AI-Master-Book
  • about AI-Master-Book
  • AI Master Book
    • 이상치 탐지 with Python
    • 베이지안 뉴럴네트워크 (BNN) with Python
    • 그래프 뉴럴네트워크 (GNN) with Python
    • 데이터 마케팅 분석 with Python
  • LLM MASTER BOOK
    • OpenAI API 쿡북 with Python
    • 기초부터 심화까지 RAG 쿡북 with Python
  • LLMs
    • OpenAI API
      • 1️⃣ChatCompletion
      • 2️⃣DALL-E
      • 3️⃣Text to Speech
      • 4️⃣Video to Transcripts
      • 5️⃣Assistants API
      • 6️⃣Prompt Engineering
      • 7️⃣OpenAI New GPT-4o
    • LangChain
      • LangChain Basic
        • 1️⃣Basic Modules
        • 2️⃣Model I/O
        • 3️⃣Prompts
        • 4️⃣Chains
        • 5️⃣Agents
        • 6️⃣Tools
        • 7️⃣Memory
      • LangChain Intermediate
        • 1️⃣OpenAI LLM
        • 2️⃣Prompt Template
        • 3️⃣Retrieval
        • 4️⃣RAG ChatBot
        • 5️⃣RAG with Gemini
        • 6️⃣New Huggingface-LangChain
        • 7️⃣Huggingface Hub
        • 8️⃣SQL Agent & Chain
        • 9️⃣Expression Language(LCEL)
        • 🔟Llama3-8B with LangChain
      • LangChain Advanced
        • 1️⃣LLM Evaluation
        • 2️⃣RAG Evaluation with RAGAS
        • 3️⃣LangChain with RAGAS
        • 4️⃣RAG Paradigms
        • 5️⃣LangChain: Advance Techniques
        • 6️⃣LangChain with NeMo-Guardrails
        • 7️⃣LangChain vs. LlamaIndex
        • 8️⃣LangChain LCEL vs. LangGraph
    • LlamaIndex
      • LlamaIndex Basic
        • 1️⃣Introduction
        • 2️⃣Customization
        • 3️⃣Data Connectors
        • 4️⃣Documents & Nodes
        • 5️⃣Naive RAG
        • 6️⃣Advanced RAG
        • 7️⃣Llama3-8B with LlamaIndex
        • 8️⃣LlmaPack
      • LlamaIndex Intermediate
        • 1️⃣QueryEngine
        • 2️⃣Agent
        • 3️⃣Evaluation
        • 4️⃣Evaluation-Driven Development
        • 5️⃣Fine-tuning
        • 6️⃣Prompt Compression with LLMLingua
      • LlamaIndex Advanced
        • 1️⃣Agentic RAG: Router Engine
        • 2️⃣Agentic RAG: Tool Calling
        • 3️⃣Building Agent Reasoning Loop
        • 4️⃣Building Multi-document Agent
    • Hugging Face
      • Huggingface Basic
        • 1️⃣Datasets
        • 2️⃣Tokenizer
        • 3️⃣Sentence Embeddings
        • 4️⃣Transformers
        • 5️⃣Sentence Transformers
        • 6️⃣Evaluate
        • 7️⃣Diffusers
      • Huggingface Tasks
        • NLP
          • 1️⃣Sentiment Analysis
          • 2️⃣Zero-shot Classification
          • 3️⃣Aspect-Based Sentiment Analysis
          • 4️⃣Feature Extraction
          • 5️⃣Intent Classification
          • 6️⃣Topic Modeling: BERTopic
          • 7️⃣NER: Token Classification
          • 8️⃣Summarization
          • 9️⃣Translation
          • 🔟Text Generation
        • Audio & Tabular
          • 1️⃣Text-to-Speech: TTS
          • 2️⃣Speech Recognition: Whisper
          • 3️⃣Audio Classification
          • 4️⃣Tabular Qustaion & Answering
        • Vision & Multimodal
          • 1️⃣Image-to-Text
          • 2️⃣Text to Image
          • 3️⃣Image to Image
          • 4️⃣Text or Image-to-Video
          • 5️⃣Depth Estimation
          • 6️⃣Image Classification
          • 7️⃣Object Detection
          • 8️⃣Segmentatio
      • Huggingface Optimization
        • 1️⃣Accelerator
        • 2️⃣Bitsandbytes
        • 3️⃣Flash Attention
        • 4️⃣Quantization
        • 5️⃣Safetensors
        • 6️⃣Optimum-ONNX
        • 7️⃣Optimum-NVIDIA
        • 8️⃣Optimum-Intel
      • Huggingface Fine-tuning
        • 1️⃣Transformer Fine-tuning
        • 2️⃣PEFT Fine-tuning
        • 3️⃣PEFT: Fine-tuning with QLoRA
        • 4️⃣PEFT: Fine-tuning Phi-2 with QLoRA
        • 5️⃣Axoltl Fine-tuning with QLoRA
        • 6️⃣TRL: RLHF Alignment Fine-tuning
        • 7️⃣TRL: DPO Fine-tuning with Phi-3-4k-instruct
        • 8️⃣TRL: ORPO Fine-tuning with Llama3-8B
        • 9️⃣Convert GGUF gemma-2b with llama.cpp
        • 🔟Apple Silicon Fine-tuning Gemma-2B with MLX
        • 🔢LLM Mergekit
    • Agentic LLM
      • Agentic LLM
        • 1️⃣Basic Agentic LLM
        • 2️⃣Multi-agent with CrewAI
        • 3️⃣LangGraph: Multi-agent Basic
        • 4️⃣LangGraph: Agentic RAG with LangChain
        • 5️⃣LangGraph: Agentic RAG with Llama3-8B by Groq
      • Autonomous Agent
        • 1️⃣LLM Autonomous Agent?
        • 2️⃣AutoGPT: Worldcup Winner Search with LangChain
        • 3️⃣BabyAGI: Weather Report with LangChain
        • 4️⃣AutoGen: Writing Blog Post with LangChain
        • 5️⃣LangChain: Autonomous-agent Debates with Tools
        • 6️⃣CAMEL Role-playing Autonomous Cooperative Agents
        • 7️⃣LangChain: Two-player Harry Potter D&D based CAMEL
        • 8️⃣LangChain: Multi-agent Bid for K-Pop Debate
        • 9️⃣LangChain: Multi-agent Authoritarian Speaker Selection
        • 🔟LangChain: Multi-Agent Simulated Environment with PettingZoo
    • Multimodal
      • 1️⃣PaliGemma: Open Vision LLM
      • 2️⃣FLUX.1: Generative Image
    • Building LLM
      • 1️⃣DSPy
      • 2️⃣DSPy RAG
      • 3️⃣DSPy with LangChain
      • 4️⃣Mamba
      • 5️⃣Mamba RAG with LangChain
      • 7️⃣PostgreSQL VectorDB with pgvorco.rs
Powered by GitBook
On this page
  1. LLMs
  2. Hugging Face
  3. Huggingface Tasks
  4. NLP

Feature Extraction

Feature Extraction

NLP에서 특징 추출은 원시 텍스트 데이터를 기계 학습 모델에서 사용할 수 있는 수치 표현으로 변환하는 프로세스를 말합니다.

텐서의 모양은 입력 텍스트의 토큰 수와 BERT 모델에서 숨겨진 레이어의 크기에 따라 달라집니다.

"bert-base-uncased" 모델의 경우 각 토큰은 768차원 벡터로 표현되므로 결과의 모양은 (1, number_of_tokens, 768)이 되며, 여기서 1은 배치 크기가 1임을 나타냅니다(단일 문장을 처리)

from transformers import pipeline

extractor = pipeline(
    model="google-bert/bert-base-uncased", 
    task="feature-extraction"
)
result = extractor("This is a simple test.", return_tensors=True)
result.shape
> torch.Size([1, 8, 768])
from transformers import pipeline

checkpoint = "facebook/bart-base"
feature_extractor = pipeline(
    "feature-extraction",
    framework="pt",
    model=checkpoint
)
text = "Transformers is an awesome library!"

#Reducing along the first dimension to get a 768 dimensional array
result = feature_extractor(text,return_tensors = "pt")[0].numpy().mean(axis=0)
result
array([ 2.65979469e-01, -6.56846821e-01,  1.25627622e-01,  1.36601961e+00,
       -7.59062350e-01,  7.56494999e-01, -1.05783379e+00, -1.87183488e-02,
       -6.71836197e-01, -1.96252692e+00,  2.81317592e-01,  1.35964262e+00,
       -7.21477158e-03,  9.12114024e-01,  3.43531281e-01, -1.35358238e+00,
       -7.05923915e-01,  1.40108848e+00,  9.22238588e-01, -1.89351216e-01,
        2.40267381e-01, -2.19243467e-01,  9.30557787e-01,  7.81986415e-01,
       -1.01491988e+00, -5.85219324e-01,  6.38441265e-01, -4.86978149e+00,
       -1.68945134e-01, -3.85040236e+00, -9.25553560e-01, -1.71881378e-01,
        5.59025332e-02, -6.87882125e-01, -1.03691256e+00, -3.37629229e-01,
        4.45500523e-01,  2.23247707e-01,  5.60621440e-01,  4.59313184e-01,
        2.38988829e+00,  6.36417508e-01, -8.32767725e-01,  3.60407889e-01,
       -2.03047842e-01,  9.03682828e-01, -1.65837502e+00,  1.19581902e+00,
       -2.48081341e-01,  8.12374890e-01, -4.88460273e-01,  2.99310595e-01,
        5.61737835e-01, -3.31205308e-01, -1.28257060e+00, -4.96809697e-03,
        7.20308483e-01, -4.77443188e-02,  4.61459041e-01,  1.31555617e+00,
       -2.52127767e-01, -2.13813365e-01, -8.04339498e-02, -7.61193752e-01,
       -5.84884942e-01, -7.94288933e-01, -7.32236505e-01,  1.45702612e+00,
        7.83099055e-01, -4.38804209e-01,  3.54020238e-01,  7.75779635e-02,
        1.90865564e+00, -1.15279332e-01, -2.89797974e+00,  4.35574591e-01,
        1.95469415e+00, -1.57004988e+00, -6.98929250e-01,  1.97130308e-01,
        2.16544896e-01,  1.79927900e-01, -4.03966725e-01, -5.33977710e-02,
        4.32771295e-02, -6.61527812e-01, -7.12189525e-02, -4.96350408e-01,
        3.23695511e-01,  3.63479644e-01, -7.13343740e-01,  9.46864963e-01,
       -4.25351888e-01, -1.62067080e+00,  4.80425507e-01,  6.41202092e-01,
        2.59259790e-01,  5.41801155e-01, -3.12652916e-01, -7.57046163e-01,
        1.49013770e+00,  2.58979249e+00, -1.38797617e+00,  6.22176707e-01,
       -4.93953610e-03,  1.16229892e+00,  1.27965415e+00, -7.59564459e-01,
        4.44990993e-01, -1.19959974e+00, -2.21594334e-01, -1.01445842e+00,
       -1.89106047e-01, -3.31210196e-01,  8.83097172e-01, -1.40546516e-01,
        3.58635008e-01, -4.30313945e-01,  1.20075715e+00, -3.19993407e-01,
        1.29259288e+00,  1.60403013e+00, -1.70512187e+00,  4.31367129e-01,
       -4.54189837e-01,  1.58637106e-01, -7.68502951e-01,  5.44382691e-01,
       -7.79063642e-01,  7.52042174e-01,  3.63594383e-01,  1.50957608e+00,
        4.73330081e-01,  5.57976484e-01,  3.91978323e-01, -1.21539712e+00,
       -1.84798136e-01,  8.70272458e-01, -1.24984539e+00, -3.81815940e-01,
       -2.21528202e-01, -2.32662141e-01,  1.01794183e+00,  4.10634369e-01,
       -2.76572406e-01, -7.22168162e-02,  7.32802749e-02, -4.83183920e-01,
        2.79749595e-02, -1.90762311e-01, -2.02779293e+00, -1.51625490e-02,
        2.34252289e-01,  1.20170951e+00,  5.16299963e-01, -1.41489655e-01,
       -2.46584147e-01,  1.91934600e-01, -2.89843321e-01,  7.42309749e-01,
       -2.88151741e-01, -8.30338299e-01,  1.08718443e+00,  7.55699515e-01,
        1.44116759e-01, -1.84054241e-01, -9.06795204e-01,  2.49311715e-01,
       -3.42900723e-01,  8.78563583e-01,  3.04132570e-02,  1.34575880e+00,
       -1.85381448e+00,  6.08434156e-02, -1.42206645e+00,  6.67507231e-01,
        1.34545970e+00,  4.45354342e-01,  2.74102747e-01, -1.42620504e+00,
        1.66471791e+00,  4.83929425e-01, -3.44172597e-01,  7.31653988e-01,
       -1.34243101e-01,  3.96722883e-01,  6.22375488e-01,  3.31292927e-01,
       -3.20087582e-01, -1.07533097e+00,  9.76715624e-01, -3.52599472e-01,
       -1.31508529e-01, -1.68829501e-01, -1.38619637e+00,  8.90181839e-01,
       -5.07539250e-02,  3.07513736e-02, -3.93885255e-01, -1.71657515e+00,
        1.08871162e+00,  4.53436106e-01,  1.21618068e+00,  5.08168757e-01,
       -6.95172369e-01, -9.72706795e-01, -6.98203564e-01, -3.77047509e-02,
        5.14743686e-01,  7.26660311e-01,  1.42454967e-01,  1.70566046e+00,
        2.11760625e-01,  7.90104866e-01, -1.11895156e+00,  9.30911675e-02,
       -8.53542924e-01, -3.47614557e-01, -8.34994614e-01, -2.87239719e-02,
        1.74926448e+00, -1.14972591e+00,  7.42217958e-01, -5.98051906e-01,
        4.17294323e-01,  9.28575099e-02, -1.65214360e+00,  2.74253279e-01,
       -6.29156351e-01, -6.69885874e-01, -6.18473053e-01,  4.31621671e-01,
       -5.20275354e-01,  1.36028194e+00,  1.14785850e+00, -7.18702853e-01,
        7.51556270e-03,  1.28168547e+00,  4.58477229e-01, -6.67250633e-01,
       -8.71611178e-01,  1.54485416e+00,  9.50916409e-01,  1.56043792e+00,
        4.03211296e-01,  1.39242363e+00, -1.03513181e-01,  5.22949100e-01,
        6.63481355e-01, -2.78556813e-02, -7.29202271e-01, -3.53717595e-01,
       -1.51166081e+00, -1.01856375e+00,  6.35557413e-01, -1.74585164e+00,
        1.15104437e+00,  1.59911370e+00,  1.51997781e+00,  3.44798148e-01,
       -1.03017783e+00, -3.45427930e-01, -3.04957658e-01, -1.52942911e-01,
       -5.44561386e-01,  2.72866637e-01, -4.63098109e-01,  1.88960946e+00,
       -6.39785528e-01,  4.92546350e-01,  3.46445173e-01,  1.83566287e-01,
        3.37440491e-01,  8.80383372e-01, -1.57478404e+00,  2.41358012e-01,
        3.42918366e-01, -1.77611008e-01, -1.44021642e+00, -4.05111313e-01,
        9.36335549e-02,  1.14355230e+00,  5.38416505e-01, -1.01677418e+00,
       -6.91158831e-01,  1.24062963e-01,  2.08523095e-01,  9.39111337e-02,
        9.54437554e-01, -9.73225534e-02,  1.23226786e+00,  1.23787582e+00,
       -3.61190140e-01,  9.04121161e-01,  1.90456226e-01,  3.66956741e-01,
       -8.37930024e-01,  5.74374676e-01, -1.30378753e-01,  1.15551674e+00,
        7.32960701e-01, -6.51617289e-01,  1.14175498e+00,  4.53374743e-01,
       -1.02572456e-01,  1.62613416e+00,  2.79911518e-01, -9.57020462e-01,
       -1.28427362e+00, -3.79492283e-01,  1.21321309e+00, -6.52288198e-01,
       -7.16957331e-01,  8.21545362e-01, -2.78386503e-01,  1.13116992e+00,
        1.09877765e+00, -9.30897295e-01,  3.44983310e-01,  8.14800978e-01,
       -4.39907372e-01,  3.38642240e-01, -4.09136683e-01,  1.71418175e-01,
        5.85774332e-02,  6.26304924e-01,  1.14245892e+00, -7.30417669e-02,
        4.70405728e-01,  2.00712815e-01, -1.10264778e+00,  6.69935524e-01,
        1.26470280e+00,  9.37951952e-02, -9.04536903e-01,  7.71072134e-02,
        6.09872103e-01,  4.51007843e-01,  9.30153728e-01, -3.91749442e-01,
        6.32270098e-01,  2.91940570e-01,  6.22446001e-01,  5.41503489e-01,
        4.00291502e-01,  5.99478722e-01,  8.12314451e-02,  8.11395943e-02,
        4.41665947e-01,  1.10141015e+00, -9.73566175e-01, -1.45525181e+00,
        5.64943314e-01, -2.38152429e-01,  1.41102523e-01, -7.86105156e-01,
        8.86056185e-01, -7.10839093e-01,  3.27000290e-01,  5.97459912e-01,
       -1.95455417e-01, -5.74278355e-01,  1.28140464e-01,  8.05716038e-01,
       -4.99953538e-01,  3.17015976e-01,  3.11146140e-01,  2.31078458e+00,
       -6.16076440e-02, -4.31919515e-01, -1.45730332e-01,  6.16032660e-01,
       -9.90341902e-01, -8.43507424e-02, -8.87243569e-01, -3.43032390e-01,
        1.29097724e+00, -6.21537507e-01,  5.30034304e-01,  2.00960255e+00,
       -3.07885379e-01, -3.07402015e-01, -2.99035162e-01,  1.58422029e+00,
        6.66770160e-01,  9.46137309e-01,  1.00991011e+00, -4.43267673e-01,
       -1.53554392e+00, -3.11964750e-01, -9.86443758e-02,  2.98534557e-02,
       -2.61718333e-01, -2.58908939e+00, -4.89793241e-01,  2.49610424e+00,
        5.08717000e-01,  6.50168598e-01,  3.08177888e-01,  4.27572727e-01,
        2.46583670e-01, -3.61945689e-01,  3.36508095e-01,  1.83386886e+00,
       -3.30805369e-02,  4.69550401e-01, -2.56731361e-01, -5.89862108e-01,
        1.82226884e+00,  4.07911837e-01, -5.95187187e-01,  8.44984949e-02,
        2.93200195e-01,  2.97689795e+00,  1.81511497e+00,  1.09478426e+00,
        6.32766724e-01, -3.60258728e-01, -1.17961836e+00, -6.41382575e-01,
       -1.11094289e-01, -3.20806690e-02,  6.67168975e-01, -3.80334437e-01,
        2.60429978e-01,  2.25518560e+00,  6.14036143e-01, -1.76677987e-01,
       -1.77043125e-01,  2.15038013e+00, -7.47117698e-01, -6.94940984e-01,
       -8.47300291e-01, -8.48803222e-02,  8.57784092e-01,  5.92671931e-01,
        8.59628022e-01, -2.30567765e+00, -7.14714766e-01,  1.16440272e+00,
       -1.58269092e-01,  1.78712055e-01,  6.23534322e-01,  5.20011425e-01,
        1.34862280e+00,  4.30682451e-01,  1.31071520e+00,  7.05770731e-01,
        8.79335642e-01,  1.04244995e+00, -4.27514404e-01,  1.63319424e-01,
       -1.31811261e+00,  3.38772506e-01,  6.16423368e-01,  9.51805353e-01,
       -4.15812321e-02,  1.17271984e+00, -1.08082747e+00, -3.87872487e-01,
       -6.65958166e-01, -8.24503243e-01, -1.04688489e+00,  3.82272780e-01,
       -4.10087854e-01,  6.21690869e-01, -3.57830405e-01, -2.88853496e-01,
       -3.67524803e-01, -6.23959303e-01, -1.07978083e-01, -7.72217512e-01,
       -3.48409325e-01,  5.99236190e-01, -3.06323647e-01,  6.41631603e-01,
        1.45477855e+00, -2.87821621e-01,  1.15087187e+00, -8.66089165e-01,
        1.25644231e+00,  1.78619817e-01, -3.97542268e-01,  1.23447955e+00,
        3.76234233e-01,  2.45014504e-01,  1.80363189e-02, -1.10603869e+00,
        8.92204285e-01,  6.63499773e-01,  1.25622559e+00, -1.29876032e-01,
       -5.97272754e-01, -3.75236988e-01,  4.66275960e-01,  2.05121726e-01,
        8.55793536e-01, -4.02110481e+00,  1.44663119e+00,  2.47043297e-01,
        6.41006827e-01, -2.76829362e-01,  9.26169932e-01,  4.13330823e-01,
        3.42508614e-01,  1.37350821e+00, -1.03768575e+00,  2.95568883e-01,
        7.84581363e-01, -6.02763034e-02, -1.78127453e-01, -8.99840117e-01,
        1.44706023e+00,  1.71805096e+00,  1.68698475e-01, -5.69088042e-01,
       -1.05945960e-01, -2.80161768e-01, -6.83287263e-01, -1.20706224e+00,
       -6.09945178e-01, -1.30536163e+00, -2.01481447e-01, -2.13613780e-03,
        8.87428403e-01, -7.25608051e-01,  1.03411388e+00,  6.20733440e-01,
       -2.92526424e-01, -7.22831935e-02,  2.99865872e-01, -9.95002985e-01,
        9.87491369e-01,  2.62370020e-01,  3.95637065e-01, -8.62838387e-01,
       -3.04033637e-01,  3.50857317e-01, -4.46875505e-02,  9.52754319e-02,
       -1.81237090e+00,  1.28286946e+00, -1.01697290e+00,  3.33020598e-01,
        5.44873066e-02,  8.81249309e-02, -8.30183446e-01, -1.25741291e+00,
        1.13769807e-01,  4.17068511e-01, -3.84530395e-01, -2.24328697e-01,
        1.64285719e-01,  2.24695787e-01, -1.52362788e+00, -1.03049445e+00,
        5.07936776e-01,  6.00863695e-01,  6.78084612e-01,  5.73505223e-01,
        9.01764214e-01,  2.14455634e-01, -8.64829645e-02,  5.21133542e-01,
       -3.84391618e+00, -4.32377815e-01,  3.86084586e-01, -5.74579954e-01,
       -6.64480180e-02,  5.68239272e-01,  1.40972987e-01,  4.91561055e-01,
       -1.59076011e+00,  4.58789200e-01, -3.81277895e+00, -3.56776267e-01,
       -7.03040540e-01,  4.01638448e-01, -4.14998889e-01,  9.21632528e-01,
       -2.88249195e-01,  5.43611228e-01, -1.87583610e-01,  1.77437171e-01,
        3.09019029e-01, -8.21537554e-01,  9.64059532e-01, -4.18440819e-01,
       -3.74807954e-01, -7.94256628e-01,  1.79041013e-01,  1.07198226e+00,
        7.37931907e-01, -5.71660399e-01,  4.73279476e-01, -1.12489057e+00,
       -8.20027113e-01,  1.16821051e+00,  1.13403749e+00, -8.37848783e-01,
       -6.94696903e-01, -5.87071300e-01, -9.37979996e-01,  2.85793412e-02,
       -1.03942998e-01,  8.15052748e-01,  2.77216405e-01,  1.07841063e+00,
        8.49105597e-01, -1.22114134e+00,  1.31575480e-01,  4.65513058e-02,
       -9.81023669e-01, -1.33689547e+00,  9.66923177e-01, -1.29471350e+00,
        1.38094172e-01,  6.99724734e-01, -3.32960755e-01,  1.33537698e+00,
       -3.91558230e-01,  5.63477397e-01, -2.24221781e-01, -2.44365379e-01,
       -1.46850929e-01, -1.11731577e+00, -6.03061169e-02, -4.82764870e-01,
        9.59276438e-01,  2.19666290e+00, -9.22762871e-01,  3.80298674e-01,
        3.07932347e-01,  1.24999022e+00,  3.42894405e-01,  9.13683604e-03,
       -2.37314433e-01, -5.78179881e-02, -1.28482133e-01,  8.15542936e-01,
        4.25670408e-02,  1.63982916e+00, -2.73039818e-01, -1.85594022e-01,
       -8.31672192e-01,  9.44418490e-01, -8.35879385e-01,  8.57529119e-02,
       -2.79763341e-01,  9.06917334e-01,  3.24907333e-01,  3.79474759e-02,
        1.75888743e-02, -3.66290390e-01, -5.95124885e-02,  6.87758029e-01,
        5.07672548e-01,  8.89262438e-01,  1.69256181e-01, -2.01039672e-01,
        5.85848987e-01,  1.67553282e+00,  4.36515212e-01, -3.62415373e-01,
        8.22778121e-02,  4.75326419e-01,  1.03972077e+00,  1.17408001e+00,
        1.08113861e+00,  1.17782259e+00,  7.19923496e-01, -6.69488251e-01,
        1.36961073e-01,  1.56454515e+00, -4.26551163e-01, -9.29266691e-01,
        7.18483865e-01,  3.78707618e-01,  7.51538515e-01,  9.56296325e-02,
        2.21472085e-01,  2.28069518e-02,  5.18463790e-01,  1.02798796e+00,
        2.11316586e-01,  1.61214125e+00,  7.54863799e-01, -9.27332759e-01,
       -1.86410949e-01, -1.12251091e+00,  3.68848145e-01,  4.18064177e-01,
        7.76560783e-01,  2.79443599e-02, -3.13859940e-01, -4.47741836e-01,
        4.84260559e-01,  1.09082341e+00,  1.55656290e+00, -3.40694696e-01,
        4.41435456e-01,  7.73885310e-01,  6.59892440e-01, -1.66957736e+00,
       -1.25799298e+00,  7.65725374e-01,  7.07073510e-01, -1.03319263e+00,
        5.52317381e-01,  3.94287586e-01,  3.92046362e-01, -1.49054527e-01,
       -5.83782732e-01, -5.09692848e-01,  5.62452197e-01,  6.57410741e-01,
       -5.00763655e-01,  4.95452195e-01,  3.58396247e-02, -7.79511034e-01,
        4.42655563e-01, -2.28039667e-01,  8.00651312e-02,  1.79112375e+00,
        3.67924958e-01,  1.30259705e+00,  1.01184118e+00, -8.26271296e-01,
       -1.88019410e-01,  1.58395958e+00, -5.49807787e-01,  5.25783479e-01,
       -8.63637924e-01, -5.48161387e-01,  4.79667455e-01,  4.72317845e-01,
       -3.08872998e-01,  8.75118449e-02, -8.59509230e-01,  7.50889957e-01,
       -2.64737934e-01,  7.62561083e-01, -4.49841231e-01,  1.45644456e-01,
        1.01997578e+00,  3.77051145e-01, -2.72834718e-01,  4.99737054e-01,
       -1.25988948e+00, -5.44752657e-01,  2.32418612e-01,  9.05423999e-01,
       -5.48643887e-01,  7.33481586e-01, -5.10832906e-01,  1.30848140e-01,
        7.81549454e-01, -4.22700197e-01, -3.84838343e-01,  1.11054122e+00,
       -9.99024928e-01,  1.36228776e+00, -6.92071795e-01, -3.62365395e-01,
       -2.13862792e-01,  7.48088360e-01, -6.61671400e-01,  3.47315609e-01],
      dtype=float32)

  • feature_extractor(text, return_tensors="pt": 입력 텍스트를 특징 추출 파이프라인으로 전달하고 결과를 PyTorch 텐서로 반환하도록 지정합니다. [0]: 입력 텍스트의 특징 표현에 해당하는 출력의 첫 번째 요소를 선택합니다.

  • .numpy(): 파이토치 텐서를 NumPy 배열로 변환합니다.

  • .mean(axis=0): 시퀀스 차원(축 0)에 걸쳐 특징 벡터의 평균을 계산합니다. 이는 모든 토큰의 특징을 768차원의 단일 벡터로 집계하는 방법입니다(BART 베이스의 숨겨진 크기가 768이므로).

PreviousAspect-Based Sentiment AnalysisNextIntent Classification

Last updated 11 months ago

4️⃣