Instructions to use mnmly/anycalib-mlx with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- MLX
How to use mnmly/anycalib-mlx with MLX:
# Download the model from the Hub pip install huggingface_hub[hf_xet] huggingface-cli download --local-dir anycalib-mlx mnmly/anycalib-mlx
- Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- LM Studio
- Atomic Chat
Upload folder using huggingface_hub
Browse files
anycalib_edit/config.json
ADDED
|
@@ -0,0 +1,25 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architecture": "anycalib",
|
| 3 |
+
"backbone": "dinov2_vitl14",
|
| 4 |
+
"embed_dim": 1024,
|
| 5 |
+
"depth": 24,
|
| 6 |
+
"num_heads": 16,
|
| 7 |
+
"patch_size": 14,
|
| 8 |
+
"out_index": [
|
| 9 |
+
4,
|
| 10 |
+
11,
|
| 11 |
+
17,
|
| 12 |
+
23
|
| 13 |
+
],
|
| 14 |
+
"norm_layer": false,
|
| 15 |
+
"decoder_dim_dhat": 256,
|
| 16 |
+
"up_factor": 7,
|
| 17 |
+
"resolution": 102400,
|
| 18 |
+
"edge_divisible_by": 14,
|
| 19 |
+
"ar_range": [
|
| 20 |
+
0.5,
|
| 21 |
+
2.0
|
| 22 |
+
],
|
| 23 |
+
"dtype": "float32",
|
| 24 |
+
"source_checkpoint": "anycalib_edit.pt"
|
| 25 |
+
}
|
anycalib_edit/weights.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3796de458570c9de71f2c16c33cb3d48ae640388aebab6be20c2a20de53ecc28
|
| 3 |
+
size 1281083116
|