Qwen

/

Qwen3-32B

like 685

Follow

Qwen 80.4k

In the model card/README, why is it
--rope-scaling '{"rope_type":"yarn","factor":4.0,"original_max_position_embeddings":32768}' --max-model-len 131072
instead of
--rope-scaling '{"rope_type":"yarn","factor":4.0,"original_max_position_embeddings":40960}' --max-model-len 163840 ?