Top Rated Weights & Biases Alternatives
I like that W&B provides an on-premise solution (in our cloud environment) that allows us to manage our data completely internally. Their python library is easy to use and can integrate quickly with our existing workflows for ML research. Specifically, we're able to automatically collect relevant ML data and display appropriate visualizations to help us find the best models. More generally, W&B lets us better keep track of our models and test various experiments easily.
They have a great reporting feature as well, where we can easily create and share reports relating to our ML experiments. The visualizations also flexible, and we can basically create whatever visuals we want (although with some effort)
I also like that they keep adding more features to help us accelerate and manage all of our ML operations easily. To my knowledge, we haven't made use of all of these features yet (at least Artifacts and Tables), but they will definitely help us with our workflows as we grow and mature our teams. Review collected by and hosted on G2.com.
Since we're running on-premise in our cloud, it takes a little effort to maintain the product in our environment. Review collected by and hosted on G2.com.
40 out of 41 Total Reviews for Weights & Biases
Overall Review Sentiment for Weights & Biases
Log in to view review sentiment.
I like the flexibility to log custom parameters, and the colorful comparisons between models in for instance confusion matrices. Review collected by and hosted on G2.com.
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 Review collected by and hosted on G2.com.
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. Review collected by and hosted on G2.com.
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. Review collected by and hosted on G2.com.
Everything works almost out of the box and it has a "nice look" once logged in Review collected by and hosted on G2.com.
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 Review collected by and hosted on G2.com.
The hosted aspect is great and if you avoid storing large artifacts, too many histograms, or too many images, it's very cheap or free even for heavy usage. I often run training on remote systems and checking the eval outputs remotely can be a pain with other software like TensorFlow. You need to be sure to run the server on your instance to view the dashboard. With W&B you get comparably powerful features, and you don't need to do anything but run your train script. In addition to hosting, the Sweeps functionality is excellent for hyper-parameter sweeping and pre-determined groups of runs in multi-task settings. Review collected by and hosted on G2.com.
Some UI elements, like the Runs data table, can be laggy, and sometimes the auto-refresh seems finicky. Review collected by and hosted on G2.com.
Ease of setting it up, the dashboard is pretty much each to use and allows to visualize a wide range of features at once. The integration is pretty much solid and works out of box for any script that I worked on Review collected by and hosted on G2.com.
I used to pre train certain checkpoints in a sequential manner or sometimes my runs used to break in between due to memory/connection issues from there on it was quite difficult to visualize all the previous run in a single curve using the dashboard, setting the x axis as wall time helped but the curve was still not a single continuous graph Review collected by and hosted on G2.com.
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! Review collected by and hosted on G2.com.
Can't think of anything that I dislike, my experience has been very positive until now. Review collected by and hosted on G2.com.
The ability to track everything including gradients is a wonderful aspect. Review collected by and hosted on G2.com.
One problem is if there's a zombie wandb process and I try to kill it, it makes other functioning trainig runs terminate. Review collected by and hosted on G2.com.
I like the WEB UI, especially the manipulation of plots and reports as they simplify and visualize many metrics and parameters.
I also like the artifactory and the model registry, they help manage the countless number of models created during an ML/DL project.
Sweep management is also cool! We build an automation tool around it that simplifies ML sweeps and thus helps us get better results.
Finally, I love the prompt and kind assistance we (Nvidia) get on the dedicated Slack channel. Really appreciated! Review collected by and hosted on G2.com.
Not too much actually :)
I guess sometimes the web UI is a bit slow. Review collected by and hosted on G2.com.
The online API is really helpful to organize results and projects. Review collected by and hosted on G2.com.
In the online website, the graphs showing evolution of a metric for example can't be dezoomed easily with the scroll wheel and we can't move in the graph after zooming. Review collected by and hosted on G2.com.
Its easy and seamless integration with PyTorch lightening and simple API usage. The model (artifact) and logs logging also help me trace back a model that was training months ago. Review collected by and hosted on G2.com.
Nothing as of now. Would appreciate more dark modes and API control to give experiment names, rather than having my experiment named 'Tasty-Aadvark'. Review collected by and hosted on G2.com.
