from transformers import AutoModelForVision2Seq, AutoProcessor
from qwen_vl_utils import process_vision_info
import torch

model = AutoModelForVision2Seq.from_pretrained(
    "TerraSense-CASM/TerraSense-Base",
    torch_dtype=torch.bfloat16,
    device_map="auto",
    trust_remote_code=True
)
processor = AutoProcessor.from_pretrained("TerraSense-CASM/TerraSense-Base", trust_remote_code=True)

messages = [{"role": "user", "content": [
    {"type": "image", "image": "path/to/image.jpg"},
    {"type": "text", "text": "Describe this remote sensing image."},
]}]

text = processor.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
image_inputs, _ = process_vision_info(messages)
inputs = processor(text=[text], images=image_inputs, padding=True, return_tensors="pt").to("cuda")
output = model.generate(**inputs, max_new_tokens=512)
print(processor.batch_decode(output, skip_special_tokens=True)[0])

📜 License

Apache 2.0