stduhpf commited on
Commit
1de7f71
·
verified ·
1 Parent(s): fc71e94

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +6 -3
README.md CHANGED
@@ -22,13 +22,15 @@ Only the second layers of both MLPs in each MMDiT block of SD3.5 Large models ha
22
 
23
  - [sd3.5_large_turbo-q2_k_4_0.gguf](https://huggingface.co/stduhpf/SD3.5-Large-Turbo-GGUF-mixed-sdcpp/blob/main/sd3.5_large_turbo-q2_k_4_0.gguf): Smallest quantization yet. Use this if you can't afford anything bigger
24
  - [sd3.5_large_turbo-q3_k_4_0.gguf](https://huggingface.co/stduhpf/SD3.5-Large-Turbo-GGUF-mixed-sdcpp/blob/main/sd3.5_large_turbo-q3_k_4_0.gguf): Smaller than q4_0, acceptable degradation.
25
- - [sd3.5_large_turbo-q4_k_4_0.gguf](https://huggingface.co/stduhpf/SD3.5-Large-Turbo-GGUF-mixed-sdcpp/blob/main/sd3.5_large_turbo-q4_k_4_0.gguf): Exacty same size as q4_0, but with slightly less degradation. Recommended
26
  - [sd3.5_large_turbo-q4_k_4_1.gguf](https://huggingface.co/stduhpf/SD3.5-Large-Turbo-GGUF-mixed-sdcpp/blob/main/sd3.5_large_turbo-q4_k_4_1.gguf): Smaller than q4_1, and with comparable degradation. Recommended
27
  - [sd3.5_large_turbo-q4_k_5_0.gguf](https://huggingface.co/stduhpf/SD3.5-Large-Turbo-GGUF-mixed-sdcpp/blob/main/sd3.5_large_turbo-q4_k_5_0.gguf): Smaller than q5_0, and with comparable degradation. Very close to the original f16 already. Recommended
28
 
29
  ### Legacy types:
30
 
31
- - [sd3.5_large_turbo-q4_0.gguf](https://huggingface.co/stduhpf/SD3.5-Large-Turbo-GGUF-mixed-sdcpp/blob/main/legacy/sd3.5_large_turbo-q4_0.gguf): Same size as q4_k_4_0, Not recommended (use q4_k_4_0 instead)
 
 
32
  - [sd3.5_large_turbo-q4_1.gguf](https://huggingface.co/stduhpf/SD3.5-Large-Turbo-GGUF-mixed-sdcpp/blob/main/legacy/sd3.5_large_turbo-q4_1.gguf): Not recommended (q4_k_4_1 is better and smaller)
33
  - [sd3.5_large_turbo-q5_0.gguf](https://huggingface.co/stduhpf/SD3.5-Large-Turbo-GGUF-mixed-sdcpp/blob/main/legacy/sd3.5_large_turbo-q5_0.gguf): Barely better and bigger than q4_k_5_0
34
  - [sd3.5_large_turbo-q5_1.gguf](https://huggingface.co/stduhpf/SD3.5-Large-Turbo-GGUF-mixed-sdcpp/blob/main/legacy/sd3.5_large_turbo-q5_1.gguf): Better and bigger than q5_0
@@ -36,7 +38,7 @@ Only the second layers of both MLPs in each MMDiT block of SD3.5 Large models ha
36
 
37
  ## Outputs:
38
 
39
- Sorted by model size (Note that q4_0 and q4_k_4_0 are the exact same size)
40
 
41
  | Quantization | Robot girl | Text | Cute kitten |
42
  | ------------------ | -------------------------------- | ---------------------------------- | ---------------------------------- |
@@ -44,6 +46,7 @@ Sorted by model size (Note that q4_0 and q4_k_4_0 are the exact same size)
44
  | q3_k_4_0 | ![q3_k_4_0](Images/q3_k_4_0.png) | ![q3_k_4_0](Images/1_q3_k_4_0.png) | ![q3_k_4_0](Images/2_q3_k_4_0.png) |
45
  | q4_0 | ![q4_0](Images/q4_0.png) | ![q4_0](Images/1_q4_0.png) | ![q4_0](Images/2_q4_0.png) |
46
  | q4_k_4_0 | ![q4_k_4_0](Images/q4_k_4_0.png) | ![q4_k_4_0](Images/1_q4_k_4_0.png) | ![q4_k_4_0](Images/2_q4_k_4_0.png) |
 
47
  | q4_k_4_1 | ![q4_k_4_1](Images/q4_k_4_1.png) | ![q4_k_4_1](Images/1_q4_k_4_1.png) | ![q4_k_4_1](Images/2_q4_k_4_1.png) |
48
  | q4_1 | ![q4_1](Images/q4_1.png) | ![q4_1](Images/1_q4_1.png) | ![q4_1](Images/2_q4_1.png) |
49
  | q4_k_5_0 | ![q4_k_5_0](Images/q4_k_5_0.png) | ![q4_k_5_0](Images/1_q4_k_5_0.png) | ![q4_k_5_0](Images/2_q4_k_5_0.png) |
 
22
 
23
  - [sd3.5_large_turbo-q2_k_4_0.gguf](https://huggingface.co/stduhpf/SD3.5-Large-Turbo-GGUF-mixed-sdcpp/blob/main/sd3.5_large_turbo-q2_k_4_0.gguf): Smallest quantization yet. Use this if you can't afford anything bigger
24
  - [sd3.5_large_turbo-q3_k_4_0.gguf](https://huggingface.co/stduhpf/SD3.5-Large-Turbo-GGUF-mixed-sdcpp/blob/main/sd3.5_large_turbo-q3_k_4_0.gguf): Smaller than q4_0, acceptable degradation.
25
+ - [sd3.5_large_turbo-q4_k_4_0.gguf](https://huggingface.co/stduhpf/SD3.5-Large-Turbo-GGUF-mixed-sdcpp/blob/main/sd3.5_large_turbo-q4_k_4_0.gguf): Exacty same size as q4_0 and iq4_nl, I recommend using iq4_nl instead.
26
  - [sd3.5_large_turbo-q4_k_4_1.gguf](https://huggingface.co/stduhpf/SD3.5-Large-Turbo-GGUF-mixed-sdcpp/blob/main/sd3.5_large_turbo-q4_k_4_1.gguf): Smaller than q4_1, and with comparable degradation. Recommended
27
  - [sd3.5_large_turbo-q4_k_5_0.gguf](https://huggingface.co/stduhpf/SD3.5-Large-Turbo-GGUF-mixed-sdcpp/blob/main/sd3.5_large_turbo-q4_k_5_0.gguf): Smaller than q5_0, and with comparable degradation. Very close to the original f16 already. Recommended
28
 
29
  ### Legacy types:
30
 
31
+ - [sd3.5_large_turbo-iq4_nl.gguf](https://huggingface.co/stduhpf/SD3.5-Large-Turbo-GGUF-mixed-sdcpp/blob/main/legacy/sd3.5_large_turbo-iq4_nl.gguf): Same size as q4_k_4_0 and q4_0, runs faster than q4_k_4_0 (on Vulkan at least), and provides better image quality. Recommended
32
+
33
+ - [sd3.5_large_turbo-q4_0.gguf](https://huggingface.co/stduhpf/SD3.5-Large-Turbo-GGUF-mixed-sdcpp/blob/main/legacy/sd3.5_large_turbo-q4_0.gguf): Same size as q4_k_4_0, Not recommended (use iqk_nl q4_k_4_0 instead)
34
  - [sd3.5_large_turbo-q4_1.gguf](https://huggingface.co/stduhpf/SD3.5-Large-Turbo-GGUF-mixed-sdcpp/blob/main/legacy/sd3.5_large_turbo-q4_1.gguf): Not recommended (q4_k_4_1 is better and smaller)
35
  - [sd3.5_large_turbo-q5_0.gguf](https://huggingface.co/stduhpf/SD3.5-Large-Turbo-GGUF-mixed-sdcpp/blob/main/legacy/sd3.5_large_turbo-q5_0.gguf): Barely better and bigger than q4_k_5_0
36
  - [sd3.5_large_turbo-q5_1.gguf](https://huggingface.co/stduhpf/SD3.5-Large-Turbo-GGUF-mixed-sdcpp/blob/main/legacy/sd3.5_large_turbo-q5_1.gguf): Better and bigger than q5_0
 
38
 
39
  ## Outputs:
40
 
41
+ Sorted by model size (Note that q4_0, q4_k_4_0, and iq4_nl are the exact same size)
42
 
43
  | Quantization | Robot girl | Text | Cute kitten |
44
  | ------------------ | -------------------------------- | ---------------------------------- | ---------------------------------- |
 
46
  | q3_k_4_0 | ![q3_k_4_0](Images/q3_k_4_0.png) | ![q3_k_4_0](Images/1_q3_k_4_0.png) | ![q3_k_4_0](Images/2_q3_k_4_0.png) |
47
  | q4_0 | ![q4_0](Images/q4_0.png) | ![q4_0](Images/1_q4_0.png) | ![q4_0](Images/2_q4_0.png) |
48
  | q4_k_4_0 | ![q4_k_4_0](Images/q4_k_4_0.png) | ![q4_k_4_0](Images/1_q4_k_4_0.png) | ![q4_k_4_0](Images/2_q4_k_4_0.png) |
49
+ | iq4_nl | ![iq4_nl](Images/iq4_nl.png) | ![iq4_nl](Images/1_iq4_nl.png) | ![iq4_nl](Images/2_iq4_nl.png) |
50
  | q4_k_4_1 | ![q4_k_4_1](Images/q4_k_4_1.png) | ![q4_k_4_1](Images/1_q4_k_4_1.png) | ![q4_k_4_1](Images/2_q4_k_4_1.png) |
51
  | q4_1 | ![q4_1](Images/q4_1.png) | ![q4_1](Images/1_q4_1.png) | ![q4_1](Images/2_q4_1.png) |
52
  | q4_k_5_0 | ![q4_k_5_0](Images/q4_k_5_0.png) | ![q4_k_5_0](Images/1_q4_k_5_0.png) | ![q4_k_5_0](Images/2_q4_k_5_0.png) |