BinSaqban commited on
Commit
0ccebf7
·
verified ·
1 Parent(s): 53cb240

🔧 Remove training time and M2 hardware references

Browse files
Files changed (1) hide show
  1. README.md +1 -3
README.md CHANGED
@@ -2,7 +2,7 @@
2
  tags: [hayula, rushd, arabic, qwen, reasoning]
3
  ---
4
  # Part of **Hayula Agent OS**
5
- > **15 Specialists. 1 M2 Ultra. Zero Cloud.**
6
  ---
7
  license: other
8
  license_name: Hayula Research License v1.0
@@ -17,9 +17,7 @@ Routes queries to the appropriate Rushd specialist. 99% routing accuracy on held
17
  | Property | Value |
18
  |----------|-------|
19
  | Base Model | Qwen3-8B + LoRA (r=16) |
20
- | Training | 22 min on M2 Ultra, val loss 0.014 |
21
  | Performance | 99% routing accuracy |
22
- | Framework | MLX (Apple Silicon) |
23
  | License | Research Only — No Commercial Use |
24
  ## Usage
25
  ```python
 
2
  tags: [hayula, rushd, arabic, qwen, reasoning]
3
  ---
4
  # Part of **Hayula Agent OS**
5
+ > **15 Specialists — Zero Cloud.**
6
  ---
7
  license: other
8
  license_name: Hayula Research License v1.0
 
17
  | Property | Value |
18
  |----------|-------|
19
  | Base Model | Qwen3-8B + LoRA (r=16) |
 
20
  | Performance | 99% routing accuracy |
 
21
  | License | Research Only — No Commercial Use |
22
  ## Usage
23
  ```python