Create asdf
Browse files
asdf
ADDED
|
@@ -0,0 +1,21 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
from transformers import LlavaNextProcessor, LlavaNextForConditionalGeneration
|
| 2 |
+
import torch
|
| 3 |
+
from PIL import Image
|
| 4 |
+
import requests
|
| 5 |
+
|
| 6 |
+
processor = LlavaNextProcessor.from_pretrained("llava-hf/llava-v1.6-mistral-7b-hf")
|
| 7 |
+
|
| 8 |
+
model = LlavaNextForConditionalGeneration.from_pretrained("llava-hf/llava-v1.6-mistral-7b-hf", torch_dtype=torch.float16, low_cpu_mem_usage=True)
|
| 9 |
+
model.to("cuda:0")
|
| 10 |
+
|
| 11 |
+
# prepare image and text prompt, using the appropriate prompt template
|
| 12 |
+
url = "https://github.com/haotian-liu/LLaVA/blob/1a91fc274d7c35a9b50b3cb29c4247ae5837ce39/images/llava_v1_5_radar.jpg?raw=true"
|
| 13 |
+
image = Image.open(requests.get(url, stream=True).raw)
|
| 14 |
+
prompt = "[INST] <image>\nWhat is shown in this image? [/INST]"
|
| 15 |
+
|
| 16 |
+
inputs = processor(prompt, image, return_tensors="pt").to("cuda:0")
|
| 17 |
+
|
| 18 |
+
# autoregressively complete prompt
|
| 19 |
+
output = model.generate(**inputs, max_new_tokens=100)
|
| 20 |
+
|
| 21 |
+
print(processor.decode(output[0], skip_special_tokens=True))
|