sharpenb commited on
Commit
3920125
1 Parent(s): 51d6c2b

a53eecaa316f75d30ff366f5d820294a9d503905c949f7b342de3f4b178b4a5c

Browse files
Files changed (4) hide show
  1. README.md +45 -18
  2. config.json +1 -1
  3. model/smash_config.json +3 -0
  4. plots.png +0 -0
README.md CHANGED
@@ -19,40 +19,67 @@ metrics:
19
  </div>
20
  <!-- header end -->
21
 
 
 
 
 
 
22
  # Simply make AI models cheaper, smaller, faster, and greener!
23
 
 
 
 
 
 
 
24
  ## Results
25
 
26
  ![image info](./plots.png)
27
 
 
 
 
 
 
28
  ## Setup
29
 
30
- You can run the smashed model by:
31
- 1. Installing and importing the `pruna-engine` (version 0.2.9) package. Use `pip install pruna-engine==0.2.9 --extra-index-url https://pypi.nvidia.com --extra-index-url https://pypi.ngc.nvidia.com` for installation. See [Pypi](https://pypi.org/project/pruna-engine/) for details on the package.
32
- 2. Downloading the model files. This can be done using the Hugging Face CLI with the following commands:
 
33
  ```bash
34
- mkdir CompVis-stable-diffusion-v1-4-smashed
35
- huggingface-cli download PrunaAI/segmind-SSD-1B-turbo-tiny-green-smashed --local-dir segmind-SSD-1B-turbo-tiny-green-smashed --local-dir-use-symlinks False
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
36
  ```
37
- Alternatively, you can download them manually.
38
- 3. Loading the model.
39
- 4. Running the model.
40
- You can achieve this by running the following code:
41
-
42
- ```python
43
- from pruna_engine.PrunaModel import PrunaModel # Step (1): install and import `pruna-engine` package.
44
- model_path = "segmind-SSD-1B-turbo-tiny-green-smashed/model" # Step (2): specify the downloaded model path.
45
- smashed_model = PrunaModel.load_model(model_path) # Step (3): load the model.
46
- y = smashed_model(prompt="an astronaut riding a horse on mars", image_height=1024, image_width=1024)[0] # Step (4): run the model.
47
- ```
48
 
49
  ## Configurations
50
 
51
  The configuration info are in `config.json`.
52
 
53
- ## License
54
 
55
- We follow the same license as the original model. Please check the license of the original model before using this model.
56
 
57
  ## Want to compress other models?
58
 
 
19
  </div>
20
  <!-- header end -->
21
 
22
+ [![Twitter](https://img.shields.io/twitter/follow/PrunaAI?style=social)](https://twitter.com/PrunaAI)
23
+ [![GitHub](https://img.shields.io/github/followers/PrunaAI?label=Follow%20%40PrunaAI&style=social)](https://github.com/PrunaAI)
24
+ [![LinkedIn](https://img.shields.io/badge/LinkedIn-Connect-blue)](https://www.linkedin.com/company/93832878/admin/feed/posts/?feedType=following)
25
+ [![Discord](https://img.shields.io/badge/Discord-Join%20Us-blue?style=social&logo=discord)](https://discord.gg/CP4VSgck)
26
+
27
  # Simply make AI models cheaper, smaller, faster, and greener!
28
 
29
+ - Give a thumbs up if you like this model!
30
+ - Contact us and tell us which model to compress next [here](https://www.pruna.ai/contact).
31
+ - Request access to easily compress your *own* AI models [here](https://z0halsaff74.typeform.com/pruna-access?typeform-source=www.pruna.ai).
32
+ - Read the documentations to know more [here](https://pruna-ai-pruna.readthedocs-hosted.com/en/latest/)
33
+ - Join Pruna AI community on Discord [here](https://discord.gg/CP4VSgck) to share feedback/suggestions or get help.
34
+
35
  ## Results
36
 
37
  ![image info](./plots.png)
38
 
39
+ **Important remarks:**
40
+ - The quality of the model output might slightly vary compared to the base model. There might be minimal quality loss.
41
+ - These results were obtained on NVIDIA A100-PCIE-40GB with configuration described in config.json and are obtained after a hardware warmup. Efficiency results may vary in other settings (e.g. other hardware, image size, batch size, ...).
42
+ - You can request premium access to more compression methods and tech support for your specific use-cases [here](https://z0halsaff74.typeform.com/pruna-access?typeform-source=www.pruna.ai).
43
+
44
  ## Setup
45
 
46
+ You can run the smashed model with these steps:
47
+
48
+ 0. Check cuda, torch, packaging requirements are installed. For cuda, check with `nvcc --version` and install with `conda install nvidia/label/cuda-12.1.0::cuda`. For packaging and torch, run `pip install packaging torch`.
49
+ 1. Install the `pruna-engine` available [here](https://pypi.org/project/pruna-engine/) on Pypi. It might take 15 minutes to install.
50
  ```bash
51
+ pip install pruna-engine[gpu]==0.6.0 --extra-index-url https://pypi.nvidia.com --extra-index-url https://pypi.ngc.nvidia.com --extra-index-url https://prunaai.pythonanywhere.com/
52
+ ```
53
+ 3. Download the model files using one of these three options.
54
+ - Option 1 - Use command line interface (CLI):
55
+ ```bash
56
+ mkdir segmind-SSD-1B-turbo-tiny-green-smashed
57
+ huggingface-cli download PrunaAI/segmind-SSD-1B-turbo-tiny-green-smashed --local-dir segmind-SSD-1B-turbo-tiny-green-smashed --local-dir-use-symlinks False
58
+ ```
59
+ - Option 2 - Use Python:
60
+ ```python
61
+ import subprocess
62
+ repo_name = "segmind-SSD-1B-turbo-tiny-green-smashed"
63
+ subprocess.run(["mkdir", repo_name])
64
+ subprocess.run(["huggingface-cli", "download", 'PrunaAI/'+ repo_name, "--local-dir", repo_name, "--local-dir-use-symlinks", "False"])
65
+ ```
66
+ - Option 3 - Download them manually on the HuggingFace model page.
67
+ 3. Load & run the model.
68
+ ```python
69
+ from pruna_engine.PrunaModel import PrunaModel
70
+
71
+ model_path = "segmind-SSD-1B-turbo-tiny-green-smashed/model" # Specify the downloaded model path.
72
+ smashed_model = PrunaModel.load_model(model_path) # Load the model.
73
+ smashed_model(prompt='Beautiful fruits in trees', height=1024, width=1024)[0][0] # Run the model where x is the expected input of.
74
  ```
 
 
 
 
 
 
 
 
 
 
 
75
 
76
  ## Configurations
77
 
78
  The configuration info are in `config.json`.
79
 
80
+ ## Credits & License
81
 
82
+ We follow the same license as the original model. Please check the license of the original model segmind/SSD-1B before using this model which provided the base model.
83
 
84
  ## Want to compress other models?
85
 
config.json CHANGED
@@ -1 +1 @@
1
- {"pruner": "None", "pruning_ratio": 0.0, "factorizer": "None", "quantizer": "None", "n_quantization_bits": 32, "output_deviation": 0.005, "compiler": "diffusers", "static_batch": true, "static_shape": true, "controlnet": "None", "unet_dim": 4, "max_batch_size": 1, "image_height": 1024, "image_width": 1024, "version": "xl-1.0"}
 
1
+ {"pruners": "None", "pruning_ratio": 0.0, "factorizers": "None", "quantizers": "None", "n_quantization_bits": 32, "output_deviation": 0.005, "compilers": "['diffusers2', 'tiling', 'step_caching']", "static_batch": true, "static_shape": false, "controlnet": "None", "unet_dim": 4, "device": "cuda", "batch_size": 1, "max_batch_size": 1, "image_height": 1024, "image_width": 1024, "version": "xl-1.0", "scheduler": "DDIM", "task": "txt2imgxl", "model_name": "segmind/SSD-1B", "weight_name": "None", "save_load_fn": "stable_fast"}
model/smash_config.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9b204d22b90e6eed84e4762b0f1e04a52cded9a8b7a9539facdde272a60a4859
3
+ size 695
plots.png CHANGED