add adamw + params to training hyperparam
Browse files
README.md
CHANGED
|
@@ -88,6 +88,7 @@ This model used weights pretrained by [lxj616](https://huggingface.co/lxj616/mak
|
|
| 88 |
* **Batch size:** 1 x 4
|
| 89 |
* **Image size:** 512 x 512
|
| 90 |
* **Frame count:** 24
|
|
|
|
| 91 |
* **Schedule:**
|
| 92 |
* 2 x 10 epochs: LR warmup for 1 epochs then held constant at 5e-5 (10,000 samples per ep)
|
| 93 |
* 2 x 20 epochs: LR warmup for 1 epochs then held constant at 5e-5 (10,000 samples per ep)
|
|
|
|
| 88 |
* **Batch size:** 1 x 4
|
| 89 |
* **Image size:** 512 x 512
|
| 90 |
* **Frame count:** 24
|
| 91 |
+
* **Optimizer:** AdamW (beta_1 = 0.9, beta_2 = 0.999, weight decay = 0.02)
|
| 92 |
* **Schedule:**
|
| 93 |
* 2 x 10 epochs: LR warmup for 1 epochs then held constant at 5e-5 (10,000 samples per ep)
|
| 94 |
* 2 x 20 epochs: LR warmup for 1 epochs then held constant at 5e-5 (10,000 samples per ep)
|