zyh3826

Does there have any example to remove duplicate docs using MinHash？

4

Does there have any example to remove duplicate docs using MinHash？

TSNE freezes on certain data[QST]

3

**What is your question?** like question #3360 [my dataset](https://drive.google.com/file/d/1P3UcuCrO6FU_NKv4E59goDzqQP36CxHD/view?usp=sharing) # Steps/Code to reproduce ```python import numpy as np from cuml import TSNE vectors = np.load(your_path) tsne = TSNE(n_components=2, init='random', random_state=42,...

无法复现结果

2

batch size: 32 GPU: RTX 2080Ti torch: 1.7.1 其他都一样，参数用的默认params.json里的，能分享一下你的超参嘛，谢谢

compile fitsne error

3

I use this command to compile from the main page: ```bash g++ -std=c++11 -O3 src/sptree.cpp src/tsne.cpp src/nbodyfft.cpp -o bin/fast_tsne -pthread -lfftw3 -lm ``` but got a warning, and the program...

Seems like does't support Chinese keys bettween named-databases?

2

### Affected Operating Systems * Linux ### Affected py-lmdb Version 1.3.0/1.0.0 ### py-lmdb Installation Method sudo pip install lmdb ### Machine "free -m" output ``` total used free shared buff/cache...

我用两块RTX 2080Ti还是会OOM，还有用你们的代码使用三块gpu必报错，多gpu混合精度也会报错，估计是pytorch-lightning的原因？没用过这个库

直接执行lcqmc_bert.sh结果如下  按照论文中先微调bert，再frozen bert训练田字格-CNN，最后两者一起训练，得出的结果还不如单纯bert高 调参后的bert结果  frozen bert训练田字格-CNN的结果：  两者一起训练结果：  除了bert经过了调参所有参数均是源码中的参数

使用RoFormerTokenizer出现错误

4

config = RoFormerTokenizer.from_pretrained(XXX) ModuleNotFoundError: No module named 'transformers.models.imagegpt' 使用transformers的没有错误

结果比较烂

6

博主，你的结果咋样啊，我用你的参数，验证集上f1一直在0.2左右，这也太低了吧

**Describe the bug** I use this command, `deepspeed --include localhost:"1" finetune_wav2vec2_for_ASR.py`, to finetune a Wav2Vec2 model using transformers trainer, it works well, but when change `--include` to `--include localhost:"1,3"`, everything...