Update README.md
Browse files
README.md
CHANGED
|
@@ -14,21 +14,26 @@ Google/Gemma3-4b-it
|
|
| 14 |
|
| 15 |
Jina Ai/Jina Clip v2
|
| 16 |
|
| 17 |
-
Total: approximately 5b
|
| 18 |
|
| 19 |
>Networks:
|
| 20 |
|
| 21 |
-
Next-DIT (modified from 26 layers to 36 layers) 3.5b
|
| 22 |
|
| 23 |
>VAE:
|
| 24 |
|
| 25 |
-
Flux 1 Dev-VAE
|
| 26 |
|
| 27 |
Pretrain Dataset: full dan + 1m e621
|
|
|
|
| 28 |
Use 8*h200 train for four months (23000 h200 hour)
|
|
|
|
| 29 |
Restructure text data use XML format
|
|
|
|
| 30 |
The model is currently in the train phase (60% complete).
|
|
|
|
| 31 |
Overall progress is approximately 80%.
|
|
|
|
| 32 |
The model is expected to be open sourced on Hugging Face on December 31, 2025.
|
| 33 |
|
| 34 |
example：
|
|
|
|
| 14 |
|
| 15 |
Jina Ai/Jina Clip v2
|
| 16 |
|
| 17 |
+
Total: approximately 5b<
|
| 18 |
|
| 19 |
>Networks:
|
| 20 |
|
| 21 |
+
Next-DIT (modified from 26 layers to 36 layers) 3.5b<
|
| 22 |
|
| 23 |
>VAE:
|
| 24 |
|
| 25 |
+
Flux 1 Dev-VAE<
|
| 26 |
|
| 27 |
Pretrain Dataset: full dan + 1m e621
|
| 28 |
+
|
| 29 |
Use 8*h200 train for four months (23000 h200 hour)
|
| 30 |
+
|
| 31 |
Restructure text data use XML format
|
| 32 |
+
|
| 33 |
The model is currently in the train phase (60% complete).
|
| 34 |
+
|
| 35 |
Overall progress is approximately 80%.
|
| 36 |
+
|
| 37 |
The model is expected to be open sourced on Hugging Face on December 31, 2025.
|
| 38 |
|
| 39 |
example：
|