"Neptune works flawlessly, and integrating it with PyTorch Lightning was very smooth."
"Clearly, handling the training of more than 7000 separate machine learning models without any specialized tool is practically impossible. We definitely needed a framework able to group and manage the experiments."
“The last few hours have been my first w/ Neptune and I’m really appreciative of how much time it’s saved me not having to fiddle w/ matplotlib in addition to everything else.“
"Weights and Biases went from being reasonably priced to being way too much. Especially since more than half the people we wanted to be able to see our models weren’t doing modeling. When we looked for an alternative, Neptune was the only one that could offer us everything we needed."
“I have been pleasantly surprised with how easy it was to set up Neptune in my PyTorch Lightning projects."
"We initially aimed for a GKE deployment for our experiment tracking tool. However, the other solution we explored had a rigid installation process and limited support, making it unsuitable for our needs. Thankfully, Neptune’s on-premise installation offered the flexibility and adjustability we required. The process was well-prepared, and their engineers were incredibly helpful, answering all our questions and even guiding us through a simpler deployment approach. Neptune’s on-prem solution and supportive team saved the day, making it a win for us."
"We use PyTorch Lightning, and it was just a matter of changing the tracker from Weights and Biases to Neptune. It’s like two lines of code. It’s actually quite easy."
"Self-hosted deployment for ML solutions will become more and more important. People don't feel comfortable with valuable intellectual property being stored in 3rd party DBs. For us, such deployment was too difficult and time-consuming in the previous solution. We could achieve that with Neptune, and it allowed us to close important deals that had stringent security requirements."
"At some point, one of my students tried doing the tracking process manually, and he was very frustrated after one project. Any manual change can mess up information organization and how you track it. And if you do not build it well, then you suffer, you need to recode, etc. I think it’s just a waste of time."
"We have a mantra: always be learning. We apply this primarily to our model, which means we’re always running experiments. So me, our CEO, other people in the team—we’re constantly checking the monitoring tool. It has to be nice, smooth, and be able to handle our training data streams consistently."
“I am super messy with my experiments, but now I have everything organized for me automatically. I love it."
“Neptune allows us to keep all of our experiments organized in a single space. Being able to see my team’s work results any time I need makes it effortless to track progress and enables easier coordination.”
"When working with projects that had thousands of runs, loading the interface and sorting through data was super slow in Weights & Biases. Neptune is better at handling large-scale. We’re happy with this choice."
"We have all the metrics in our shared file storage as a backup, but we don’t really have a nice way to access them, to sort them, etc. We don’t have a setup for it because Neptune has been stable enough for us not to need it."
"No more DevOps needed for logging. No more starting VMs just to look at some old logs. No more moving data around to compare TensorBoards."