Visual Document Retrieval
sentence-transformers
Safetensors
Arabic
qwen3_vl
sentence-similarity
feature-extraction
Generated from Trainer
dataset_size:48002
loss:MatryoshkaLoss
loss:CachedMultipleNegativesRankingLoss
Instructions to use Omartificial-Intelligence-Space/Qwen3-VL-Embedding-2B-Arabic-VDR with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- sentence-transformers
How to use Omartificial-Intelligence-Space/Qwen3-VL-Embedding-2B-Arabic-VDR with sentence-transformers:
from sentence_transformers import SentenceTransformer model = SentenceTransformer("Omartificial-Intelligence-Space/Qwen3-VL-Embedding-2B-Arabic-VDR") sentences = [ "بناءً على ما يظهر في الصورة، كيف يمكن تفسير تكيف هذا الطائر مع بيئته الصخرية والصحراوية؟", "Fauna", "Flora", "Oman" ] embeddings = model.encode(sentences) similarities = model.similarity(embeddings, embeddings) print(similarities.shape) # [4, 4] - Notebooks
- Google Colab
- Kaggle
Upload finetuned Qwen3-VL-Embedding-2B Arabic VDR (1 epoch Pearl-vdr-ar, dev NDCG@10=0.109, test NDCG@10=0.115)
c7cd538 verified Download tokenizer_config.json from Omartificial-Intelligence-Space/Qwen3-VL-Embedding-2B-Arabic-VDR: direct link, hf CLI and curl.
- Browser
- Download file 652 Bytes
-
https://huggingface.co/Omartificial-Intelligence-Space/Qwen3-VL-Embedding-2B-Arabic-VDR/resolve/main/tokenizer_config.json
- Command line
-
hf download hf://Omartificial-Intelligence-Space/Qwen3-VL-Embedding-2B-Arabic-VDR/tokenizer_config.json
-
curl -L -o tokenizer_config.json https://huggingface.co/Omartificial-Intelligence-Space/Qwen3-VL-Embedding-2B-Arabic-VDR/resolve/main/tokenizer_config.json
652 Bytes
| { | |
| "add_prefix_space": false, | |
| "backend": "tokenizers", | |
| "bos_token": null, | |
| "clean_up_tokenization_spaces": false, | |
| "eos_token": "<|im_end|>", | |
| "errors": "replace", | |
| "is_local": true, | |
| "max_length": 262144, | |
| "max_pixels": 360000, | |
| "min_pixels": 784, | |
| "model_max_length": 262144, | |
| "pad_to_multiple_of": null, | |
| "pad_token": "<|endoftext|>", | |
| "pad_token_type_id": 0, | |
| "padding_side": "right", | |
| "processor_class": "Qwen3VLProcessor", | |
| "split_special_tokens": false, | |
| "stride": 0, | |
| "tokenizer_class": "Qwen2Tokenizer", | |
| "truncation_side": "right", | |
| "truncation_strategy": "longest_first", | |
| "unk_token": null | |
| } | |