michaelgold commited on
Commit
98ab7cf
·
1 Parent(s): 10f8a38

feat: add big-lama.safetensors

Browse files
Files changed (3) hide show
  1. README.md +30 -0
  2. big-lama.safetensors +3 -0
  3. config.json +7 -0
README.md ADDED
@@ -0,0 +1,30 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: mit
3
+ tags:
4
+ - image-inpainting
5
+ - LaMa
6
+ - PyTorch
7
+ - safetensors
8
+ library_name: none
9
+ ---
10
+
11
+ # Big-LaMa Inpainting Model (Safetensors)
12
+
13
+ This is the `big-lama` image inpainting model, converted to the [safetensors](https://github.com/huggingface/safetensors) format from the original `.pt` release by [Sanster](https://github.com/Sanster/models).
14
+
15
+ - ✅ Only the `generator` weights are included
16
+ - ❌ Evaluation code and FID scoring models have been removed
17
+ - ✅ TorchScript converted to state_dict
18
+ - ✅ 100% compatible with inference pipelines based on [LaMa](https://github.com/advimman/lama)
19
+
20
+ ## How to use
21
+
22
+ ```python
23
+ from safetensors.torch import load_file
24
+ from huggingface_hub import hf_hub_download
25
+
26
+ # Download from Hub
27
+ weights_path = hf_hub_download("michaelgold/big-lama-safetensors", "big-lama.safetensors")
28
+
29
+ # Load into your model
30
+ state_dict = load_file(weights_path)
big-lama.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:326ae1889b92d07c8819c35eaf561f8b290d057536eeb12524b72997f31005ec
3
+ size 204342884
config.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "format": "safetensors",
3
+ "architecture": "LaMaGenerator",
4
+ "source": "https://github.com/Sanster/models",
5
+ "converted_by": "michaelgold",
6
+ "inference_only": true
7
+ }