zazd

@naibaf7 I am very interesting in the LibDNN. It gets a good capability. For I am not familiar with opencl , I just glance over the LibDNN, it seems that...

@yangxudong 提供一个参考案例给你，希望有帮助, 因为只是简单体验uniem，可能有理解不对的地方。假设你已经跑通单机finetune的情况下， 1. 通过accelerate启动多卡: a.生成一个config.yaml文件，复制以下配置: (其中num_processes:4 --表示4卡跑任务，其他配置自行查找accelerate配置) compute_environment: LOCAL_MACHINE distributed_type: MULTI_GPU downcast_bf16: 'no' gpu_ids: all machine_rank: 0 main_training_function: main mixed_precision: fp16 num_machines: 1 num_processes: 4 rdzv_backend: static same_network: true...