MolmoE-1B-0924 / example.py
chrisc36's picture
Add files using upload-large-folder tool
3f119eb verified
from transformers import AutoProcessor, AutoModelForCausalLM, GenerationConfig
from PIL import Image
import requests
def main():
load_path = "."
# load the processor
print("Loading processor")
processor = AutoProcessor.from_pretrained(
load_path,
trust_remote_code=True,
torch_dtype='auto',
device_map='auto'
)
# load the model
print("Loading model")
model = AutoModelForCausalLM.from_pretrained(
load_path,
trust_remote_code=True,
torch_dtype='auto',
device_map='auto'
)
# process the image and text
print("Processing...")
inputs = processor.process(
images=[Image.open(requests.get("https://picsum.photos/id/237/536/354", stream=True).raw)],
text="Describe this image."
)
# move inputs to the correct device and make a batch of size 1
inputs = {k: v.to(model.device).unsqueeze(0) for k, v in inputs.items()}
# generate output; maximum 200 new tokens; stop generation when <|endoftext|> is generated
print("Generating....")
output = model.generate_from_batch(
inputs,
GenerationConfig(max_new_tokens=200, stop_strings="<|endoftext|>"),
tokenizer=processor.tokenizer
)
# only get generated tokens; decode them to text
generated_tokens = output[0,inputs['input_ids'].size(1):]
generated_text = processor.tokenizer.decode(generated_tokens, skip_special_tokens=True)
# print the generated text
print(generated_text)
if __name__ == '__main__':
main()