No opinion yet, but I'm planning to temporarily switch and see how it goes. I really hope the grass is greener there, but I'm keeping my expectations in check for now.
I recently switched to using for some large volume inference with small, fine-tuned models and was surprised how nice and painless things were. vLLM seems to have a lot more gotchas and kludged together stuff once you get outside of anything straightforward.
> What's you opnion vs. SGLang in this regard?
No opinion yet, but I'm planning to temporarily switch and see how it goes. I really hope the grass is greener there, but I'm keeping my expectations in check for now.
SGLang also accepts vibe coded PRs, but seems to be more careful about what goes in
I’ve had to debug DSV4 issues with both SGLang and vLLM in the last month. FWIW current SGLang nightly seems fine.
I recently switched to using for some large volume inference with small, fine-tuned models and was surprised how nice and painless things were. vLLM seems to have a lot more gotchas and kludged together stuff once you get outside of anything straightforward.