vermarjun/chad-pretrain download history
vermarjun/chad-pretrain is a dataset on the Hugging Face Hub. In the last 30 days it was downloaded 475 times (59 in the last 7 days), and 2,548 times in total. It ranks #34,487 among datasets by monthly downloads.
Chad pretrain corpus The pretraining corpus for Chad, a small from scratch Hinglish chat model. Around 4.13 billion tokens of romanized Hinglish and Indian English, split into 46 shards, one JSON object per line as {"text": ..., "source": ...}. Where it came from Reddi