cleanup: drop mlx files, align with lemer gguf-playground pattern

Mirrors the 160384c lemer cleanup for the rest of the family. HF was
rendering this repo with mlx detection because mlx-formatted safetensors
lived alongside the gguf files — hurting gguf/Ollama discoverability
for users searching for llama.cpp-compatible weights.

Changes:
- model-*.safetensors / model.safetensors.index.json removed (these
were mlx affine-quant format, not transformers — duplicated content
now in lthn/lemma-mlx-*)
- README title suffixed '(GGUF)'
- README body: removed mlx inline quick-starts, added redirect block
pointing at lthn/lemma-mlx / -mlx-8bit / -mlx-bf16
- frontmatter tags dropped mlx / apple-silicon, added llama.cpp / ollama
- library_name preserved gguf (lemmy changed from mlx → gguf)
- stale model-index block (pre-reset smoke numbers) removed — the
canon in .eval_results/ is authoritative

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

README.md

@@ -7,54 +7,31 @@ base_model:
 base_model_relation: quantized
 tags:
   - gemma4
-  - gemma
+  - lemma
   - gguf
-  - safetensors
-  - mlx
+  - llama.cpp
+  - ollama
   - multimodal
   - vision
   - audio
-  - lemma
-  - lethean
-  - lem
-  - apple-silicon
   - on-device
   - conversational
-model-index:
-  - name: lemma
-    results:
-      - task:
-          type: text-generation
-          name: mmlu_pro (8-PAC)
-        dataset:
-          type: TIGER-Lab/MMLU-Pro
-          name: MMLU-Pro
-          split: test
-          config: mmlu_pro
-        metrics:
-          - type: acc
-            value: 31.25
-            name: Per-round accuracy (%)
-          - type: acc
-            value: 25.0
-            name: Majority-vote accuracy (8-PAC)
-          - type: confidence
-            value: 0.6875
-            name: Mean per-question confidence (max-share)
-        source:
-          name: Raw per-iteration results (parquet + latest.md)
-          url: https://huggingface.co/lthn/lemma/tree/main/results
 ---
 <!--
 This content is subject to the European Union Public Licence (EUPL-1.2).
 For full licence details, please refer to: https://huggingface.co/lthn/lemma/tree/main/LICENSE
 Origin URL: https://huggingface.co/lthn/lemma/tree/main
 -->
-# Lemma — Gemma 4 E4B
+# Lemma — Gemma 4 E4B (GGUF)
+
+The mid-sized member of the [Lemma model family](https://huggingface.co/collections/lthn/lemma) by [Lethean](https://lthn.ai). An EUPL-1.2 fork of [Gemma 4 E4B](https://huggingface.co/google/gemma-4-E4B-it) with the **Lethean Ethical Kernel (LEK) merged into the weights** — consent-based reasoning baked into the attention projections via LoRA finetune, then merged so inference uses a single standalone model with no PEFT runtime required.
 
-The mid-sized member of the [Lemma model family](https://huggingface.co/collections/lthn/lemma) by [Lethean](https://lthn.ai). An EUPL-1.2 fork of [Gemma 4 E4B](https://huggingface.co/google/gemma-4-E4B-it) with the **Lethean Ethical Kernel (LEK) merged into the weights** — consent-based reasoning baked into the attention projections via LoRA finetune, then merged so inference uses a single standalone model with no PEFT runtime required. GGUF and MLX builds with full multimodal support — text, image, and audio — distributed from a single repo. Use GGUF with Ollama, llama.cpp, GPT4All, or LM Studio. Use MLX safetensors with `mlx-lm` and `mlx-vlm` for native Apple Silicon inference. The unmodified Gemma 4 E4B fork lives at [LetheanNetwork/lemma](https://huggingface.co/LetheanNetwork/lemma) for users who want the raw Google weights without the LEK shift.
+This repo ships the **GGUF multi-quant build** — five quants from Q4_K_M up to BF16, with full multimodal support (text, image, audio). Use with Ollama, llama.cpp, GPT4All, or LM Studio. The unmodified Gemma 4 E4B fork lives at [LetheanNetwork/lemma](https://huggingface.co/LetheanNetwork/lemma) for users who want the raw Google weights without the LEK shift.
 
-**MLX-only quant variants:** [Q8](https://huggingface.co/lthn/lemma-mlx-q8) | [BF16](https://huggingface.co/lthn/lemma-mlx-bf16) — Q4 multimodal is bundled in this repo.
+**Looking for MLX?** The native Apple Silicon builds live in sibling repos:
+[`lthn/lemma-mlx`](https://huggingface.co/lthn/lemma-mlx) (4-bit default) |
+[`lthn/lemma-mlx-8bit`](https://huggingface.co/lthn/lemma-mlx-8bit) |
+[`lthn/lemma-mlx-bf16`](https://huggingface.co/lthn/lemma-mlx-bf16) (full precision)
 
 > A **lemma** is "something assumed" — an intermediate theorem on the path to a larger proof, or a heading that signals the subject of what follows. The Lemma model family is named for that role: each variant is a stepping stone between raw capability and ethical application.
 

model-00001-of-00002.safetensors

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:574cb733a9a381602fbe121fc36281b167c504cc1ea397f787d34e3a90061ef7
-size 4280494837

model-00002-of-00002.safetensors

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:b55534903049d53498f1db96c02bfc162330fa0a7a078ba26bdea9894822e85d
-size 2588373297