Yz

granite3.1-dense

The IBM Granite 2B and 8B models are text-only dense LLMs trained on over 12 trillion tokens of data, demonstrated significant improvements over their predecessors in performance and speed in IBM’s initial testing.

Araç Kullanımı 2b 8b
Hızlı Kurulum (Ollama kuruluysa) 
ollama run granite3.1-dense

Ollama kurulu değil mi? ollama.com/download — Windows, macOS ve Linux için ücretsiz. İlk çalıştırmada model indirilir, sonrası tamamen çevrimdışıdır.

Varyantlar

Boyut büyüdükçe kalite artar, donanım ihtiyacı yükselir. Başlangıç için küçük varyantı deneyin.

EtiketBoyutBağlamGirdiKomut
latest 5.0GB 128K Text ollama run granite3.1-dense:latest
2b 1.6GB 128K Text ollama run granite3.1-dense:2b
8b 5.0GB 128K Text ollama run granite3.1-dense:8b
2b-instruct-q2_K 1.0GB 128K Text ollama run granite3.1-dense:2b-instruct-q2_K
2b-instruct-q3_K_S 1.2GB 128K Text ollama run granite3.1-dense:2b-instruct-q3_K_S
2b-instruct-q3_K_M 1.3GB 128K Text ollama run granite3.1-dense:2b-instruct-q3_K_M
2b-instruct-q3_K_L 1.4GB 128K Text ollama run granite3.1-dense:2b-instruct-q3_K_L
2b-instruct-q4_0 1.5GB 128K Text ollama run granite3.1-dense:2b-instruct-q4_0
2b-instruct-q4_1 1.6GB 128K Text ollama run granite3.1-dense:2b-instruct-q4_1
2b-instruct-q4_K_S 1.5GB 128K Text ollama run granite3.1-dense:2b-instruct-q4_K_S
2b-instruct-q4_K_M 1.6GB 128K Text ollama run granite3.1-dense:2b-instruct-q4_K_M
2b-instruct-q5_0 1.8GB 128K Text ollama run granite3.1-dense:2b-instruct-q5_0
2b-instruct-q5_1 1.9GB 128K Text ollama run granite3.1-dense:2b-instruct-q5_1
2b-instruct-q5_K_S 1.8GB 128K Text ollama run granite3.1-dense:2b-instruct-q5_K_S
2b-instruct-q5_K_M 1.8GB 128K Text ollama run granite3.1-dense:2b-instruct-q5_K_M
2b-instruct-q6_K 2.1GB 128K Text ollama run granite3.1-dense:2b-instruct-q6_K
2b-instruct-q8_0 2.7GB 128K Text ollama run granite3.1-dense:2b-instruct-q8_0
2b-instruct-fp16 5.1GB 128K Text ollama run granite3.1-dense:2b-instruct-fp16
8b-instruct-q2_K 3.2GB 128K Text ollama run granite3.1-dense:8b-instruct-q2_K
8b-instruct-q3_K_S 3.6GB 128K Text ollama run granite3.1-dense:8b-instruct-q3_K_S
8b-instruct-q3_K_M 4.0GB 128K Text ollama run granite3.1-dense:8b-instruct-q3_K_M
8b-instruct-q3_K_L 4.4GB 128K Text ollama run granite3.1-dense:8b-instruct-q3_K_L
8b-instruct-q4_0 4.7GB 128K Text ollama run granite3.1-dense:8b-instruct-q4_0
8b-instruct-q4_1 5.2GB 128K Text ollama run granite3.1-dense:8b-instruct-q4_1
8b-instruct-q4_K_S 4.7GB 128K Text ollama run granite3.1-dense:8b-instruct-q4_K_S

8 varyant daha var — tümü için ollama.com.

Model Detayları ve Benchmarklar (kaynak: ollama.com)

Granite dense models

The IBM Granite 2B and 8B models are text-only dense LLMs trained on over 12 trillion tokens of data, demonstrated significant improvements over their predecessors in performance and speed in IBM’s initial testing.

They are designed to support tool-based use cases and for retrieval augmented generation (RAG), streamlining code generation, translation and bug fixing.

Parameter Sizes

2B:

ollama run granite3.1-dense:2b

8B:

ollama run granite3.1-dense:8b

Supported Languages

English, German, Spanish, French, Japanese, Portuguese, Arabic, Czech, Italian, Korean, Dutch, Chinese (Simplified)

Capabilities

  • Summarization
  • Text classification
  • Text extraction
  • Question-answering
  • Retrieval Augmented Generation (RAG)
  • Code related tasks
  • Function-calling tasks
  • Multilingual dialog use cases
  • Long-context tasks including long document/meeting summarization, long document QA, etc.

Granite mixture of experts models

The Granite mixture of experts models are available in 1B and 3B parameter sizes designed for low latency usage.

See model page

Learn more