Yz

yarn-llama2

An extension of Llama 2 that supports a context of up to 128k tokens.

7b 13b
Hızlı Kurulum (Ollama kuruluysa) 
ollama run yarn-llama2

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 3.8GB 64K Text ollama run yarn-llama2:latest
7b 3.8GB 64K Text ollama run yarn-llama2:7b
13b 7.4GB 64K Text ollama run yarn-llama2:13b
7b-128k 3.8GB 128K Text ollama run yarn-llama2:7b-128k
7b-128k-q2_K 2.8GB 128K Text ollama run yarn-llama2:7b-128k-q2_K
7b-128k-q3_K_S 2.9GB 128K Text ollama run yarn-llama2:7b-128k-q3_K_S
7b-128k-q3_K_M 3.3GB 128K Text ollama run yarn-llama2:7b-128k-q3_K_M
7b-128k-q3_K_L 3.6GB 128K Text ollama run yarn-llama2:7b-128k-q3_K_L
7b-128k-q4_0 3.8GB 128K Text ollama run yarn-llama2:7b-128k-q4_0
7b-128k-q4_1 4.2GB 128K Text ollama run yarn-llama2:7b-128k-q4_1
7b-128k-q4_K_S 3.9GB 128K Text ollama run yarn-llama2:7b-128k-q4_K_S
7b-128k-q4_K_M 4.1GB 128K Text ollama run yarn-llama2:7b-128k-q4_K_M
7b-128k-q5_0 4.7GB 128K Text ollama run yarn-llama2:7b-128k-q5_0
7b-128k-q5_1 5.1GB 128K Text ollama run yarn-llama2:7b-128k-q5_1
7b-128k-q5_K_S 4.7GB 128K Text ollama run yarn-llama2:7b-128k-q5_K_S
7b-128k-q5_K_M 4.8GB 128K Text ollama run yarn-llama2:7b-128k-q5_K_M
7b-128k-q6_K 5.5GB 128K Text ollama run yarn-llama2:7b-128k-q6_K
7b-128k-q8_0 7.2GB 128K Text ollama run yarn-llama2:7b-128k-q8_0
7b-128k-fp16 13GB 128K Text ollama run yarn-llama2:7b-128k-fp16
7b-64k 3.8GB 64K Text ollama run yarn-llama2:7b-64k
7b-64k-q2_K 2.8GB 64K Text ollama run yarn-llama2:7b-64k-q2_K
7b-64k-q3_K_S 2.9GB 64K Text ollama run yarn-llama2:7b-64k-q3_K_S
7b-64k-q3_K_M 3.3GB 64K Text ollama run yarn-llama2:7b-64k-q3_K_M
7b-64k-q3_K_L 3.6GB 64K Text ollama run yarn-llama2:7b-64k-q3_K_L
7b-64k-q4_0 3.8GB 64K Text ollama run yarn-llama2:7b-64k-q4_0

42 varyant daha var — tümü için ollama.com.

Model Detayları ve Benchmarklar (kaynak: ollama.com)

Yarn Llama 2 is a model based on Llama2 that extends its context size up to 128k context. It is developed by Nous Research by implementing the YaRN method to further train the model to support larger context windows.

CLI

64k context size:

ollama run yarn-llama2

128k context size:

ollama run yarn-llama2:7b-128k

API

Example:

curl -X POST http://localhost:11434/api/generate -d '{
  "model": "yarn-llama2:7b-128k",
  "prompt":"Here is a story about llamas eating grass"
 }'

References

Hugging Face

YaRN: Efficient Context Window Extension of Large Language Models