Paik, Young Sang (Keanu)

Suffering to make 'tars base Korean version'

2

I want to make a Korean base tars model that replaces 'tars-base-v8.pt' based on the Korean dataset and the Korean bert pretrained model on the huggingface. The train loss curve...

What is the major difference between ChemBERTa_zinc250k_v2_40k and v1 (or others)

2

Hi seyonec, I am using several version of models published ChemBERTa, and I have tried several versions of ChemBERTa tokenizer/pretrained model, and found result using "ChemBERTa_zinc250k_v2_40k" quite better than "seyonec/ChemBERTa-zinc-base-v1"...

### Describe your feature request Hi, starting with version 1.24, Weaviate provides gse and trigram tokenization methods, mentioned in your document. Gse implements the "Jieba" algorithm, a popular Chinese text...