Update README.md
Browse files
README.md
CHANGED
|
@@ -1,3 +1,17 @@
|
|
| 1 |
-
---
|
| 2 |
-
license: creativeml-openrail-m
|
| 3 |
-
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: creativeml-openrail-m
|
| 3 |
+
base_model: nucleus-image
|
| 4 |
+
tags:
|
| 5 |
+
- text-to-image
|
| 6 |
+
- fp8
|
| 7 |
+
---
|
| 8 |
+
|
| 9 |
+
# Nucleus-Image-FP8-Native (Heavyweight Edition)
|
| 10 |
+
|
| 11 |
+
This is a native FP8 (float8_e4m3fn) quantization of the 17B Nucleus-Image model.
|
| 12 |
+
|
| 13 |
+
## ⚠️ VRAM Warning
|
| 14 |
+
This model is **EXTREMELY HEAVY**. Even in FP8, the weights alone take up ~13-14GB of VRAM.
|
| 15 |
+
- **16GB VRAM (RTX 4060 Ti / 4070 Ti Super / 4080):** Recommended to use with `sequential_cpu_offload` for stability. Pure GPU inference might OOM at 1024x1024.
|
| 16 |
+
- **24GB VRAM (RTX 3090 / 4090):** Best experience. Can run Pure GPU without offloading.
|
| 17 |
+
- **12GB VRAM or less:** **NOT RECOMMENDED** unless using heavy CPU offloading (will be slow).
|