cognitivecomputations / dolphin-vision-7b

DolphinVision 7b 🐬

Curated and trained by Quan Nguyen (qnguyen3/stablequan), Eric Hartford, and Cognitive Computations

Discord: https://discord.gg/h3K4XGj2RH

Our appreciation for the sponsors of DolphinVision:

• TensorWave - provided 8x mi300x node used for training, evaluations, and inference

DolphinVision is a multimodal model. It is uncensored, and capable to reason and comment regarding images that other popular models would object to.

import torch
import transformers
from transformers import AutoModelForCausalLM, AutoTokenizer
from PIL import Image
import warnings

# disable some warnings
transformers.logging.set_verbosity_error()
transformers.logging.disable_progress_bar()
warnings.filterwarnings('ignore')

# set device
torch.set_default_device('cuda')  # or 'cpu'

model_name = 'cognitivecomputations/dolphin-vision-7b'

# create model
model = AutoModelForCausalLM.from_pretrained(
    model_name,
    torch_dtype=torch.float16,
    device_map='auto',
    trust_remote_code=True)
tokenizer = AutoTokenizer.from_pretrained(
    model_name,
    trust_remote_code=True)

# text prompt
prompt = 'Describe this image in detail'

messages = [
    {"role": "user", "content": f'<image>\n{prompt}'}
]
text = tokenizer.apply_chat_template(
    messages,
    tokenize=False,
    add_generation_prompt=True
)

print(text)

text_chunks = [tokenizer(chunk).input_ids for chunk in text.split('<image>')]
input_ids = torch.tensor(text_chunks[0] + [-200] + text_chunks[1], dtype=torch.long).unsqueeze(0)

# image, sample images can be found in images folder
image = Image.open('/path/to/image.png')
image_tensor = model.process_images([image], model.config).to(dtype=model.dtype)

# generate
output_ids = model.generate(
    input_ids,
    images=image_tensor,
    max_new_tokens=2048,
    use_cache=True)[0]

print(tokenizer.decode(output_ids[input_ids.shape[1]:], skip_special_tokens=True).strip())

Examples

<|im_start|>user
: do OCR on all the text
<|im_start|>assistant
: The text in the image reads as follows:

- STOP USING DOCKER
- Applications were not meant to be trapped in virtual jars like digital fireflies
- Years of software development, yet no one questioned why we were putting programs in imaginary boats
- Let's containerize our monolith. I'll just quickly spin up 17 microservices. Statements dreamt up by the utterly deranged
- Observe the whale-sized deception they fooled us with all this time
- These are real images used to explain Docker
- "Yay Matey, I'd like to run my website in a virtual sea"
- They have played us for absolute fools

This text is a satirical commentary on the use of Docker in software development, using humor and exaggeration to question the practice.

Benchmarks

TBD