feat: add Q4_K_M + Q8_0 + BF16 gguf via LFS (xet-accelerated)

Converted from model.safetensors via llama.cpp convert_hf_to_gguf.py
(bf16 intermediate) and llama-quantize (Q4_K_M + Q8_0). Same gguf set
as the lemer/lemmy/lemrd family members.

.gitattributes now has *.gguf under LFS filter so the push goes
through HF's Xet content-defined dedup instead of a raw blob upload
at <1MB/s.

Ollama pull paths:
ollama pull hf.co/LetheanNetwork/lemma:Q4_K_M
ollama pull hf.co/LetheanNetwork/lemma:Q8_0
ollama pull hf.co/LetheanNetwork/lemma:BF16

Co-Authored-By: Virgil <virgil@lethean.io>

.gitattributes

@@ -34,3 +34,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
 tokenizer.json filter=lfs diff=lfs merge=lfs -text
+*.gguf filter=lfs diff=lfs merge=lfs -text

lemma-bf16.gguf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:966017977865dc4437769fb5275ea5c0a40ed9c25a3984a254c2507d3273d174
+size 15053090880

lemma-q4_k_m.gguf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ec9c82cf82b8fb23fab5191338dc8126c068fe8e29cc7aa3e97b82bf98efaef6
+size 5335285824

lemma-q8_0.gguf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7f22360a35824f991a4d3df519f94275aa9fc519e1c2611018ee118fd418933b
+size 8031236160