Spaces:
Running
Running
Update app.py
Browse files
app.py
CHANGED
@@ -1,9 +1,33 @@
|
|
1 |
import spaces
|
2 |
import gradio as gr
|
3 |
-
import transformers
|
4 |
-
from transformers import AutoTokenizer
|
5 |
-
from transformers import pipeline
|
|
|
|
|
|
|
6 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
7 |
tokenizer = AutoTokenizer.from_pretrained("EleutherAI/gpt-neox-20b")
|
8 |
|
9 |
model = transformers.AutoModelForCausalLM.from_pretrained(
|
@@ -31,6 +55,9 @@ PROMPT_FOR_GENERATION_FORMAT = """{intro}
|
|
31 |
example = "James decides to run 3 sprints 3 times a week. He runs 60 meters each sprint. How many total meters does he run a week? Explain before answering."
|
32 |
fmt_ex = PROMPT_FOR_GENERATION_FORMAT.format(instruction=example)
|
33 |
|
|
|
|
|
|
|
34 |
@spaces.GPU
|
35 |
def run():
|
36 |
with torch.autocast('cuda', dtype=torch.bfloat16):
|
@@ -39,12 +66,13 @@ def run():
|
|
39 |
max_new_tokens=100,
|
40 |
do_sample=True,
|
41 |
use_cache=True))
|
42 |
-
|
43 |
|
44 |
|
45 |
with gr.Blocks() as app:
|
46 |
btn = gr.Button()
|
47 |
-
outp=gr.Textbox()
|
|
|
48 |
btn.click(run,None,outp)
|
49 |
app.launch()
|
50 |
|
|
|
1 |
import spaces
|
2 |
import gradio as gr
|
3 |
+
#import transformers
|
4 |
+
#from transformers import AutoTokenizer
|
5 |
+
#from transformers import pipeline
|
6 |
+
from diffusers import StableDiffusionXLPipeline, UNet2DConditionModel, EulerDiscreteScheduler
|
7 |
+
from huggingface_hub import hf_hub_download
|
8 |
+
from safetensors.torch import load_file
|
9 |
|
10 |
+
base = "stabilityai/stable-diffusion-xl-base-1.0"
|
11 |
+
repo = "ByteDance/SDXL-Lightning"
|
12 |
+
ckpt = "sdxl_lightning_4step_unet.safetensors" # Use the correct ckpt for your step setting!
|
13 |
+
|
14 |
+
# Load model.
|
15 |
+
unet = UNet2DConditionModel.from_config(base, subfolder="unet").to("cuda", torch.float16)
|
16 |
+
unet.load_state_dict(load_file(hf_hub_download(repo, ckpt), device="cuda"))
|
17 |
+
pipe = StableDiffusionXLPipeline.from_pretrained(base, unet=unet, torch_dtype=torch.float16, variant="fp16").to("cuda")
|
18 |
+
|
19 |
+
# Ensure sampler uses "trailing" timesteps.
|
20 |
+
pipe.scheduler = EulerDiscreteScheduler.from_config(pipe.scheduler.config, timestep_spacing="trailing")
|
21 |
+
|
22 |
+
@spaces.GPU
|
23 |
+
def run():
|
24 |
+
# Ensure using the same inference steps as the loaded model and CFG set to 0.
|
25 |
+
return pipe("A cat", num_inference_steps=4, guidance_scale=0).images[0].save("output.png")
|
26 |
+
|
27 |
+
|
28 |
+
|
29 |
+
|
30 |
+
'''
|
31 |
tokenizer = AutoTokenizer.from_pretrained("EleutherAI/gpt-neox-20b")
|
32 |
|
33 |
model = transformers.AutoModelForCausalLM.from_pretrained(
|
|
|
55 |
example = "James decides to run 3 sprints 3 times a week. He runs 60 meters each sprint. How many total meters does he run a week? Explain before answering."
|
56 |
fmt_ex = PROMPT_FOR_GENERATION_FORMAT.format(instruction=example)
|
57 |
|
58 |
+
|
59 |
+
|
60 |
+
|
61 |
@spaces.GPU
|
62 |
def run():
|
63 |
with torch.autocast('cuda', dtype=torch.bfloat16):
|
|
|
66 |
max_new_tokens=100,
|
67 |
do_sample=True,
|
68 |
use_cache=True))
|
69 |
+
'''
|
70 |
|
71 |
|
72 |
with gr.Blocks() as app:
|
73 |
btn = gr.Button()
|
74 |
+
#outp=gr.Textbox()
|
75 |
+
outp=gr.Image()
|
76 |
btn.click(run,None,outp)
|
77 |
app.launch()
|
78 |
|