kevin

tensor.numpy()执行大量数据从GPU拷贝到CPU速度缓慢

2

### 需求描述 Feature Description tensor.numpy()执行大量数据从GPU拷贝到CPU速度缓慢，5M数据执行tensor.numpy()耗费了1.4s，完全不可接受！啥原因呢？ ### 替代实现 Alternatives _No response_

[BUG]: ChatGPT: why total reward is reward = r - kl_coef * kl, not total_reward=r + gamma * critic(next_states)?

1

### 🐛 Describe the bug ChatGPT: why total reward is reward = r - kl_coef * kl, not total_reward=r + gamma * critic(next_states)?  ### Environment _No response_

where is the train file?

1

where is the train file? I want to learn how to train.

TencentPretrain pretrain.py of llama-7b model, Exception: : Current loss scale already at minimum

2

when I run pretrain.py of llama-7b model, it has exception below （not user zero3）: ExceptionException: : Current loss scale already at minimum - cannot decrease scale anymore. Exiting run.Current loss...

### Is your feature request related to a problem? Please describe. 为啥我的4090显卡跑chatglm-6b的GPU只有20%？ 如何提升？ ### Solutions who can help me？ ### Additional context _No response_

when to support chatglm2-6b?

2

chatglm-6b(chatglm2-6b) is a very popular Chinese LLM. Do you have a plan？

pretrain.py of llama-7b model, Exception: : Current loss scale already at minimum

2

when I run pretrain.py of llama-7b model, it has exception below: ExceptionException: : Current loss scale already at minimum - cannot decrease scale anymore. Exiting run.Current loss scale already at...