Vertex AI LLM Model Honest Reviews
Overall Satisfaction with Vertex AI
I use Vertex AI to solve the problems of finding insight from multiple pdf document using a generative AI chatbot, GCP Document AI used for OCR which is costlier we later realized then we used their Gemini model as a large language model for answer generation but it also failed in generating answer at production, overall i am not satisfied with the Vertex AI services. It is ok for POC but not for production.
Pros
- Document AI is good at extracting content from pdf
- Model Garden is a good option for exploring different models at one place.
- Kubeflow pipeline for automating training and serving model
Cons
- their Large language model accuracy
- vector index creation time is at least 40 mins which is not good
- Model garden should include OpenAI integration
- it is good for faster development by leveraging their API key and SDKs
- Costing is high, it is not worth the amount of accuracy they are providing
- Computer Vision related task are easy to implement using this
I have used OpenAI for their LLM and Vector Embedding service, they are really good at it. But Vertex AI has other better services like training pipeline , depolyment creation etc.
Do you think Vertex AI delivers good value for the price?
No
Are you happy with Vertex AI's feature set?
No
Did Vertex AI live up to sales and marketing promises?
No
Did implementation of Vertex AI go as expected?
No
Would you buy Vertex AI again?
No

Comments
Please log in to join the conversation