OpenRaiser/Pager

1---2library_name: transformers3pipeline_tag: image-text-to-text4tags:5- qwen3-vl6- vision-language7- multimodal8- image-text-to-text9---10 11# Pager12 13This repository contains the model weights, tokenizer, processor, and configuration files for **Pager**, a vision-language model based on the Qwen3-VL architecture.14 15## Files16 17The repository includes:18 19- `config.json`20- `generation_config.json`21- `tokenizer.json`22- `tokenizer_config.json`23- `vocab.json`24- `merges.txt`25- `special_tokens_map.json`26- `added_tokens.json`27- `preprocessor_config.json`28- `video_preprocessor_config.json`29- `chat_template.jinja`30- `model.safetensors.index.json`31- `model-00001-of-00004.safetensors`32- `model-00002-of-00004.safetensors`33- `model-00003-of-00004.safetensors`34- `model-00004-of-00004.safetensors`35 36## Usage37 38Install dependencies:39 40```bash41pip install -U transformers accelerate safetensors pillow42```43 44Load the model:45 46```python47import torch48from transformers import AutoProcessor, AutoModelForImageTextToText49 50model_id = "OpenRaiser/Pager"51 52processor = AutoProcessor.from_pretrained(53    model_id,54    trust_remote_code=True55)56 57model = AutoModelForImageTextToText.from_pretrained(58    model_id,59    torch_dtype="auto",60    device_map="auto",61    trust_remote_code=True62)63 64print("Model loaded successfully.")65```66 67If your local `transformers` version does not support this model class, please upgrade `transformers` first.68 69## Notes70 71- The model weights are stored in four `.safetensors` shards.72- `model.safetensors.index.json` maps model parameters to the corresponding weight shards.73- This repository is intended for research and development use.74 75## Citation76 77If you use this model, please cite or link to this repository:78 79```text80https://huggingface.co/OpenRaiser/Pager81```