Junhua LU
Junhua LU
when running the example, and using the draw-polygon. after clicking the "first point to close shape", the dashed line is still there. the console also says "Uncaught TypeError: Cannot read...
首先非常感谢贵司提供的数据集和模型，这两天摩拳擦掌跃跃欲试 但是发现量化版本这种模型占用资源小的，都没法像原版那样直接用LlamaForCausalLM加载 我自己找的云服务器配置各种设置半天没弄出来... 请问贵司会提供LlamaForCausalLM这样方便调用的接口吗
MSEloss by default uses 'mean' as reduction method, so I think `epoch_loss += (loss.detach().item() / batchsize)` is incorrect
 the number in the red circle should be 31 since the index starts from 0