Even though @eugr has been the main focus lately, you guys—as the extended team—have simply done a fantastic job. You’ve also helped so many of us out of a bind along the way (like when we had to completely reset Spark at the beginning because we were a bit too hasty with outdated playbooks). I feel like the forum is getting harder and harder to navigate because of the subject matter. Of course, this is systemic—different LLMs, quants, inferences, and the latest advancements (MTP, etc.). Maybe you could also use your chemistry to brainstorm a bit more here about the structure (flares?) and so on … Nevertheless, thanks for the great work you community experts have done for Spark, which many of us quietly benefit from because it’s saved us hours of time, a lot of gray hairs, and kept our better halves in good spirits ;)
Congrats, @eugr_nv !!! All the best in your future endeavors!!!
Congratulations!
Congrats! Smart hire by NVIDIA! It does bring some confidence in sticking with these devices to see them grab some community experts and bring them in full time!
For some reason I just saw this in my email. That is super good news, Gene. Congratulations and thanks for all your hard work on the Spark so far. It’s helped me and many others a lot!