Update README.md
Browse files
README.md
CHANGED
@@ -6,30 +6,31 @@ language:
|
|
6 |
massive thank you to [@silveroxides](https://huggingface.co/silveroxides) for phenomenal work collecting pristine state dicts and related information
|
7 |
|
8 |
#
|
|
|
|
|
9 |
|
10 |
-
# MIR (Machine Intelligence Resource)
|
11 |
-
#
|
12 |
MIR is a naming standard, a proposed schema for ML work.<br>
|
13 |
In its current incarnation, it looks like this:
|
14 |
|
15 |
-
|
16 |
-
|
17 |
-
|
18 |
-
|
19 |
-
|
20 |
-
|
21 |
-
|
|
|
22 |
```
|
23 |
|
24 |
|
25 |
-
|
26 |
-
and the super-resolution registry code from [Spandrel](https://github.com/chaiNNer-org/spandrel/blob/main/libs/spandrel/spandrel/__helpers/registry.py) library.
|
27 |
|
28 |
## Goals
|
29 |
- Standard identification scheme for all ML-related development
|
30 |
- Simplification of code for model-related logistics
|
31 |
- Rapid retrieval of resources and metadata
|
32 |
- Efficient and reliable compatability checks
|
33 |
-
-
|
34 |
|
35 |
|
|
|
6 |
massive thank you to [@silveroxides](https://huggingface.co/silveroxides) for phenomenal work collecting pristine state dicts and related information
|
7 |
|
8 |
#
|
9 |
+
> [!IMPORTANT]
|
10 |
+
> # MIR (Machine Intelligence Resource)
|
11 |
|
|
|
|
|
12 |
MIR is a naming standard, a proposed schema for ML work.<br>
|
13 |
In its current incarnation, it looks like this:
|
14 |
|
15 |
+
> [!NOTE]
|
16 |
+
> # mir : model . transformer . clip-l : stable-diffusion-xl
|
17 |
+
|
18 |
+
|
19 |
+
```
|
20 |
+
mir : model . lora . hyper : flux-1
|
21 |
+
↑ ↑ ↑ ↑ ↑
|
22 |
+
mir:[domain].[architecture].[implementation]:[compatibility]
|
23 |
```
|
24 |
|
25 |
|
26 |
+
This work is inspired by the AIR standard from CivitAi [AIR-URN](https://github.com/civitai/civitai/wiki/AIR-%E2%80%90-Uniform-Resource-Names-for-AI) project<br>
|
27 |
+
and by the super-resolution registry code from [Spandrel](https://github.com/chaiNNer-org/spandrel/blob/main/libs/spandrel/spandrel/__helpers/registry.py) library.
|
28 |
|
29 |
## Goals
|
30 |
- Standard identification scheme for all ML-related development
|
31 |
- Simplification of code for model-related logistics
|
32 |
- Rapid retrieval of resources and metadata
|
33 |
- Efficient and reliable compatability checks
|
34 |
+
- Organized hyperparameter management
|
35 |
|
36 |
|