Qwen3-32B resident model portions distributed by Limonata

Original publisher: Qwen.
Original artifact: Qwen/Qwen3-32B-GGUF, Qwen3-32B-Q4_K_M.gguf.
Revision: 938a7432affaec9157f883a87164e2646ae17555
Source: https://huggingface.co/Qwen/Qwen3-32B-GGUF/tree/938a7432affaec9157f883a87164e2646ae17555
Original SHA-256: efd971561896866f0e910cce52761ca77b1b138090c7f15fe284676d57d1f689
License: Apache License, Version 2.0; complete text accompanies this notice in
QWEN3-32B-LICENSE.txt and at https://limonata.xyz/ai/shared/license.

Modified by Limonata on 2026-10-17 for resident layer execution. The GGUF file
has been split into contiguous transformer ranges [0,48) and [48,64), local
layer numbering and stage metadata have been added, and required global tensors
have been duplicated. Original quantized tensor payloads are preserved; this is
not a trained or fine-tuned model. The splitting recipe and runtime modifications
are included in the provider source distribution under ai/swarm.

Portion 0 SHA-256: 88761b057a62f7f6b9889ccec94d3451982ea023893e9b1f7bade521ba48a56b
Portion 1 SHA-256: f31cedb9cb78b787fcb918bce24ec484007a049a6a548400d4d0f56f7a029472

This notice identifies the original work and Limonata's modifications. It does
not imply endorsement by Qwen or change the original model's license terms.
