hibana2077
Remove old train.parquet file and update new file with larger size; enhance convert.py to drop duplicates and reset index; add PTT article crawler script
39c419c Download data/train.parquet from DoggiAI/Taiwan-Netizen: direct link, hf CLI and curl.
- Browser
- Download file 636 kB
-
https://huggingface.co/datasets/DoggiAI/Taiwan-Netizen/resolve/main/data/train.parquet
- Command line
-
hf download hf://datasets/DoggiAI/Taiwan-Netizen/data/train.parquet
-
curl -L -o train.parquet https://huggingface.co/datasets/DoggiAI/Taiwan-Netizen/resolve/main/data/train.parquet
636 kB
- Xet hash:
- 706c373ff471f2654c4378edd078125e12053fbcfdb4cb5ce29cfec82380d726
- Size of remote file:
- 636 kB
- SHA256:
- 61b17507086708ca1fce53157fdd6fa9bc36049750b1e776adf6cdc134bab144
·
Xet efficiently stores Large Files inside Git, intelligently splitting files into unique chunks and accelerating uploads and downloads. More info.