Reviews from AWS Marketplace
0 AWS reviews
-
5 star0
-
4 star0
-
3 star0
-
2 star0
-
1 star0
External reviews
External reviews are not included in the AWS star rating for the product.
Very useful quite powerful tool
What do you like best about the product?
Easy of use, ease of implementation, the possibility to gather all my results, ease of sharing results with teammates, It can compare a lot of data interactively which in other cases could be hard to implement
What do you dislike about the product?
It is online approach which is both strong and weak side, sometimes servers are bit laggy.
What problems is the product solving and how is that benefiting you?
It makes easy for me to store and analyze experiments results, which in case of using own implementation approach using matplotlib for example would require quite a lot of work.
- Leave a Comment |
- Mark review as helpful
Seamless Integration and Reliable Support: A Daily Essential for Machine Learning
What do you like best about the product?
It is highly and well integrated with libraries I am using like PyTorch Lightning. Many times I have set that up because it was very easy, and after a while, it has happened that I lost part of my results, and W&B helped me to easily recover them through the logs, which without it, I probably wouldn't have. The next and very important feature to me is that I can use many different machines and servers at the same time and without being worried about gathering all results together, then using tools like TensorBoard, having them online without any effort(most of the time saves me when I am using supercomputers). It is part of my daily tools, and when I am teaching students machine learning, in very early sessions after teaching them visualization, I will have them use W&B to repeat whatever they have learned so far. I have never had any problem with W&B, but I have heard from one of my friends, whom I recommended he use W&B, that customer support is very fast and experienced.
What do you dislike about the product?
Sometimes it is bothering me when I am looking for very basic functionality of W&B and it doesn't provided good documentation for that.
What problems is the product solving and how is that benefiting you?
Monitoring training models is an integral part of my research ad W&B made it easy for me.
Review for wandb
What do you like best about the product?
I like the flexibility to log custom parameters, and the colorful comparisons between models in for instance confusion matrices.
What do you dislike about the product?
There are a couple of features I would have liked, such as the ability to set a global normalization flag, or controlling which windows stay on screen on reload, maybe these features exist, in that case I just haven't been able to find them
What problems is the product solving and how is that benefiting you?
We build image classification software, and for that it is very useful to track model performance
The fastest way to log your training runs
What do you like best about the product?
Everything works almost out of the box and it has a "nice look" once logged in
What do you dislike about the product?
I really can't resize images and masked images the way I want inside the GUI and looking at images on mobile is quite a nightmare
What problems is the product solving and how is that benefiting you?
Logging my training runs. I could achieve the same result with tensorboard (and alike) + saving images and masks and periodically generating a webpage that shows them, but W&B already offers that and much more
Great platform for DNN training support
What do you like best about the product?
Visualization of data from different perspective
What do you dislike about the product?
Nothing really, maybe it can be good to customize the hardware monitoring interval
What problems is the product solving and how is that benefiting you?
It helps me learn the progress of my machine learning training jobs
WandB - the best online tool for experiment logging so far
What do you like best about the product?
We are a small team of 15 researchers. After the years of usage tensorboard, we decided to try online tools. We tried a few of them and find out that WandB suits us best. We really like easy of use, the fact that experiments are easily sharable, and hyperparameter sweep option. Also the option to tag all your experiments and subsequent filtration of them is also great.
What do you dislike about the product?
WandB provides an option to create a report from your experiments. It would be nice to use these reports in our papers, however, during the review period, papers should be anonymized but there is no option to anonymize the reports.
What problems is the product solving and how is that benefiting you?
Experiments sharing and logging
Weights and biases changed the way I interact with my models.
What do you like best about the product?
The best thing about W&B is that you don't need to think about performance visualization anymore. W&B handles that for you, no matter how many metrics you have or how complex they are. It's also very simple to use!
What do you dislike about the product?
Can't think of anything that I dislike, my experience has been very positive until now.
What problems is the product solving and how is that benefiting you?
Helps a lot with visualization and performance tracking.
WandMe
What do you like best about the product?
The ability to track everything including gradients is a wonderful aspect.
What do you dislike about the product?
One problem is if there's a zombie wandb process and I try to kill it, it makes other functioning trainig runs terminate.
What problems is the product solving and how is that benefiting you?
To keep track of all the experiments and look at the comparisons anytime.
Also share the plots and reports with the rest of the world.
Also share the plots and reports with the rest of the world.
Great platform for collaboration of experiments with intuitive coding steps
What do you like best about the product?
The API is self explanatory and the UI is smart
What do you dislike about the product?
Caching management is not clear, dis not find a way to clean old logs
What problems is the product solving and how is that benefiting you?
Sharing results, experiment tracking management
Best software for ML experiment tracking
What do you like best about the product?
- Ability to view in browser
- Ability to share with collaborators
- Ability to add figures to experiment reports
- Ability to share with collaborators
- Ability to add figures to experiment reports
What do you dislike about the product?
- Annoying bug with visualising segmentation labels - red speckles appear in the label masks which can make it hard to evaluate.
What problems is the product solving and how is that benefiting you?
Tracking ML experiments and sharing with collaborators
showing 1 - 10