ZeroCool94
commited on
Commit
•
737beff
1
Parent(s):
4aec445
Update README.md
Browse files
README.md
CHANGED
@@ -67,7 +67,7 @@ image.save("fantasy_forest_illustration.png")
|
|
67 |
|
68 |
## Available Checkpoints:
|
69 |
- [Sygil Diffusion v0.1](https://huggingface.co/Sygil/Sygil-Diffusion/blob/main/sygil-diffusion-v0.1.ckpt): Trained for 800,000 steps
|
70 |
-
- [sygil-diffusion-v0.
|
71 |
|
72 |
## Training
|
73 |
|
@@ -77,12 +77,12 @@ The model was trained on the following dataset:
|
|
77 |
|
78 |
**Hardware and others**
|
79 |
- **Hardware:** 1 x Nvidia RTX 3050 8GB GPU
|
80 |
-
- **Hours Trained:**
|
81 |
- **Optimizer:** AdamW
|
82 |
-
- **Gradient Accumulations**:
|
83 |
- **Batch:** 1
|
84 |
- **Learning rate:** warmup to 1e-7 for 10,000 steps and then kept constant
|
85 |
-
- **Total Training Steps:** 1,
|
86 |
|
87 |
Developed by: [ZeroCool94](https://huggingface.co/ZeroCool94) at [Sygil-Dev](https://github.com/Sygil-Dev/)
|
88 |
|
|
|
67 |
|
68 |
## Available Checkpoints:
|
69 |
- [Sygil Diffusion v0.1](https://huggingface.co/Sygil/Sygil-Diffusion/blob/main/sygil-diffusion-v0.1.ckpt): Trained for 800,000 steps
|
70 |
+
- [sygil-diffusion-v0.2_1344635_lora.ckpt](https://huggingface.co/Sygil/Sygil-Diffusion/blob/main/sygil-diffusion-v0.2_1344635_lora.ckpt): Resumed from Sygil Diffusion v0.1 and now up to 1.34 million steps.
|
71 |
|
72 |
## Training
|
73 |
|
|
|
77 |
|
78 |
**Hardware and others**
|
79 |
- **Hardware:** 1 x Nvidia RTX 3050 8GB GPU
|
80 |
+
- **Hours Trained:** 496 hours approximately.
|
81 |
- **Optimizer:** AdamW
|
82 |
+
- **Gradient Accumulations**: 2
|
83 |
- **Batch:** 1
|
84 |
- **Learning rate:** warmup to 1e-7 for 10,000 steps and then kept constant
|
85 |
+
- **Total Training Steps:** 1,344,635
|
86 |
|
87 |
Developed by: [ZeroCool94](https://huggingface.co/ZeroCool94) at [Sygil-Dev](https://github.com/Sygil-Dev/)
|
88 |
|