young_chao issues

Results 10 issues of


                                            young_chao

直接使用fate镜像新建容器，启动后没办法使用fateboard

采用python ./examples/pipeline/demo/pipeline-mini-demo.py方式运行示例任务，fateflow的logs文件夹中有相应日志记录，而fateboard文件夹内无日志生成。在浏览器中无法打开127.0.0.1:8080无反应。

RuntimeError: 'weight' must be 2-D

### Describe the bug When I run the example of text_to_image.py, I got the problem shown in logs. I'm pretty sure I have it configured and running as the reademe.md...

bug

stale

[BUG]: LoRA does not support the training of reward models

### 🐛 Describe the bug ![image](https://user-images.githubusercontent.com/34190033/220584377-0467d526-d1c4-46a6-b4c8-c56a088bff96.png) ### Environment ### env OS：ubuntu 20.04 GPU：4 x A10 python==3.9.0 torch==1.13.1-cu116 colossalai==0.2.5 ### command `python train_reward_model.py --pretrain "bigscience/bloom-560m" --lora_rank 16`

bug

[BUG]: LoRA still does not support the training of reward models

### 🐛 Describe the bug Although [ht-zhou](https://github.com/ht-zhou) said that the LoRA problem has been fixed, according to the latest code and experimental tests of ColossalAI, LoRA still does not support...

bug

Question about the num-gpu-batches and gpu-batch-size

According to batch_size_table.md, from 144=48 x 3 (144 from batch_size_table.md and 48 x 3 from bench_suite.py) I can think that batch-size is composed of num-gpu-batches and gpu-batch-size together in FlexGen....

[Chatllama] Errors when training actor model based on LLaMA-7B

root@b787722dc2e1:/workspace/workfile/Projects/chatllama# python artifacts/main.py artifacts/config/config.yaml --type ACTOR Current device used :cuda local_rank: -1 world_size: -1 Traceback (most recent call last): File "/workspace/workfile/Projects/chatllama/artifacts/main.py", line 50, in actor_trainer = ActorTrainer(config.actor) File "/usr/local/lib/python3.9/site-packages/chatllama/rlhf/actor.py", line...

中文生成代码评测问题

在七月份官方公众号的博客上，能够看到对于中文生成代码评测问题的描述：目前的生成代码任务的基准评估集合都是英文的，为了客观地评估模型利用中文指令生成代码的能力，我们从stackoverflow和其他中文问答社区收集了与代码相关的问题。这些问题涵盖了Python、C++、Java、PHP、SQL等多种编程语言，包括代码纠错、代码生成、代码问题解答、代码解释等共计75道题目。我们使用这个评测集进行了Case-by-Case的人工评估。对于生成的代码问题，我们以运行结果是否正确作为通过标准；对于代码问答和解释的回答，我们以正确性和全面性为通过标准。但是以上描述的评测问题似乎并没有开源？如果有开源的话，希望能够提供一下链接。目前在huggingface和github并没有找到。

young_chao