DeepSeek-R1-Distill-Qwen-32B config.json

Fields

id

sources/mbl-cfg-r1-distill-qwen-32b

URL

https://huggingface.co/deepseek-ai/DeepSeek-R1-Distill-Qwen-32B/raw/main/config.json

Publisher

Hugging Face

Kind

repo

Accessed

2026-10-04

notes

num_hidden_layers 64, hidden_size 5120, rope_theta 1000000, max_position_embeddings 131072, rope_scaling null. Native 32768 with the model card advertising 131072; no rope_scaling in config, so the extension is not a config-level interpolation. Read raw 2026-10-04.

Notes

num_hidden_layers 64, hidden_size 5120, rope_theta 1000000, max_position_embeddings 131072, rope_scaling null. Native 32768 with the model card advertising 131072; no rope_scaling in config, so the extension is not a config-level interpolation. Read raw 2026-10-04.

Cross-references

This record references no other records.