Winnow-E4B

Base model: Gemma 4 E4B IT by Google DeepMind
Source: https://huggingface.co/google/gemma-4-E4B-it
Source revision: ee0ef6023621cff504d758262d4e04895a5af4a2
Upstream license: Apache License, Version 2.0
License reference: https://ai.google.dev/gemma/docs/gemma_4_license

Winnow modifications are developed by EldanRing. The released Q8_0 GGUF
modifies the base through task-specific LoRA fine-tuning, adapter merging and
quantization. Its checksum and provenance are recorded in SHA256SUMS and
release-manifest.json. The private training dataset and native checkpoint are
not included in this model repository.

The separate Winnow inference repository builds on llama.cpp by Georgi
Gerganov and contributors under the MIT license. No llama.cpp source or binary
is included in this model repository. Winnow is an independent project; no
upstream endorsement is implied.