cc

The score of CIDER decreses when the model is test on test server of MSCOCO

4

Thanks for your code first. I trained the model of top-down, the score on karpathy split is similar to the paper. However, when I want to test the model in...

The first stage training, lora is included.

1

Thanks for your excellent code! When I train the data in the first stage. I find that the lora layers are included in the first stage. I think it is...

InternVL3.5-1B sft的时候报错 RuntimeError: The size of tensor a (64) must match the size of tensor b (128) at non-singleton dimension 3

9

同样的数据在internvl3-1B训练ok，看了下配置也基本保持是一致的，但是在训练刚开始就报错 纬度不匹配，可以帮忙看下是哪里的问题嘛？ 报错如下： [rank18]: outputs: BaseModelOutputWithPast = self.model( [rank18]: return self._call_impl(*args, **kwargs) [rank18]: ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ [rank18]: File "/code/xxx/envs/internvl3/lib/python3.11/site-packages/torch/nn/modules/module.py", line 1844, in _call_impl [rank18]: return inner() [rank18]: ^^^^^^^ [rank18]: File "/code/xxx/envs/internvl3/lib/python3.11/site-packages/torch/nn/modules/module.py", line...