Hugging Face's logo

Bedovyy
/
Anima-FP8

Text-to-Image
Diffusion Single File
comfyui
Model card Files
xet
 Community
5
Bedovyy commited on
Commit
8f000df
·
verified ·
1 Parent(s): 2098531

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +24 -25
README.md CHANGED
@@ -13,6 +13,30 @@ There are two models - FP8 and FP8Mixed.
13
 
14
  FP8Mixed quantized fewer layers, so expects better quality.
15
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
16
  ## Quantized layers
17
 
18
  ### fp8
@@ -40,28 +64,3 @@ FP8Mixed quantized fewer layers, so expects better quality.
40
  ]
41
  }
42
  ```
43
-
44
- ## Generation speed
45
-
46
- Tested on
47
- - RTX5090 (400W), ComfyUI with torch2.10.0+cu130
48
- - Generates 832x1216, 30steps, cfg 4.0, er sde, simple
49
-
50
- | Dtype (Quant) | Sage Attn | it/s | Time (s) | Relative Speedup (%) |
51
- | ------------- | --------- | ---- | -------- | -------------------- |
52
- | bf16 | X | 4.33 | 7.07 | 0% |
53
- | bf16 | O | 4.66 | 6.59 | +7.6% |
54
- | fp8mixed | X | 4.79 | 6.42 | +10.6% |
55
- | fp8mixed | O | 5.25 | 5.87 | +21.3% |
56
- | fp8 | X | 4.94 | 6.23 | +14.1% |
57
- | fp8 | O | 5.40 | 5.72 | +24.7% |
58
-
59
-
60
- ## Sample
61
-
62
- | bf16 | fp8mixed | fp8 |
63
- |-----|------|------|
64
- |![ComfyUI_temp_onhsj_00007_](https://cdn-uploads.huggingface.co/production/uploads/63fbf6951b4b1bd4e706fed1/4OzIFBhl_FipyILF3ZXyb.png)|![ComfyUI_temp_gqpfh_00024_](https://cdn-uploads.huggingface.co/production/uploads/63fbf6951b4b1bd4e706fed1/GiVW4VHm1AnMB73q1mv1w.png)|![ComfyUI_temp_onhsj_00006_](https://cdn-uploads.huggingface.co/production/uploads/63fbf6951b4b1bd4e706fed1/Bk0o1slEWvdbO9NynBWGr.png)|
65
- |![anima_bf16_1](https://cdn-uploads.huggingface.co/production/uploads/63fbf6951b4b1bd4e706fed1/_u2hrgEulSMOPiNux9bYI.png)|![anima_fp8mixed_1](https://cdn-uploads.huggingface.co/production/uploads/63fbf6951b4b1bd4e706fed1/dWEUFK3lFLK8_wsiRDMzJ.png)|![anima_fp8_1](https://cdn-uploads.huggingface.co/production/uploads/63fbf6951b4b1bd4e706fed1/b769101lzVGLjPIGJApJp.png)|
66
-
67
-
 
13
 
14
  FP8Mixed quantized fewer layers, so expects better quality.
15
 
16
+
17
+ ## Generation speed
18
+
19
+ Tested on
20
+ - RTX5090 (400W), ComfyUI with torch2.10.0+cu130
21
+ - Generates 832x1216, 30steps, cfg 4.0, er sde, simple
22
+
23
+ | Dtype (Quant) | Sage Attn | it/s | Time (s) | Relative Speedup (%) |
24
+ | ------------- | --------- | ---- | -------- | -------------------- |
25
+ | bf16 | X | 4.33 | 7.07 | 0% |
26
+ | bf16 | O | 4.66 | 6.59 | +7.6% |
27
+ | fp8mixed | X | 4.79 | 6.42 | +10.6% |
28
+ | fp8mixed | O | 5.25 | 5.87 | +21.3% |
29
+ | fp8 | X | 4.94 | 6.23 | +14.1% |
30
+ | fp8 | O | 5.40 | 5.72 | +24.7% |
31
+
32
+
33
+ ## Sample
34
+
35
+ | bf16 | fp8mixed | fp8 |
36
+ |------|----------|-----|
37
+ |<img src="https://cdn-uploads.huggingface.co/production/uploads/63fbf6951b4b1bd4e706fed1/4OzIFBhl_FipyILF3ZXyb.png" width="832">|<img src="https://cdn-uploads.huggingface.co/production/uploads/63fbf6951b4b1bd4e706fed1/GiVW4VHm1AnMB73q1mv1w.png" width="832">|<img src="https://cdn-uploads.huggingface.co/production/uploads/63fbf6951b4b1bd4e706fed1/Bk0o1slEWvdbO9NynBWGr.png" width="832">|
38
+ |<img src="https://cdn-uploads.huggingface.co/production/uploads/63fbf6951b4b1bd4e706fed1/_u2hrgEulSMOPiNux9bYI.png" width="832">|<img src="https://cdn-uploads.huggingface.co/production/uploads/63fbf6951b4b1bd4e706fed1/dWEUFK3lFLK8_wsiRDMzJ.png" width="832">|<img src="https://cdn-uploads.huggingface.co/production/uploads/63fbf6951b4b1bd4e706fed1/b769101lzVGLjPIGJApJp.png" width="832">|
39
+
40
  ## Quantized layers
41
 
42
  ### fp8
 
64
  ]
65
  }
66
  ```