[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M7D2XBXAN469KZWXM8PDWJS8":3,"skill-same:01M7D2XBXAN469KZWXM8PDWJS8":65},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":15,"categories":16,"keywords":17,"publishedAt":18,"linkOnly":19,"markdownContent":20,"markdownBody":21,"frontMatter":22,"refName":39,"commitSha":40,"skillDirectory":41,"forksCount":13,"repositoryHtmlUrl":42,"repositoryHomepageUrl":6,"ownerHtmlUrl":43,"ownerBlogUrl":6,"ownerDescription":6,"updatedAt":44,"install":45,"resources":58},"01M7D2XBXAN469KZWXM8PDWJS8","sentence-transformers",null,"Framework for state-of-the-art sentence, text, and image embeddings. Provides 5000+ pre-trained models for semantic similarity, clustering, and retrieval. Supports multilingual, domain-specific, and multimodal models. Use for generating embeddings for RAG, semantic search, or similarity tasks. Best for production embedding generation.","beyondai\u002Fsandbox","beyondai","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F2845958?v=4","01M7D24SBNM47BJGJ9J2HB13BP","01M7D24SCDC9DYE005XH6R553S",0,"MIT","https:\u002F\u002Fgithub.com\u002Fbeyondai\u002Fsandbox\u002Fblob\u002Fmain\u002Flabs\u002F.codex\u002Fskills\u002Fsentence-transformers\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: sentence-transformers\ndescription: Framework for state-of-the-art sentence, text, and image embeddings. Provides 5000+ pre-trained models for semantic similarity, clustering, and retrieval. Supports multilingual, domain-specific, and multimodal models. Use for generating embeddings for RAG, semantic search, or similarity tasks. Best for production embedding generation.\nversion: 1.0.0\nauthor: Orchestra Research\nlicense: MIT\ntags: [Sentence Transformers, Embeddings, Semantic Similarity, RAG, Multilingual, Multimodal, Pre-Trained Models, Clustering, Semantic Search, Production]\ndependencies: [sentence-transformers, transformers, torch]\n---\n\n# Sentence Transformers - State-of-the-Art Embeddings\n\nPython framework for sentence and text embeddings using transformers.\n\n## When to use Sentence Transformers\n\n**Use when:**\n- Need high-quality embeddings for RAG\n- Semantic similarity and search\n- Text clustering and classification\n- Multilingual embeddings (100+ languages)\n- Running embeddings locally (no API)\n- Cost-effective alternative to OpenAI embeddings\n\n**Metrics**:\n- **15,700+ GitHub stars**\n- **5000+ pre-trained models**\n- **100+ languages** supported\n- Based on PyTorch\u002FTransformers\n\n**Use alternatives instead**:\n- **OpenAI Embeddings**: Need API-based, highest quality\n- **Instructor**: Task-specific instructions\n- **Cohere Embed**: Managed service\n\n## Quick start\n\n### Installation\n\n```bash\npip install sentence-transformers\n```\n\n### Basic usage\n\n```python\nfrom sentence_transformers import SentenceTransformer\n\n# Load model\nmodel = SentenceTransformer('all-MiniLM-L6-v2')\n\n# Generate embeddings\nsentences = [\n    \"This is an example sentence\",\n    \"Each sentence is converted to a vector\"\n]\n\nembeddings = model.encode(sentences)\nprint(embeddings.shape)  # (2, 384)\n\n# Cosine similarity\nfrom sentence_transformers.util import cos_sim\nsimilarity = cos_sim(embeddings[0], embeddings[1])\nprint(f\"Similarity: {similarity.item():.4f}\")\n```\n\n## Popular models\n\n### General purpose\n\n```python\n# Fast, good quality (384 dim)\nmodel = SentenceTransformer('all-MiniLM-L6-v2')\n\n# Better quality (768 dim)\nmodel = SentenceTransformer('all-mpnet-base-v2')\n\n# Best quality (1024 dim, slower)\nmodel = SentenceTransformer('all-roberta-large-v1')\n```\n\n### Multilingual\n\n```python\n# 50+ languages\nmodel = SentenceTransformer('paraphrase-multilingual-MiniLM-L12-v2')\n\n# 100+ languages\nmodel = SentenceTransformer('paraphrase-multilingual-mpnet-base-v2')\n```\n\n### Domain-specific\n\n```python\n# Legal domain\nmodel = SentenceTransformer('nlpaueb\u002Flegal-bert-base-uncased')\n\n# Scientific papers\nmodel = SentenceTransformer('allenai\u002Fspecter')\n\n# Code\nmodel = SentenceTransformer('microsoft\u002Fcodebert-base')\n```\n\n## Semantic search\n\n```python\nfrom sentence_transformers import SentenceTransformer, util\n\nmodel = SentenceTransformer('all-MiniLM-L6-v2')\n\n# Corpus\ncorpus = [\n    \"Python is a programming language\",\n    \"Machine learning uses algorithms\",\n    \"Neural networks are powerful\"\n]\n\n# Encode corpus\ncorpus_embeddings = model.encode(corpus, convert_to_tensor=True)\n\n# Query\nquery = \"What is Python?\"\nquery_embedding = model.encode(query, convert_to_tensor=True)\n\n# Find most similar\nhits = util.semantic_search(query_embedding, corpus_embeddings, top_k=3)\nprint(hits)\n```\n\n## Similarity computation\n\n```python\n# Cosine similarity\nsimilarity = util.cos_sim(embedding1, embedding2)\n\n# Dot product\nsimilarity = util.dot_score(embedding1, embedding2)\n\n# Pairwise cosine similarity\nsimilarities = util.cos_sim(embeddings, embeddings)\n```\n\n## Batch encoding\n\n```python\n# Efficient batch processing\nsentences = [\"sentence 1\", \"sentence 2\", ...] * 1000\n\nembeddings = model.encode(\n    sentences,\n    batch_size=32,\n    show_progress_bar=True,\n    convert_to_tensor=False  # or True for PyTorch tensors\n)\n```\n\n## Fine-tuning\n\n```python\nfrom sentence_transformers import InputExample, losses\nfrom torch.utils.data import DataLoader\n\n# Training data\ntrain_examples = [\n    InputExample(texts=['sentence 1', 'sentence 2'], label=0.8),\n    InputExample(texts=['sentence 3', 'sentence 4'], label=0.3),\n]\n\ntrain_dataloader = DataLoader(train_examples, batch_size=16)\n\n# Loss function\ntrain_loss = losses.CosineSimilarityLoss(model)\n\n# Train\nmodel.fit(\n    train_objectives=[(train_dataloader, train_loss)],\n    epochs=10,\n    warmup_steps=100\n)\n\n# Save\nmodel.save('my-finetuned-model')\n```\n\n## LangChain integration\n\n```python\nfrom langchain_community.embeddings import HuggingFaceEmbeddings\n\nembeddings = HuggingFaceEmbeddings(\n    model_name=\"sentence-transformers\u002Fall-mpnet-base-v2\"\n)\n\n# Use with vector stores\nfrom langchain_chroma import Chroma\n\nvectorstore = Chroma.from_documents(\n    documents=docs,\n    embedding=embeddings\n)\n```\n\n## LlamaIndex integration\n\n```python\nfrom llama_index.embeddings.huggingface import HuggingFaceEmbedding\n\nembed_model = HuggingFaceEmbedding(\n    model_name=\"sentence-transformers\u002Fall-mpnet-base-v2\"\n)\n\nfrom llama_index.core import Settings\nSettings.embed_model = embed_model\n\n# Use in index\nindex = VectorStoreIndex.from_documents(documents)\n```\n\n## Model selection guide\n\n| Model | Dimensions | Speed | Quality | Use Case |\n|-------|------------|-------|---------|----------|\n| all-MiniLM-L6-v2 | 384 | Fast | Good | General, prototyping |\n| all-mpnet-base-v2 | 768 | Medium | Better | Production RAG |\n| all-roberta-large-v1 | 1024 | Slow | Best | High accuracy needed |\n| paraphrase-multilingual | 768 | Medium | Good | Multilingual |\n\n## Best practices\n\n1. **Start with all-MiniLM-L6-v2** - Good baseline\n2. **Normalize embeddings** - Better for cosine similarity\n3. **Use GPU if available** - 10× faster encoding\n4. **Batch encoding** - More efficient\n5. **Cache embeddings** - Expensive to recompute\n6. **Fine-tune for domain** - Improves quality\n7. **Test different models** - Quality varies by task\n8. **Monitor memory** - Large models need more RAM\n\n## Performance\n\n| Model | Speed (sentences\u002Fsec) | Memory | Dimension |\n|-------|----------------------|---------|-----------|\n| MiniLM | ~2000 | 120MB | 384 |\n| MPNet | ~600 | 420MB | 768 |\n| RoBERTa | ~300 | 1.3GB | 1024 |\n\n## Resources\n\n- **GitHub**: https:\u002F\u002Fgithub.com\u002FUKPLab\u002Fsentence-transformers ⭐ 15,700+\n- **Models**: https:\u002F\u002Fhuggingface.co\u002Fsentence-transformers\n- **Docs**: https:\u002F\u002Fwww.sbert.net\n- **License**: Apache 2.0\n\n\n","# Sentence Transformers - State-of-the-Art Embeddings\n\nPython framework for sentence and text embeddings using transformers.\n\n## When to use Sentence Transformers\n\n**Use when:**\n- Need high-quality embeddings for RAG\n- Semantic similarity and search\n- Text clustering and classification\n- Multilingual embeddings (100+ languages)\n- Running embeddings locally (no API)\n- Cost-effective alternative to OpenAI embeddings\n\n**Metrics**:\n- **15,700+ GitHub stars**\n- **5000+ pre-trained models**\n- **100+ languages** supported\n- Based on PyTorch\u002FTransformers\n\n**Use alternatives instead**:\n- **OpenAI Embeddings**: Need API-based, highest quality\n- **Instructor**: Task-specific instructions\n- **Cohere Embed**: Managed service\n\n## Quick start\n\n### Installation\n\n```bash\npip install sentence-transformers\n```\n\n### Basic usage\n\n```python\nfrom sentence_transformers import SentenceTransformer\n\n# Load model\nmodel = SentenceTransformer('all-MiniLM-L6-v2')\n\n# Generate embeddings\nsentences = [\n    \"This is an example sentence\",\n    \"Each sentence is converted to a vector\"\n]\n\nembeddings = model.encode(sentences)\nprint(embeddings.shape)  # (2, 384)\n\n# Cosine similarity\nfrom sentence_transformers.util import cos_sim\nsimilarity = cos_sim(embeddings[0], embeddings[1])\nprint(f\"Similarity: {similarity.item():.4f}\")\n```\n\n## Popular models\n\n### General purpose\n\n```python\n# Fast, good quality (384 dim)\nmodel = SentenceTransformer('all-MiniLM-L6-v2')\n\n# Better quality (768 dim)\nmodel = SentenceTransformer('all-mpnet-base-v2')\n\n# Best quality (1024 dim, slower)\nmodel = SentenceTransformer('all-roberta-large-v1')\n```\n\n### Multilingual\n\n```python\n# 50+ languages\nmodel = SentenceTransformer('paraphrase-multilingual-MiniLM-L12-v2')\n\n# 100+ languages\nmodel = SentenceTransformer('paraphrase-multilingual-mpnet-base-v2')\n```\n\n### Domain-specific\n\n```python\n# Legal domain\nmodel = SentenceTransformer('nlpaueb\u002Flegal-bert-base-uncased')\n\n# Scientific papers\nmodel = SentenceTransformer('allenai\u002Fspecter')\n\n# Code\nmodel = SentenceTransformer('microsoft\u002Fcodebert-base')\n```\n\n## Semantic search\n\n```python\nfrom sentence_transformers import SentenceTransformer, util\n\nmodel = SentenceTransformer('all-MiniLM-L6-v2')\n\n# Corpus\ncorpus = [\n    \"Python is a programming language\",\n    \"Machine learning uses algorithms\",\n    \"Neural networks are powerful\"\n]\n\n# Encode corpus\ncorpus_embeddings = model.encode(corpus, convert_to_tensor=True)\n\n# Query\nquery = \"What is Python?\"\nquery_embedding = model.encode(query, convert_to_tensor=True)\n\n# Find most similar\nhits = util.semantic_search(query_embedding, corpus_embeddings, top_k=3)\nprint(hits)\n```\n\n## Similarity computation\n\n```python\n# Cosine similarity\nsimilarity = util.cos_sim(embedding1, embedding2)\n\n# Dot product\nsimilarity = util.dot_score(embedding1, embedding2)\n\n# Pairwise cosine similarity\nsimilarities = util.cos_sim(embeddings, embeddings)\n```\n\n## Batch encoding\n\n```python\n# Efficient batch processing\nsentences = [\"sentence 1\", \"sentence 2\", ...] * 1000\n\nembeddings = model.encode(\n    sentences,\n    batch_size=32,\n    show_progress_bar=True,\n    convert_to_tensor=False  # or True for PyTorch tensors\n)\n```\n\n## Fine-tuning\n\n```python\nfrom sentence_transformers import InputExample, losses\nfrom torch.utils.data import DataLoader\n\n# Training data\ntrain_examples = [\n    InputExample(texts=['sentence 1', 'sentence 2'], label=0.8),\n    InputExample(texts=['sentence 3', 'sentence 4'], label=0.3),\n]\n\ntrain_dataloader = DataLoader(train_examples, batch_size=16)\n\n# Loss function\ntrain_loss = losses.CosineSimilarityLoss(model)\n\n# Train\nmodel.fit(\n    train_objectives=[(train_dataloader, train_loss)],\n    epochs=10,\n    warmup_steps=100\n)\n\n# Save\nmodel.save('my-finetuned-model')\n```\n\n## LangChain integration\n\n```python\nfrom langchain_community.embeddings import HuggingFaceEmbeddings\n\nembeddings = HuggingFaceEmbeddings(\n    model_name=\"sentence-transformers\u002Fall-mpnet-base-v2\"\n)\n\n# Use with vector stores\nfrom langchain_chroma import Chroma\n\nvectorstore = Chroma.from_documents(\n    documents=docs,\n    embedding=embeddings\n)\n```\n\n## LlamaIndex integration\n\n```python\nfrom llama_index.embeddings.huggingface import HuggingFaceEmbedding\n\nembed_model = HuggingFaceEmbedding(\n    model_name=\"sentence-transformers\u002Fall-mpnet-base-v2\"\n)\n\nfrom llama_index.core import Settings\nSettings.embed_model = embed_model\n\n# Use in index\nindex = VectorStoreIndex.from_documents(documents)\n```\n\n## Model selection guide\n\n| Model | Dimensions | Speed | Quality | Use Case |\n|-------|------------|-------|---------|----------|\n| all-MiniLM-L6-v2 | 384 | Fast | Good | General, prototyping |\n| all-mpnet-base-v2 | 768 | Medium | Better | Production RAG |\n| all-roberta-large-v1 | 1024 | Slow | Best | High accuracy needed |\n| paraphrase-multilingual | 768 | Medium | Good | Multilingual |\n\n## Best practices\n\n1. **Start with all-MiniLM-L6-v2** - Good baseline\n2. **Normalize embeddings** - Better for cosine similarity\n3. **Use GPU if available** - 10× faster encoding\n4. **Batch encoding** - More efficient\n5. **Cache embeddings** - Expensive to recompute\n6. **Fine-tune for domain** - Improves quality\n7. **Test different models** - Quality varies by task\n8. **Monitor memory** - Large models need more RAM\n\n## Performance\n\n| Model | Speed (sentences\u002Fsec) | Memory | Dimension |\n|-------|----------------------|---------|-----------|\n| MiniLM | ~2000 | 120MB | 384 |\n| MPNet | ~600 | 420MB | 768 |\n| RoBERTa | ~300 | 1.3GB | 1024 |\n\n## Resources\n\n- **GitHub**: https:\u002F\u002Fgithub.com\u002FUKPLab\u002Fsentence-transformers ⭐ 15,700+\n- **Models**: https:\u002F\u002Fhuggingface.co\u002Fsentence-transformers\n- **Docs**: https:\u002F\u002Fwww.sbert.net\n- **License**: Apache 2.0\n\n\n",{"name":5,"tags":23,"author":34,"license":14,"version":35,"description":7,"dependencies":36},[24,25,26,27,28,29,30,31,32,33],"Sentence Transformers","Embeddings","Semantic Similarity","RAG","Multilingual","Multimodal","Pre-Trained Models","Clustering","Semantic Search","Production","Orchestra Research","1.0.0",[5,37,38],"transformers","torch","main","46e03791269608c5ac886a64c08f5d94147e8786","labs\u002F.codex\u002Fskills\u002Fsentence-transformers","https:\u002F\u002Fgithub.com\u002Fbeyondai\u002Fsandbox","https:\u002F\u002Fgithub.com\u002Fbeyondai","2026-11-08 04:18:42",{"source":8,"skillName":5,"commands":46,"prompt":56,"zipAvailable":57},[47,50,53],{"tool":48,"command":49},"npx","npx skills add beyondai\u002Fsandbox --skill sentence-transformers",{"tool":51,"command":52},"bunx","bunx skills add beyondai\u002Fsandbox --skill sentence-transformers",{"tool":54,"command":55},"pnpm","pnpm dlx skills add beyondai\u002Fsandbox --skill sentence-transformers","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fbeyondai\u002Fsandbox\u002Fblob\u002Fmain\u002Flabs\u002F.codex\u002Fskills\u002Fsentence-transformers\u002FSKILL.md",true,[59],{"path":60,"resourceType":61,"resourceTypeDesc":62,"sizeBytes":63,"sourceHtmlUrl":64},"references\u002Fmodels.md","REFERENCE","参考文档",2925,"https:\u002F\u002Fgithub.com\u002Fbeyondai\u002Fsandbox\u002Fblob\u002Fmain\u002Flabs\u002F.codex\u002Fskills\u002Fsentence-transformers\u002Freferences\u002Fmodels.md",[66,73,80,87,94,101],{"skillNo":67,"name":68,"nameZh":6,"description":69,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":70,"categories":71,"keywords":72,"publishedAt":18},"01M7D3GMPGJZRG8EPY404YF8TR","weights-and-biases","Track ML experiments with automatic logging, visualize training in real-time, optimize hyperparameters with sweeps, and manage model registry with W&B - collaborative MLOps platform","https:\u002F\u002Fgithub.com\u002Fbeyondai\u002Fsandbox\u002Fblob\u002Fmain\u002Flabs\u002F.opencode\u002Fskills\u002Fweights-and-biases\u002FSKILL.md",[],[],{"skillNo":74,"name":75,"nameZh":6,"description":76,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":77,"categories":78,"keywords":79,"publishedAt":18},"01M7D3GFZH440V2J532BSZMPVP","serving-llms-vllm","Serves LLMs with high throughput using vLLM's PagedAttention and continuous batching. Use when deploying production LLM APIs, optimizing inference latency\u002Fthroughput, or serving models with limited GPU memory. Supports OpenAI-compatible endpoints, quantization (GPTQ\u002FAWQ\u002FFP8), and tensor parallelism.","https:\u002F\u002Fgithub.com\u002Fbeyondai\u002Fsandbox\u002Fblob\u002Fmain\u002Flabs\u002F.opencode\u002Fskills\u002Fvllm\u002FSKILL.md",[],[],{"skillNo":81,"name":82,"nameZh":6,"description":83,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":84,"categories":85,"keywords":86,"publishedAt":18},"01M7D3GAE8MQCPGF72EP9G9GKY","verl-rl-training","Provides guidance for training LLMs with reinforcement learning using verl (Volcano Engine RL). Use when implementing RLHF, GRPO, PPO, or other RL algorithms for LLM post-training at scale with flexible infrastructure backends.","https:\u002F\u002Fgithub.com\u002Fbeyondai\u002Fsandbox\u002Fblob\u002Fmain\u002Flabs\u002F.opencode\u002Fskills\u002Fverl\u002FSKILL.md",[],[],{"skillNo":88,"name":89,"nameZh":6,"description":90,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":91,"categories":92,"keywords":93,"publishedAt":18},"01M7D3G7KKNE5RGNYJNXV4F0MM","unsloth","Expert guidance for fast fine-tuning with Unsloth - 2-5x faster training, 50-80% less memory, LoRA\u002FQLoRA optimization","https:\u002F\u002Fgithub.com\u002Fbeyondai\u002Fsandbox\u002Fblob\u002Fmain\u002Flabs\u002F.opencode\u002Fskills\u002Funsloth\u002FSKILL.md",[],[],{"skillNo":95,"name":96,"nameZh":6,"description":97,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":98,"categories":99,"keywords":100,"publishedAt":18},"01M7D3G2RCZQ08GS0KAW5CBWSC","fine-tuning-with-trl","Fine-tune LLMs using reinforcement learning with TRL - SFT for instruction tuning, DPO for preference alignment, PPO\u002FGRPO for reward optimization, and reward model training. Use when need RLHF, align model with preferences, or train from human feedback. Works with HuggingFace Transformers.","https:\u002F\u002Fgithub.com\u002Fbeyondai\u002Fsandbox\u002Fblob\u002Fmain\u002Flabs\u002F.opencode\u002Fskills\u002Ftrl-fine-tuning\u002FSKILL.md",[],[],{"skillNo":102,"name":103,"nameZh":6,"description":104,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":105,"categories":106,"keywords":107,"publishedAt":18},"01M7D3FX08BPB172CYGM597NZF","torchforge-rl-training","Provides guidance for PyTorch-native agentic RL using torchforge, Meta's library separating infra from algorithms. Use when you want clean RL abstractions, easy algorithm experimentation, or scalable training with Monarch and TorchTitan.","https:\u002F\u002Fgithub.com\u002Fbeyondai\u002Fsandbox\u002Fblob\u002Fmain\u002Flabs\u002F.opencode\u002Fskills\u002Ftorchforge\u002FSKILL.md",[],[]]