Upload GGUF model exports

.gitattributes

@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+gguf/granite-math-physics-f16.gguf filter=lfs diff=lfs merge=lfs -text

gguf/README.md

@@ -0,0 +1,13 @@
+# Galena-2B GGUF Artifact
+
+The GGUF export(s) live outside of Git and must be downloaded before llama.cpp
+examples can run:
+
+```bash
+python scripts/download_artifacts.py --artifact gguf
+```
+
+The helper script fetches `gguf/*.gguf` from the Galena-2B snapshot on Hugging Face
+(or from a release/CDN archive when `--source mirror` is used). After download,
+`models/math-physics/gguf/granite-math-physics-f16.gguf` becomes available for
+`llama-cli`, LM Studio, Ollama, etc.

gguf/granite-math-physics-f16.gguf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ea444fe88b45dca87bc6c67f805badb14f5cadd52049eb93444a8850ccd5a0a2
+size 5069157664