xuwei fang

Hello, I modify the url like you said. Now I got this: 48 53 1 [98444:98444:20240104,073424.146823:ERROR process_memory_range.cc:86] read out of range [98444:98444:20240104,073424.147124:ERROR elf_image_reader.cc:606] missing nul-terminator [98444:98445:20240104,073424.155000:ERROR directory_reader_posix.cc:42] opendir attachments/45cdb988-f72a-49a3-a1dc-da8109221a65: No...

@konraddysput Hello, the url is : https://fangxuwei.sp.backtrace.io:6098/post

@konraddysput Still error: Crashpad URL: https://submit.backtrace.io/fangxuwei/62880b6b65009b657f7d3cfa7aa01ff4b8bd8801d282fab4880d7afca8b65210/json Crashpad handler: /fangxuwei/GitRepo/crashpad/crashpad/cbuild/handler/handler Crashpad database: . 46 51 1 [608:608:20240106,015823.224501:ERROR process_memory_range.cc:86] read out of range [608:608:20240106,015823.224749:ERROR elf_image_reader.cc:606] missing nul-terminator [608:609:20240106,015823.232690:ERROR directory_reader_posix.cc:42] opendir attachments/e41bd5be-45ad-419f-a5ef-cf0f0d37d531: No...

@konraddysput Like this?? https://submit.backtrace.io/fangxuwei/62880b6b65009b657f7d3cfa7aa01ff4b8bd8801d282fab4880d7afca8b65210/minidump

@KrzaQ I follow the linux compile tutorial which just need curl rather than openssl

@konraddysput the same error.

我这边测试Qwen2.5-VL-7B-Instruct ``` torchrun --nproc-per-node=2 run.py --data MMStar --model Qwen2.5-VL-7B-Instruct --verbose ``` 第一次是0.603 然后我对比https://github.com/QwenLM/Qwen2.5-VL/blob/main/evaluation/mmmu/run_mmmu.py中的代码，发现 use_custom_prompt 官方设置为True, 而 VLMEvalKit 中设置为False， 我修改后重新测试，结果为0.615。

@rzw520 `git checkout apex_no_distributed` `error: pathspec 'apex_no_distributed' did not match any file(s) known to git.`

I modify the code `from pytorch_lightning.utilities.distributed import rank_zero_only` to `from pytorch_lightning.utilities.rank_zero import rank_zero_only` fix the problem