第54期：《转让持有Neuralink、Space X、某头部自动驾驶公司股份的专项基金LP份额｜资情留言板第54期》
The Daily Dot reached out to learn more.
，这一点在新收录的资料中也有详细论述
Alternating the GPUs each layer is on didn’t fix it, but it did produce an interesting result! It took longer to OOM. The memory started increasing on gpu 0, then 1, then 2, …, until eventually it came back around and OOM. This means memory is accumulating as the forward pass goes on. With each layer more memory is allocated and not freed. This could happen if we’re saving activations or gradients. Let’s try wrapping with torch.no_grad and make required_grad=False even for the LoRA.
对于企业用户来说，OpenClaw也不是部署之后就能及时启用的。据惊蛰研究所了解，包括OpenClaw在内，目前很多AI Agent还无法稳定完成复杂任务。因此，企业在部署完OpenClaw之后，往往仍然需要持续调试、维护，才能够满足“投产”的需求。所以部署OpenClaw只是第一步，真正困难的是运营——这也是很多普通用户很快放弃“养虾”的原因。