mongol-editor-llm-v2 / tokenizer.json

Commit History

V2 LoRA trained (loss 0.78, 3 epochs on augmented data)
d203ff8
verified

Tsedee commited on