Views
No views yet
pip install pruna1from pruna import PrunaModel
2
3loaded_model = PrunaModel.from_pretrained(
4 "radames/smashed-stabilityai-sd-turbo"
5)
6# we can then run inference using the methods supported by the base modelsmash_config.json file, which describes the optimization methods that were applied to the model.1{
2 "awq": false,
3 "c_generate": false,
4 "c_translate": false,
5 "c_whisper": false,
6 "deepcache": true,
7 "diffusers_int8": false,
8 "fastercache": false,
9 "flash_attn3": false,
10 "fora": false,
11 "gptq": false,
12 "half": false,
13 "hqq": false,
14 "hqq_diffusers": false,
15 "ifw": false,
16 "llm_int8": false,
17 "pab": false,
18 "qkv_diffusers": false,
19 "quanto": false,
20 "stable_fast": true,
21 "torch_compile": false,
22 "torch_dynamic": false,
23 "torch_structured": false,
24 "torch_unstructured": false,
25 "torchao": false,
26 "whisper_s2t": false,
27 "deepcache_interval": 2,
28 "batch_size": 1,
29 "device": "cuda",
30 "device_map": null,
31 "save_fns": [
32 "save_before_apply"
33 ],
34 "load_fns": [
35 "diffusers"
36 ],
37 "reapply_after_load": {
38 "deepcache": true,
39 "stable_fast": true
40 }
41}