I use Prem Studio to automate fine-tuning and evaluation of AI solutions.
Prem Studio
PremReviews from AWS customer
0 AWS reviews
-
5 star0
-
4 star0
-
3 star0
-
2 star0
-
1 star0
External reviews
1 review
from
External reviews are not included in the AWS star rating for the product.
Has accelerated AI solution development through automated evaluation and fine-tuning
What is our primary use case?
How has it helped my organization?
Prem Studio allowed me to easily automate and solve the complicated problem of exploring and identifying the best AI architecture for our problems. Prem Studio has a straightforward yet powerful approach to evaluate disparate AI architectures on our business problems and to accurately fine-tune the most promising ones.
This significantly reduced time-to-market, almost by a factor of ten in my case, for solutions tailored and optimized for customer requirements and KPIs.
What is most valuable?
The most valuable features of Prem Studio are evaluation, fine-tuning, and experiment design.
What needs improvement?
Prem Studio is perfect for text input and output. We requested support for vLM, which should be available soon.
For how long have I used the solution?
I have been using the solution for 3 months.
What was my experience with deployment of the solution?
I did not experience any deployment issues.
What do I think about the stability of the solution?
I did not encounter any stability issues.
What do I think about the scalability of the solution?
I did not experience any scalability issues.
How are customer service and support?
I did not interact with customer service.
Which solution did I use previously and why did I switch?
Before Prem Studio, I used FireworksAI and adaptive-ml.com, but they were too brittle and rigid. It was difficult to update experiments given the new models and architectures that constantly emerge. Finetunedb is very low-level if you have to make it work for real.
How was the initial setup?
The initial setup was straightforward and did not present any challenges.
What about the implementation team?
The implementation was handled internally.
What was our ROI?
I have observed significant value in terms of reduced time-to-market, although I have not formally calculated the ROI.
What's my experience with pricing, setup cost, and licensing?
To maximize the effectiveness of fine-tuning, I would suggest using the parallelism features of the platform. This allows fine-tuning and evaluating multiple models at the same time, making comparisons more informative and helping to choose the best configuration among a representative set of models.
Which other solutions did I evaluate?
I considered and used FireworksAI, adaptive-ml.com, finetunedb.com, Claude, ChatGPT, Optuna, and Neptune.ai.
What other advice do I have?
If you want specific features, I recommend contacting the Prem Studio team since they are happy to listen to your needs.
showing 1 - 1