Update README.md
Browse files
README.md
CHANGED
|
@@ -30,8 +30,15 @@ RND1 is an experimental diffusion language model with 30B parameters and 3B acti
|
|
| 30 |
- Architecture: Sparse Mixture-of-Experts
|
| 31 |
- Training: Converted from pretrained autoregressive base (Qwen3-30BA3B)
|
| 32 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 33 |
**Note:** RND1-Base-0910 has not been post-trained. Expect occasional repetition with greedy samplers.
|
| 34 |
|
|
|
|
| 35 |
## Installation
|
| 36 |
|
| 37 |
```bash
|
|
|
|
| 30 |
- Architecture: Sparse Mixture-of-Experts
|
| 31 |
- Training: Converted from pretrained autoregressive base (Qwen3-30BA3B)
|
| 32 |
|
| 33 |
+
For more details, see:
|
| 34 |
+
|
| 35 |
+
- Code: https://github.com/RadicalNumerics/RND1
|
| 36 |
+
- Report: https://www.radicalnumerics.ai/rnd1-report
|
| 37 |
+
- Blog: https://www.radicalnumerics.ai/blog/rnd1
|
| 38 |
+
|
| 39 |
**Note:** RND1-Base-0910 has not been post-trained. Expect occasional repetition with greedy samplers.
|
| 40 |
|
| 41 |
+
|
| 42 |
## Installation
|
| 43 |
|
| 44 |
```bash
|