Hi sir, I want to create an AI power chat application which will query with MS SQL server database. Is this possible to implement using free ai models? Plz make a video for this. Don't want to download models on local system also.plz make a video like this also if feasible.
This is amazing content. A vast amount of knowledge is shared in just 3 hours. The knowledge I gained in this video will elevate my professional career.
Hi Krish, Thank you so much for sharing your knowledge through your videos. They have been incredibly helpful in my learning journey. I appreciate all the effort you put into them. I have a few questions that I believe will also benefit the community: Best Practices for Deployment: What are the best practices for deploying an open-source LLM (Large Language Model) to handle multiple requests from different users simultaneously? Handling Multiple Requests: How can I deploy an open-source LLM to efficiently handle many requests from different users at the same time? Reducing Response Time: What strategies can I use to reduce the response time of the deployed LLM? Accuracy with RAG Implementation: How can I check the accuracy of an LLM model when using Retrieval-Augmented Generation (RAG) implementation? Creating Instruction-Response Pairs: How can I convert raw text into instruction and response pairs for fine-tuning models like Mistral, LLaMA, or any other open-source model? I believe a video addressing these questions and providing solutions would be very helpful for the community. Thank you once again for your support and guidance. Best regards, Jay
1.Best Practices for Deployment: Deploying a Large Language Model (LLM) in a production environment to handle multiple requests simultaneously requires careful planning. Here are some best practices: Load Balancing: Use a load balancer to distribute incoming requests across multiple instances of your model. This can help ensure that no single instance becomes a bottleneck. Scaling: Use auto-scaling to automatically adjust the number of model instances based on the load. This can help ensure that you have enough capacity to handle peaks in demand, while not wasting resources during quieter periods. Monitoring and Logging: Implement comprehensive monitoring and logging to track the performance of your model and to quickly identify and resolve any issues. 2.Handling Multiple Requests: To handle multiple requests efficiently, consider using asynchronous processing or queuing systems. This allows your service to accept a request, place it in a queue, and then process it when resources are available, rather than forcing each request to wait for the previous one to complete. 3.Reducing Response Time: To reduce response time, consider the following strategies: Optimization: Optimize your model for inference. This could involve techniques like quantization, pruning, or using a more efficient implementation of the model. Caching: Implement caching for common requests. If you notice that certain requests are made frequently, you can store the results in a cache and return them immediately on subsequent requests. 4.Accuracy with RAG Implementation: To check the accuracy of an LLM model when using Retrieval-Augmented Generation (RAG), you can use a validation set of data. Run the model on the validation data and compare the model's outputs to the actual outputs. You can then calculate accuracy metrics based on this comparison. 5.Creating Instruction-Response Pairs: To convert raw text into instruction and response pairs for fine-tuning models, you can use a script that processes the raw text and splits it into pairs. The specifics of this will depend on the format of your raw text and the requirements of the model you're using. Typically, you would use a natural language processing (NLP) library like NLTK or SpaCy to help with this.
super nice and detail tutorial guide that found in youtube!!! finally i can usderstand where to get the API key and end point URL and a lot of anwsers for my questions for deploying genarative AI👍👍👍👍. Thanks you so much!!
Can we use azure for free ? just for the purpose of project development for a better resume. Please guide me if possible,I am a big fan of Krish's channel and his immnese contribution.thnks from thr core of my heart for teaching us ❤
does anybody knows about student subscriotion does it work on it ? if yes then whats the procedure i have tried with my document its giving me the wrong ans only
What are the differences in approaches between creating a custom chatbot using a shoes PDF file and later applying a RAG architecture with the same document? Specifically, how are the document chunks stored in Azure Blob? Are they first converted into embedded and then stored? If a vector database isn’t used, how is the information retrieved? Could you please clarify these questions.
Blob storage is to store your pdf file as a blob object in azure. Then this file will be read by azure service and it will divide the entire content present in the file as small chunks and then these chunks will be converted into embeddings and that will be stored in a database. That database is nothing but index which you are creating and to search the database you need a search capabilities right... That's called as Azure AI search service.
I see you are first cerating a model via GUI and then coding to call them via api? Do we have to create a model first via GUI? I am thinking how to create infra and deploy automatically from one environment to another via scripts with nothing manual from the GUI?
Good job . Howevr, its too lengthy and I felt lil bit the trainer was over acting and dragging... make It small nuggets so that it will be more interesting and to the point. It took just 5 mins for me to get the bigpicture from here ..ua-cam.com/video/3SRh2nzN2DM/v-deo.html . No offense its my view point
Lets keep the target of the likes to 2000. Also dont forget to subscribe his YT channel.
Hi sir, I want to create an AI power chat application which will query with MS SQL server database. Is this possible to implement using free ai models? Plz make a video for this. Don't want to download models on local system also.plz make a video like this also if feasible.
1 like added
This is a high quality content, I have been working on Azure AI from last few months, I can say you covered all key features very well
He should be awarded Bharat Ratna for his outstanding contribution to AI learning in India
PM bna dena chahiye
@@yogeshyts that works too
@@yogeshyts Bhakt chorenge nahi tumko aisi gustaki ke liye
This is amazing content. A vast amount of knowledge is shared in just 3 hours. The knowledge I gained in this video will elevate my professional career.
Hi Krish,
Thank you so much for sharing your knowledge through your videos. They have been incredibly helpful in my learning journey. I appreciate all the effort you put into them. I have a few questions that I believe will also benefit the community:
Best Practices for Deployment: What are the best practices for deploying an open-source LLM (Large Language Model) to handle multiple requests from different users simultaneously?
Handling Multiple Requests: How can I deploy an open-source LLM to efficiently handle many requests from different users at the same time?
Reducing Response Time: What strategies can I use to reduce the response time of the deployed LLM?
Accuracy with RAG Implementation: How can I check the accuracy of an LLM model when using Retrieval-Augmented Generation (RAG) implementation?
Creating Instruction-Response Pairs: How can I convert raw text into instruction and response pairs for fine-tuning models like Mistral, LLaMA, or any other open-source model?
I believe a video addressing these questions and providing solutions would be very helpful for the community.
Thank you once again for your support and guidance.
Best regards,
Jay
1.Best Practices for Deployment: Deploying a Large Language Model (LLM) in a production environment to handle multiple requests simultaneously requires careful planning. Here are some best practices:
Load Balancing: Use a load balancer to distribute incoming requests across multiple instances of your model. This can help ensure that no single instance becomes a bottleneck.
Scaling: Use auto-scaling to automatically adjust the number of model instances based on the load. This can help ensure that you have enough capacity to handle peaks in demand, while not wasting resources during quieter periods.
Monitoring and Logging: Implement comprehensive monitoring and logging to track the performance of your model and to quickly identify and resolve any issues.
2.Handling Multiple Requests: To handle multiple requests efficiently, consider using asynchronous processing or queuing systems. This allows your service to accept a request, place it in a queue, and then process it when resources are available, rather than forcing each request to wait for the previous one to complete.
3.Reducing Response Time: To reduce response time, consider the following strategies:
Optimization: Optimize your model for inference. This could involve techniques like quantization, pruning, or using a more efficient implementation of the model.
Caching: Implement caching for common requests. If you notice that certain requests are made frequently, you can store the results in a cache and return them immediately on subsequent requests.
4.Accuracy with RAG Implementation: To check the accuracy of an LLM model when using Retrieval-Augmented Generation (RAG), you can use a validation set of data. Run the model on the validation data and compare the model's outputs to the actual outputs. You can then calculate accuracy metrics based on this comparison.
5.Creating Instruction-Response Pairs: To convert raw text into instruction and response pairs for fine-tuning models, you can use a script that processes the raw text and splits it into pairs. The specifics of this will depend on the format of your raw text and the requirements of the model you're using. Typically, you would use a natural language processing (NLP) library like NLTK or SpaCy to help with this.
Worth the time... A good course covering overall most of the concepts...
super nice and detail tutorial guide that found in youtube!!! finally i can usderstand where to get the API key and end point URL and a lot of anwsers for my questions for deploying genarative AI👍👍👍👍. Thanks you so much!!
Thank you! As always everything is interesting and clear! From the bottom of my heart I wish you success and prosperity!
Thank you for this amazing resource
A perfect course which covered everything
Who's target is complete within 3 days, hit like and share thoughts. I am learning
I am listening in one go and reiterate again topic by topic
Very nice detailed session with great explanation.
Thank you sunny and krish
Great Work...Thank you so much for providing such a worthy content..
Hi Krish, I am a MERN full stack developer, from where what should i learn to become a fullstack AI developer to match with job market.
Can we use azure for free ? just for the purpose of project development for a better resume.
Please guide me if possible,I am a big fan of Krish's channel and his immnese contribution.thnks from thr core of my heart for teaching us ❤
❤ from Malaysia!!!! , never stop learning and always Day1
Thank you for detail session, covered all the points. Can you please also provide session on Copilot it will be really helpful.
This is what I looking for for a long time. Thank you guys
really ... me too
does anybody knows about student subscriotion does it work on it ? if yes then whats the procedure
i have tried with my document its giving me the wrong ans only
Maximum log kya Sikh rahe hai currently development or AI ML ? Job kisme better hoga in future aspects
What are the differences in approaches between creating a custom chatbot using a shoes PDF file and later applying a RAG architecture with the same document? Specifically, how are the document chunks stored in Azure Blob? Are they first converted into embedded and then stored? If a vector database isn’t used, how is the information retrieved? Could you please clarify these questions.
Blob storage is to store your pdf file as a blob object in azure. Then this file will be read by azure service and it will divide the entire content present in the file as small chunks and then these chunks will be converted into embeddings and that will be stored in a database. That database is nothing but index which you are creating and to search the database you need a search capabilities right... That's called as Azure AI search service.
Bht alaw...
it shows that we can register with company ebsite or something when filling form to request acess to azure Open Ai service,
Thanks Sunny and Krish =)
I see you are first cerating a model via GUI and then coding to call them via api? Do we have to create a model first via GUI? I am thinking how to create infra and deploy automatically from one environment to another via scripts with nothing manual from the GUI?
what should i do if i don't have a private email? I can't create MS Azure OpenAI :(
2:17:33 Azure AI studio is a framework like langchain?
needed this for a project
Can we give images as an input from the user to this chatbot???
it is asking for a company mail addresss for creating Azure open AI account , students cant have it , is there any other way sir?
How much is subscription cost for using Azure OpenAI in (INR) ?
Are you planning to upload Azure Gen AI created by you. You said you are going to work in parallel with AWS.
Do we need a paid account of Azure to follow this tutorial?
@53:57 when the dark lights takes over xD
It was scary
Please make video on Generative AI With Google Cloud Open AI Services Crash Course.Waiting
Azure Open AI has restricted their access, I am not getting any access with my regular Azure account.
Thanks krish , It helps us a lot :)
We need more courses where we are integrqting open source llm models in hf/colab.
Already on the track to upload the videos
Sir, can I apply for AI 900 certification after this course?
Thanks for the help
Sir asking you to create a video on Guardrails using Guardrails AI and Nemo Guardrails using Mistral or Llamma models
Brilliant
OMG, so cool, Thanks a lot :)
Thank you I requested.
Amazing
On point Knowledge ❤
Hey am here too!
Liking the session🤙
First comment ..🎉 great work sunny Savita
Shoes pdf not uploaded please
Not bad thus far.
Good work ❤
Fresher ko generative ai me job mil rahe hai?
Can you do one on Gcp ?
will be uploaded soon.
Please Sir Hindi ki playlist bhi bna do
2:34:17 and 2:36:13 - Where is it storing vectors or embeddings? Cosmos DB? There is no reference of Cosmos DB in RAG application built in this video?
we are using Azure Blob Storage it gives the capability of vector search
@@sunnysavita10 Thanks. Is it using Integrated Vectorization?
@@Marwaha_489 yup
🙏💯👍
1:06:54
2.30
All good however the presenter needs to stop fiddling with the mouse/cursor as well as the usage of the English language.
Good job . Howevr, its too lengthy and I felt lil bit the trainer was over acting and dragging... make It small nuggets so that it will be more interesting and to the point. It took just 5 mins for me to get the bigpicture from here ..ua-cam.com/video/3SRh2nzN2DM/v-deo.html . No offense its my view point
Why do you post a cringe face on a thumbnail; is it just to annoy people? I just clicked on the video to say this.