metadata

language:
  - en
license: llama2
tags:
  - facebook
  - meta
  - pytorch
  - llama
  - llama-2
model_name: Llama 2 13B Chat
base_model: meta-llama/Llama-2-13b-chat-hf
inference: false
model_creator: Meta Llama 2
model_type: llama
pipeline_tag: text-generation
quantized_by: Second State Inc.

Llama-2-13B-Chat-GGUF

Original Model

meta-llama/Llama-2-13b-chat-hf

Run with LlamaEdge

LlamaEdge version: v0.2.8 and above

Prompt template

Prompt type: llama-2-chat

Prompt string

<s>[INST] <<SYS>>
{{ system_prompt }}
<</SYS>>

{{ user_msg_1 }} [/INST] {{ model_answer_1 }} </s><s>[INST] {{ user_msg_2 }}   [/INST]

Context size: 5120

Run as LlamaEdge service

wasmedge --dir .:. --nn-preload default:GGML:AUTO:Llama-2-13b-chat-hf-Q5_K_M.gguf \
  llama-api-server.wasm \
  --prompt-template llama-2-chat \
  --ctx-size 5120 \
  --model-name llama-2-13b-chat

Run as LlamaEdge command app

wasmedge --dir .:. --nn-preload default:GGML:AUTO:Llama-2-13b-chat-hf-Q5_K_M.gguf \
  llama-chat.wasm \
  --prompt-template llama-2-chat \
  --ctx-size 5120

Quantized GGUF Models

Name	Quant method	Bits	Size	Use case
Llama-2-13b-chat-hf-Q2_K.gguf	Q2_K	2	5.43 GB	smallest, significant quality loss - not recommended for most purposes
Llama-2-13b-chat-hf-Q3_K_L.gguf	Q3_K_L	3	6.93 GB	small, substantial quality loss
Llama-2-13b-chat-hf-Q3_K_M.gguf	Q3_K_M	3	6.34 GB	very small, high quality loss
Llama-2-13b-chat-hf-Q3_K_S.gguf	Q3_K_S	3	5.66 GB	very small, high quality loss
Llama-2-13b-chat-hf-Q4_0.gguf	Q4_0	4	7.37 GB	legacy; small, very high quality loss - prefer using Q3_K_M
Llama-2-13b-chat-hf-Q4_K_M.gguf	Q4_K_M	4	7.87 GB	medium, balanced quality - recommended
Llama-2-13b-chat-hf-Q4_K_S.gguf	Q4_K_S	4	7.41 GB	small, greater quality loss
Llama-2-13b-chat-hf-Q5_0.gguf	Q5_0	5	8.97 GB	legacy; medium, balanced quality - prefer using Q4_K_M
Llama-2-13b-chat-hf-Q5_K_M.gguf	Q5_K_M	5	9.23 GB	large, very low quality loss - recommended
Llama-2-7b-chat-hf-Q5_K_S.gguf	Q5_K_S	5	8.97 GB	large, low quality loss - recommended
Llama-2-13b-chat-hf-Q6_K.gguf	Q6_K	6	10.7 GB	very large, extremely low quality loss
Llama-2-13b-chat-hf-Q8_0.gguf	Q8_0	8	13.8 GB	very large, extremely low quality loss - not recommended
Llama-2-13b-chat-hf-f16.gguf	f16	16	26 GB