PengqianHan

Results 2 issues of PengqianHan

In model.py line 149 As the v.shape is [batch,2,obs_len,num_ped], then (v.permute(0, 3, 1, 2)[0, :, :, 0]).shape is [num_ped,2], which means the location of num_ped pedestrians at the first time...

本教程参考官方 [readme](/Users/pengqianhan/Documents/GitHub/minimind-pq/README.md) 和 [不是Issue,一点个人训练minimind的记录 #26](https://github.com/jingyaogong/minimind/issues/26) 由于本人手头只有一台Macbook (M1 Pro),因此这个项目只是用来debug和学习代码,完全没有训练出一个可用的模型。在大佬的代码基础上减少了epoch,同时在一个epoch内只用很少的数据进行训练,代码可正常运行从而可以学习代码运行的逻辑。以下是我学习代码的流程: 0. train tokenizer - 下载 | **【tokenizer训练集】** | [HuggingFace](https://huggingface.co/datasets/jingyaogong/minimind_dataset/tree/main) / [百度网盘](https://pan.baidu.com/s/1yAw1LVTftuhQGAC1Y9RdYQ?pwd=6666) 文件为 tokenizer_train.jsonl - [博客或视频讲解](https://www.bilibili.com/video/BV1KZ421M7di/?spm_id_from=333.880.my_history.page.click&vd_source=e587bac74600ca53ef886eea337fe87d) - 运行 train_tokenizer.py, 运行结束后,在’model/minimind_tokenizer/‘下得到merges.txt,tokenizer_config.json,tokenizer.json,vocab.json 四个文件 1. data_process.py...