Thanks for the feedback on the hyperX, and glad you enjoyed the video

Thanks, glad you liked it

The original video from your channel was very cool. I just added my own personal perspective.

I'm actually Automating the processes. Anyone that has used a tool like VidIQ or a video editing software is Automating processes as well, this is not any different from that point of view. I checked out your channel and your videos look great, thought it seems like a couple of years since you last uploaded, I'm curious why?

@@AppyDave I just did not find the time to do more! I am always putting ways too much time into this stuff! I am having a lot that I would like to share in KZbin but its too time consuming, so I mainly post on Instagram! Shorts did not seem to work for me! I also build a couple of bots to boost my reach on Instagram but refrained on it later because social media will punish a channel so the plattform will not be overrun by this type of content! Unless you are on TikTok probably! 😁

Great, I'll be doing a few around my various KZbin Automation workflows

:) glad you liked, the next video I put out will go more into the DSL

Glad it helped!

I have more videos coming out on this topic

Getting there 😁

@@AppyDave same! Just gotta get those watch hours! Lol

Glad you liked it!

Glad it was helpful! It's an easy set up.

Thanks

:) cool, glad you liked

Yeah, it is an easy setup

Thanks Luke :)

@@AppyDave 👊🏻❤️

I tried generating funny images with Donald Trump randomly because I wanted to create some humorous work that I thought Americans might find amusing from my point of view. However, I don't take sides in politics at all. I just find Trump to be a funny guy in terms of memes and thought of testing it out. All the prompts were blocked, and weeks later, I was poorly received by MJ support on Discord. The next morning, I generated some images involving the American flag. No mockery, just beautiful images involving America. In summary, I didn't do anything against the rules, but I believe the reason for my ban was that I was perceived as a supporter of one political side or the other. The support team gave me an empty and nonsensical response, and I also lost money because there will be no refund, and I won't be able to use the remaining credits from my plan. If you have any alternative suggestions, I'm open to them. I don't intend to use MJ anymore, and I'll do everything I can to oppose this company! And I’m not American.

@@hakicapivara American politics are a joke. It's all Coke vs Pepsi nonsense that serves the interests of rich assholes. It's frightening that a corporation decides what's acceptable and what's not to generate. Sure it's AI art now, but how long until you won't be allowed to criticize people in power with any medium? I've been messing around with Flux Pro through Glif and Civitai. I've already made some cool stuff. Have fun!

I too am doing everything to appose the company, I feel they are have no integrity and today decided to start using Leonardo.AI as my main tool, I will do some videos on it when I have some spare time.

Thank you for sharing your experience! It’s always helpful to hear from others who have faced similar challenges. Wishing you the best of luck in finding a more reliable service!

I have decided to move over to Leonardo.AI

I agree, I have had a few issues with paid services causing me grief , they should be accountable and thats why I created this video, I can't make much of a difference, but every little bit helps

@@AppyDave Thank you for continuing to make content that informs us all on the latest and greatest AI apps out there. We all need to know which service(s) is the best and most fair and your perspective really helps!

@@jonhobbs5113 Thanks :) its hard work getting these videos out, but it is a passion project

I have not tried it out, but I will put it on a list to look into to

There is a download button in the top right hand corner

Glad you enjoyed, the play testing variables is a good idea

I appreciate that!, thanks for the feedback, is there a type of video that you are interested in seeing more off?

Its a shitty feeling and costly

@@AppyDave they say that they will refund me

@@sajeebmekdad4328 great to hear Sajeeb, I'm about to make a 2nd request to them as well, will let you know how I go

Thanks for the comment, and glad it hellped

Thanks Alec, appreciate the complement.

True

Hi David, Graphics dimensions can be controlled in some tools and others can use capabilities like aspect ration, eg. 16:9 Core engine is usually Stable Diffusion or some based on DALE3. I don't feel younger, but love the compliment. :)

@@AppyDave thank you

Glad you liked it!, let me know if there are any topics in Prompt Engineering you want to see covered

Glad it was helpful, let me know if there are any prompt engineering topics you are interested in

You are so welcome!, glad it was useful, let me know if you want specific prompt engineering videos.

Yes, there is the use case that makes sense

Thank you :)

Glad it was helpful Susan

Good advice, I dropped a link 🔗 too the repo, still getting the hang of KZbin videos and the hundred little details with each upload

Thanks , it's not my normal content style but I thought it was needed

cool

@@AppyDave i since updated this method : As you only need the two functions : Method : 1 one to executecode on the system... and one to save/load a file from your tools folder : So you can allow the model to generate any function for itself :( best way ) and it can add the tool to its tool collection : then it can execute the code : this way is a beautiful way , as i used the stop and think, then act technique .. if it needs to act it will execute one of its tools or create a tools : this method really does work well : as it can generate bash code and python scripts etc: you can just create tools and add them to the tools folder ! < it will load these tools into its tool box ! > Method 2: This is simular to the above but it can only execute a tool , you supply the tool names in your code and it searches the docstrings for each function and adds it to its toolbox : ( i found that the model does not like the pydantic SLOW!! ) << the docstring method makes your overal code lovely and strong : the model knows by your docstring how to operate its functions and tools : Personally i have the model making its tool call behind the scenes so it can make the call and return only the final output to the use as the map it creates ( the list of tool names and , funciton names ) is enough to parse to see if the funciton is i the box then is can be called by the model its output returned and it can continue with its response generation , so if it needs a multi function it should pla and execute in the correct order storing the values in its Response history ( messages and extract them for its final response ( ie self conversation )... This method waorks also but the above method does incorperate this: Method 3: Yes the Notebook Method: so this method does not store tools it just generates functions and scripts and you can supply a list of pydantic functions: again the main internal fuciton is to execute a code : here we use e2b sandboxes to create a session with a notebook and an environment allowing you to execute code in this surface hence work on a project : Sadly i did not alow it to save these books ( maybe it does in appdatatemp files ... but you can reload a session if you know the session id: and the final output is saved to your working folder : the main thing with funciton calling despute the methodology you use : its important to let the model Tell you in the response it is executing a funciton : so you can intercept the respose and return the answer: the answer shoud containn the result and aknowedgement that this is the response for this funciton call :: then iterate through your function list ( for error checking and execute the funciton from the list with the parameters given : For me method one is super good working well as an agent : Once you can execute code on your system then you have a open interpretor : So why give the model functions ?? .... API CALLS !< these contain your api Keys !!!> Only! If you need to search a website etc you do not need an api you need a llm !!<< it knows how to do this without paying money ! I recently spent time training my model as a function caller with many different datasts and aso i trained for being a planner... so it can be a planner only not solving tasks but providing recipes for tasks: Then i retrained it to the messaging format template ! this enabled for the model to function correctly : the most important is the model is using a messaging format template as oter output types will not work : so if a mpodel has been trained for alpaca or normal mistral you need to change the chat template for the model to be the Chatml ... OpenAI format ! then it will perform exactly as you require even if you convert if to gguf: i was having problems with out put when i realized that the chat templates are deterimental to a model ! : so for normal use ( non messaging ) then you should remove that template and put a Q/A template then make the gguf and it will perform in GPT4all etc correxctly ! << the messaging model will not perform in these apps ... they will only perform on setrver setups... so if your havinbg problems making OpenInterpretor and other OpenAI (open source stuff ) then it is this !! <<

@@xspydazx Wow, you've really delved deep into this! It sounds like you've found some effective methods for using Open Interpreter. Thanks for sharing your insights!

@@AppyDave yes it was quite hard to figure the whole thingy out as the main this was just one function !! << the execute code and return it to the model >> the rest was logistical in fact ( like why did it not work with MY model ) << on MY tiny system ... So after this ChatTemplate was solved then the model begun to work with all my ollama/lmstudiio projects : So the chatML template for output on fie tuing was the also most crucial aspect < hence only these models work on funciton calling the chat models must use the pydantic method ... (which was too slow) i als realised that my Opus model (Manager) should have stable diffsuion and other specialist model as tools : And infact we also should make the RAG be a tool and ot be an input stage , so if the model needs to use rag to get a richer response it can do a rag ! , if the model feels the rag is generating most of the responses it can generare a dataset from the rag for fine tuning ... if the model has dataset for fine tuning it can do a quickiy train of the small rag dataset!.... So our model is a manager Agent and we should train it to be a manager... ie lots of planning an grading: LOL: