nanoLLaVA-1.5

https://huggingface.co/qnguyen3/nanoLLaVA-1.5 with ONNX weights to be compatible with Transformers.js.

Usage (Transformers.js)

If you haven't already, you can install the Transformers.js JavaScript library from NPM using:

npm i @huggingface/transformers

Example:

import { AutoProcessor, AutoTokenizer, LlavaForConditionalGeneration, RawImage } from '@huggingface/transformers';

// Load tokenizer, processor and model
const model_id = 'onnx-community/nanoLLaVA-1.5';
const tokenizer = await AutoTokenizer.from_pretrained(model_id);
const processor = await AutoProcessor.from_pretrained(model_id);
const model = await LlavaForConditionalGeneration.from_pretrained(model_id, {
    dtype: {
        embed_tokens: 'fp16', // or 'fp32' or 'q8'
        vision_encoder: 'fp16', // or 'fp32' or 'q8'
        decoder_model_merged: 'q4', // or 'q8'
    },
    // device: 'webgpu',
});

// Prepare text inputs
const prompt = 'What does the text say?';
const messages = [
    { role: 'system', content: 'Answer the question.' },
    { role: 'user', content: `<image>\n${prompt}` }
]
const text = tokenizer.apply_chat_template(messages, { tokenize: false, add_generation_prompt: true });
const text_inputs = tokenizer(text);

// Prepare vision inputs
const url = 'https://huggingface.co/qnguyen3/nanoLLaVA/resolve/main/example_1.png';
const image = await RawImage.fromURL(url);
const vision_inputs = await processor(image);

// Generate response
const { past_key_values, sequences } = await model.generate({
    ...text_inputs,
    ...vision_inputs,
    do_sample: false,
    max_new_tokens: 64,
    return_dict_in_generate: true,
});

// Decode output
const answer = tokenizer.decode(
    sequences.slice(0, [text_inputs.input_ids.dims[1], null]),
    { skip_special_tokens: true },
);
console.log(answer);
// The text on the image reads "SMALL BUT MIGHTY." This phrase is likely a play on words, combining the words "small" and "mighty," suggesting that the mouse is strong and capable, despite its size.

const new_messages = [
    ...messages,
    { role: 'assistant', content: answer },
    { role: 'user', content: 'How does the text correlate to the context of the image?' }
]
const new_text = tokenizer.apply_chat_template(new_messages, { tokenize: false, add_generation_prompt: true });
const new_text_inputs = tokenizer(new_text);

// Generate another response
const output = await model.generate({
    ...new_text_inputs,
    past_key_values,
    do_sample: false,
    max_new_tokens: 256,
});
const new_answer = tokenizer.decode(
    output.slice(0, [new_text_inputs.input_ids.dims[1], null]),
    { skip_special_tokens: true },
);
console.log(new_answer);
// The text "SMALL BUT MIGHTY" correlates to the context of the image by implying that despite its size, the mouse possesses a significant amount of strength or capability. This could be a metaphor for the mouse's ability to perform tasks or overcome challenges, especially when it comes to lifting a weight.

---

Note: Having a separate repo for ONNX weights is intended to be a temporary solution until WebML gains more traction. If you would like to make your models web-ready, we recommend converting to ONNX using 🤗 Optimum and structuring your repo like this one (with ONNX weights located in a subfolder named onnx).