Views
No views yet
Qwen3-Reranker-8B and an OpenAI-style rerank API service for Apple Silicon.| Item | Value |
|---|---|
| Base model | Qwen/Qwen3-Reranker-8B |
| Task | Text reranking |
| Profiles | b1_s128 |
| Bundle path | bundles/qwen3_reranker_ane_bundle_8b |
| Default model id | qwen3-reranker-8b-ane |
| Package size (approx.) | 14G |
POST /rerank and POST /v1/rerank.1./setup_venv.sh
2./run_server.shcurl -s http://127.0.0.1:8000/health1curl -s http://127.0.0.1:8000/v1/rerank \
2 -H 'Content-Type: application/json' \
3 -d '{
4 "query": "capital of China",
5 "documents": [
6 "The capital of China is Beijing.",
7 "Gravity is a force."
8 ],
9 "top_n": 2,
10 "return_documents": true
11 }'s128) for low-power deployment.cpu_and_ne (ANE-preferred, not ANE-guaranteed).Qwen3-Reranker-8B 的预转换 CoreML bundle，以及可直接运行的文本重排服务（/v1/rerank）。| 项目 | 值 |
|---|---|
| 基础模型 | Qwen/Qwen3-Reranker-8B |
| 任务类型 | 文本重排 |
| Profile | b1_s128 |
| Bundle 路径 | bundles/qwen3_reranker_ane_bundle_8b |
| 默认模型名 | qwen3-reranker-8b-ane |
| 包体积（约） | 14G |
POST /rerank 与 POST /v1/rerank。1./setup_venv.sh
2./run_server.shcurl -s http://127.0.0.1:8000/health1curl -s http://127.0.0.1:8000/v1/rerank \
2 -H 'Content-Type: application/json' \
3 -d '{
4 "query": "capital of China",
5 "documents": [
6 "The capital of China is Beijing.",
7 "Gravity is a force."
8 ],
9 "top_n": 2,
10 "return_documents": true
11 }'s128），偏向低功耗部署。cpu_and_ne，是偏向 ANE 调度，不等于 100% 仅 ANE 执行。