Upload README.md with huggingface_hub
Browse files
README.md
CHANGED
|
@@ -97,6 +97,8 @@ measured, a 34.7B MoE at 16.3 tok/s with 2 GB free.
|
|
| 97 |
| [MiniMax-H3 Turbo](https://huggingface.co/infosave/MiniMax-H3-Turbo-cmf) | text → video **with sound**, 4 steps | 71.05 GB |
|
| 98 |
| [LTX-2.5](https://huggingface.co/infosave/LTX-2.5-cmf) | text → video **with sound**, 21B DiT + audio VAE | 20.53 GB |
|
| 99 |
| [Qwen3.8-27B](https://huggingface.co/infosave/Qwen3.8-27B-cmf) | dense, q4tp | 14.44 GB |
|
|
|
|
|
|
|
| 100 |
|
| 101 |
Every one of them runs from the same binary: `cortiq run` for the text
|
| 102 |
models, `cortiq imagine` for Lumina, `cortiq animate` for MiniMax-H3, `cortiq ltx-video` for LTX-2.5.
|
|
|
|
| 97 |
| [MiniMax-H3 Turbo](https://huggingface.co/infosave/MiniMax-H3-Turbo-cmf) | text → video **with sound**, 4 steps | 71.05 GB |
|
| 98 |
| [LTX-2.5](https://huggingface.co/infosave/LTX-2.5-cmf) | text → video **with sound**, 21B DiT + audio VAE | 20.53 GB |
|
| 99 |
| [Qwen3.8-27B](https://huggingface.co/infosave/Qwen3.8-27B-cmf) | dense, q4tp | 14.44 GB |
|
| 100 |
+
| [Qwen3.8-Flash-Next](https://huggingface.co/infosave/Qwen3.8-Flash-Next-cmf) | 176.9B hybrid MoE/PLE, mixed q4tp + q8_2f | 97.12 GB |
|
| 101 |
+
| [Granite 4.2 — 3B/8B/30B](https://huggingface.co/infosave/Granite-4.2-cmf) | dense reasoning family, q4tp + q8_2f | 2.17–29.31 GB |
|
| 102 |
|
| 103 |
Every one of them runs from the same binary: `cortiq run` for the text
|
| 104 |
models, `cortiq imagine` for Lumina, `cortiq animate` for MiniMax-H3, `cortiq ltx-video` for LTX-2.5.
|