Add provenance and archive card
Browse files
README.md
ADDED
|
@@ -0,0 +1,21 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
library_name: mlx
|
| 3 |
+
tags:
|
| 4 |
+
- mlx
|
| 5 |
+
- quantized
|
| 6 |
+
- oq
|
| 7 |
+
- lfm2
|
| 8 |
+
---
|
| 9 |
+
|
| 10 |
+
# LFM2 1.2B MLX oQ Quantization Sweep
|
| 11 |
+
|
| 12 |
+
This repository is a public preservation archive for locally generated MLX quantizations of [mlx-community/LFM2-1.2B-bf16](https://huggingface.co/mlx-community/LFM2-1.2B-bf16).
|
| 13 |
+
|
| 14 |
+
## Provenance
|
| 15 |
+
|
| 16 |
+
- Source conversion: `mlx-community/LFM2-1.2B-bf16`
|
| 17 |
+
- Original model family: `LiquidAI/LFM2-1.2B`
|
| 18 |
+
- Quantization: local oMLX/oQ affine MLX safetensors, group size 64
|
| 19 |
+
|
| 20 |
+
Each subdirectory is an independently loadable MLX artifact. The directory name is the intended quantization label. Existing standalone `TiGa-RCE/LFM2-1.2B-MLX-oQ4` and `TiGa-RCE/LFM2-1.2B-MLX-oQ8` repositories remain as directly consumable releases; this is the complete archival sweep.
|
| 21 |
+
|