tried termination messages but since it is a sequence of chats it goes into the third chat also and for human intervention my case I don't want to use it so I saw the source code of autogen conversable agent and I'm thinking to manipulate a method in autogen and use it but yeah I'm looking into any other possibilities

Chosing groq, connect, it will show you the option to insert API key.

Gonna upload here later. Stay tuned. github.com/LearnByDoing2024/UA-cam/

Gonna post it here. Stay tuned. github.com/LearnByDoing2024/UA-cam/

Thank you for the comment but i do not get it.

​@@learnbydoingwithsteven It's a mis-pronunciation of Obama by Trump and they made meme out of it .It's not Relevant to this Topic.

How to get code and also can we build an agentic system which can analyse the sentiment based on the real news for a particular stock. It should have access to yahoo finance to get stock prediction of a particular task and also real time news on that particular stock to match the prediction with thr ground truth

​@@mohammedshuaibiqbal5469 Yes I agree with you. Below are some additional comments. So far I only use freely accessible data source from different python packages. If you want more detailed data, either numerical or text, usually different companies can provide reliable api, but it has a cost. Apart from numerical data, text data has two issues. 1, identify reliable data source is difficult, there are almost no 100% objective source without bias. 2, sentiment analysis is a part of NLP, there are many ways to get sentiment score, but like any models, the result need to be carefully examined and utilized. Stock prediction, I would suggest to compare different tech trade strategy with machine learning algorithms to make prediction. Usually you get raw data, without prediction. But if you have trusted third party analytical tools you may get predictions right away. There are many existing agentic framework, using different packages like autogen, crew ai, or even DIY frameworks. But they use different data API, different function calls, to orchestrate the workflow.

Just added. Thanks.