Yue Ding

Results 2 comments of Yue Ding

Hi, did solve the problem? I met the same problem with llava vision is `1.1.3` with LLaVA vision is `v1.2.0` transformers `4.36.2` torch `2.0.1`

same question, with the original code in the class `MultiHeadAttention` in mha.py. Cause the following logic, the softmax will operate cross batch, which I don't understand. Need help. ``` #...