Feature Extraction
Transformers
Safetensors
qwen3_5
image-text-to-text
multimodal-embedding
retrieval
evaluation
custom_code
Instructions to use woooooz/WeMM-Embedding-9B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use woooooz/WeMM-Embedding-9B with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("feature-extraction", model="woooooz/WeMM-Embedding-9B", trust_remote_code=True)# pip install -U transformers accelerate # Load model directly from transformers import AutoProcessor, AutoModelForMultimodalLM processor = AutoProcessor.from_pretrained("woooooz/WeMM-Embedding-9B", trust_remote_code=True) model = AutoModelForMultimodalLM.from_pretrained("woooooz/WeMM-Embedding-9B", trust_remote_code=True, device_map="auto") - Notebooks
- Google Colab
- Kaggle
Download evaluation/run_eval.sh from woooooz/WeMM-Embedding-9B: direct link, hf CLI and curl.
- Browser
- Download file 1.1 kB
-
https://huggingface.co/woooooz/WeMM-Embedding-9B/resolve/main/evaluation/run_eval.sh
- Command line
-
hf download hf://woooooz/WeMM-Embedding-9B/evaluation/run_eval.sh
-
curl -L -o run_eval.sh https://huggingface.co/woooooz/WeMM-Embedding-9B/resolve/main/evaluation/run_eval.sh
1.1 kB
| set -euo pipefail | |
| HERE="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)" | |
| MODEL="${MODEL_PATH:-$(dirname "$HERE")}" | |
| CONFIG="${1:-image_retrieval}" | |
| case "$CONFIG" in image_retrieval|video_ret_5|moment_retrieval|visdoc_full) ;; *) echo "Unknown config: $CONFIG" >&2; exit 2;; esac | |
| : "${DATA_BASEDIR:?Set DATA_BASEDIR to the MMEB evaluation media root}" | |
| export PYTHONPATH="$HERE/vendor:$HERE/nexus:${PYTHONPATH:-}" | |
| export TOKENIZERS_PARALLELISM=false | |
| export NEXUS_WEMM_OFFICIAL=1 | |
| OUTPUT="${OUTPUT_DIR:-$PWD/eval_output/$CONFIG}" | |
| mkdir -p "$OUTPUT" "$OUTPUT/generated_media_cache" | |
| python -m torch.distributed.run --nproc_per_node="${NPROC:-1}" --master_port="${MASTER_PORT:-29521}" \ | |
| -m Nexus.evaluation.mmeb_v2.eval_embedding \ | |
| --normalize true --per_device_eval_batch_size "${BATCH_SIZE:-8}" \ | |
| --model_name_or_path "$MODEL" --dataset_config "$HERE/configs/$CONFIG.yaml" \ | |
| --encode_output_path "$OUTPUT" --data_basedir "$DATA_BASEDIR" \ | |
| --generated_media_basedir "${GENERATED_MEDIA_BASEDIR:-$OUTPUT/generated_media_cache}" \ | |
| --force_recompute "${FORCE_RECOMPUTE:-true}" | |