Instructions to use broadfield-dev/lora-adapters-qwen3-0.6b-r16-n180 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- PEFT
How to use broadfield-dev/lora-adapters-qwen3-0.6b-r16-n180 with PEFT:
Task type is invalid.
- Notebooks
- Google Colab
- Kaggle
Upload LoRA adapter: format_0036
Browse files
format_0036/training_log.json
ADDED
|
@@ -0,0 +1,47 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"rule_id": "format_0036",
|
| 3 |
+
"steps": 30,
|
| 4 |
+
"avg_loss": 0.283,
|
| 5 |
+
"final_loss": 0.0023,
|
| 6 |
+
"lr": 0.0002,
|
| 7 |
+
"rank": 16,
|
| 8 |
+
"alpha": 32,
|
| 9 |
+
"target_modules": [
|
| 10 |
+
"q_proj",
|
| 11 |
+
"k_proj",
|
| 12 |
+
"v_proj",
|
| 13 |
+
"o_proj"
|
| 14 |
+
],
|
| 15 |
+
"step_losses": [
|
| 16 |
+
2.2623,
|
| 17 |
+
1.2436,
|
| 18 |
+
1.5673,
|
| 19 |
+
1.5659,
|
| 20 |
+
0.4185,
|
| 21 |
+
0.2599,
|
| 22 |
+
0.1449,
|
| 23 |
+
0.4237,
|
| 24 |
+
0.2427,
|
| 25 |
+
0.1709,
|
| 26 |
+
0.0631,
|
| 27 |
+
0.0025,
|
| 28 |
+
0.0217,
|
| 29 |
+
0.0158,
|
| 30 |
+
0.0118,
|
| 31 |
+
0.0163,
|
| 32 |
+
0.0126,
|
| 33 |
+
0.0085,
|
| 34 |
+
0.0036,
|
| 35 |
+
0.0016,
|
| 36 |
+
0.0047,
|
| 37 |
+
0.0054,
|
| 38 |
+
0.0011,
|
| 39 |
+
0.0016,
|
| 40 |
+
0.003,
|
| 41 |
+
0.0037,
|
| 42 |
+
0.0035,
|
| 43 |
+
0.0033,
|
| 44 |
+
0.0032,
|
| 45 |
+
0.0023
|
| 46 |
+
]
|
| 47 |
+
}
|