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_0042
Browse files
format_0042/training_log.json
ADDED
|
@@ -0,0 +1,47 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"rule_id": "format_0042",
|
| 3 |
+
"steps": 30,
|
| 4 |
+
"avg_loss": 0.298,
|
| 5 |
+
"final_loss": 0.0021,
|
| 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 |
+
3.4077,
|
| 17 |
+
0.8687,
|
| 18 |
+
1.2827,
|
| 19 |
+
0.9921,
|
| 20 |
+
0.8811,
|
| 21 |
+
0.3389,
|
| 22 |
+
0.4382,
|
| 23 |
+
0.1287,
|
| 24 |
+
0.2378,
|
| 25 |
+
0.1352,
|
| 26 |
+
0.0366,
|
| 27 |
+
0.0492,
|
| 28 |
+
0.0253,
|
| 29 |
+
0.0406,
|
| 30 |
+
0.0189,
|
| 31 |
+
0.0148,
|
| 32 |
+
0.0014,
|
| 33 |
+
0.0084,
|
| 34 |
+
0.003,
|
| 35 |
+
0.0035,
|
| 36 |
+
0.0049,
|
| 37 |
+
0.0032,
|
| 38 |
+
0.0041,
|
| 39 |
+
0.0027,
|
| 40 |
+
0.0026,
|
| 41 |
+
0.0017,
|
| 42 |
+
0.0016,
|
| 43 |
+
0.0015,
|
| 44 |
+
0.0023,
|
| 45 |
+
0.0021
|
| 46 |
+
]
|
| 47 |
+
}
|