Thank you so much for making this video. I never thought you would create a video on this with such little time. I love your work and I think you are the best AI guru out here on RUclips!
@@fahdmirza I managed to get the project done. So now how to run this? Can we use LM Studio to run the model we created? Or can we use the playground in the VS code to run it?
This is a great video but I am having difficulties setting up my environment correctly. I keep getting a "Conda not detected" error in my WSL environment but I am able to make calls to conda in the command line. I updated the settings.json file in the ubuntu environment all sorts of things. It seems like an issue with the AI Toolkit not recognizing conda. Thoughts or advice?
Can you do another video on how to test out this model? I manged to create the project but I am lost there. I couldn't not find any details on how to run it. I think the playground is supposed to be running them but it is not working. Can you create the next part of this video showing how to actually run these fine tuned models? Either on Hugging face or locally.
@fahdmirza sorry, I don't think I was specific enough. What I mean is. Generating a synthetic dataset, that is seeded by a codebase (like actually code files) to fine tune an llm on a codebase. There is a github project called your_devin from a hackathon, but it doesn't look complete. And I haven't been able to find anything else that comes close. Thanks for the reply 👍
You should have shown what's in the input and what we have achieved by doing a before after example demonstrating the the improvement. We don't even know what happened and what was achieved.
I am not sure what you are referring to here. The purpose of video is to show the process of how to do fine-tuning with this ai tool kit and that's very clear from start to finish.
I think , he meant you should first ask the question to the LLM than train a model with your own dataset and then again ask the same question to the LLM and see if it gives same response as before or gives more refined response after fine tuning
🔥Install Visual Studio AI Toolkit Locally in Windows ruclips.net/video/7dZjUH28EyA/видео.htmlsi=WaiKc-Sfb5WGTECo
Man, you are on this stuff so quickly! Thanks!
thank you
Thank you so much for making this video. I never thought you would create a video on this with such little time. I love your work and I think you are the best AI guru out here on RUclips!
You are welcome
@@fahdmirza I managed to get the project done. So now how to run this? Can we use LM Studio to run the model we created? Or can we use the playground in the VS code to run it?
This is a great video but I am having difficulties setting up my environment correctly. I keep getting a "Conda not detected" error in my WSL environment but I am able to make calls to conda in the command line. I updated the settings.json file in the ubuntu environment all sorts of things. It seems like an issue with the AI Toolkit not recognizing conda. Thoughts or advice?
would need to check when time permits
how do we query the fine-tuned model? and post this how can we deploy said model to an app or a website?
Thanks
would need to check when time permits
Can you do another video on how to test out this model? I manged to create the project but I am lost there. I couldn't not find any details on how to run it. I think the playground is supposed to be running them but it is not working. Can you create the next part of this video showing how to actually run these fine tuned models? Either on Hugging face or locally.
Noted, when the time permits.
Thanks.
You're welcome
Can you find a good synthetic dataset generator for a codebase?
plz search channel, there are already many videos around that
@fahdmirza sorry, I don't think I was specific enough. What I mean is. Generating a synthetic dataset, that is seeded by a codebase (like actually code files) to fine tune an llm on a codebase. There is a github project called your_devin from a hackathon, but it doesn't look complete. And I haven't been able to find anything else that comes close. Thanks for the reply 👍
@@figs3284 I'm also looking for the same thing to generate synthetic data using LLM. Let me know if you find any resources.
You should have shown what's in the input and what we have achieved by doing a before after example demonstrating the the improvement. We don't even know what happened and what was achieved.
I am not sure what you are referring to here. The purpose of video is to show the process of how to do fine-tuning with this ai tool kit and that's very clear from start to finish.
I think , he meant you should first ask the question to the LLM than train a model with your own dataset and then again ask the same question to the LLM and see if it gives same response as before or gives more refined response after fine tuning
@@drmetroytthat was not the purpose of video. I have countless other videos showing that.