metadata
base_model: Qwen/Qwen1.5-14B-Chat
license: other
license_name: tongyi-qianwen-research
license_link: https://huggingface.co/Qwen/Qwen1.5-14B-Chat/blob/main/LICENSE
model_creator: Qwen
model_name: Qwen1.5 14B Chat
quantized_by: Second State Inc.
language:
- en
pipeline_tag: text-generation
tags:
- chat
Qwen1.5-14B-Chat-GGUF
Original Model
Run with LlamaEdge
LlamaEdge version: v0.2.15 and above
Prompt template
Prompt type:
chatml
Prompt string
<|im_start|>system {system_message}<|im_end|> <|im_start|>user {prompt}<|im_end|> <|im_start|>assistant
Context size:
32000
Run as LlamaEdge service
wasmedge --dir .:. --nn-preload default:GGML:AUTO:Qwen1.5-14B-Chat-Q5_K_M.gguf llama-api-server.wasm -p chatml
Run as LlamaEdge command app
wasmedge --dir .:. --nn-preload default:GGML:AUTO:Qwen1.5-14B-Chat-Q5_K_M.gguf llama-chat.wasm -p chatml
Quantized GGUF Models
Name | Quant method | Bits | Size | Use case |
---|---|---|---|---|
Qwen1.5-14B-Chat-Q2_K.gguf | Q2_K | 2 | 6.09 GB | smallest, significant quality loss - not recommended for most purposes |
Qwen1.5-14B-Chat-Q3_K_L.gguf | Q3_K_L | 3 | 7.84 GB | small, substantial quality loss |
Qwen1.5-14B-Chat-Q3_K_M.gguf | Q3_K_M | 3 | 7.42 GB | very small, high quality loss |
Qwen1.5-14B-Chat-Q3_K_S.gguf | Q3_K_S | 3 | 6.95 GB | very small, high quality loss |
Qwen1.5-14B-Chat-Q4_0.gguf | Q4_0 | 4 | 8.18 GB | legacy; small, very high quality loss - prefer using Q3_K_M |
Qwen1.5-14B-Chat-Q4_K_M.gguf | Q4_K_M | 4 | 9.19 GB | medium, balanced quality - recommended |
Qwen1.5-14B-Chat-Q4_K_S.gguf | Q4_K_S | 4 | 8.56 GB | small, greater quality loss |
Qwen1.5-14B-Chat-Q5_0.gguf | Q5_0 | 5 | 9.85 GB | legacy; medium, balanced quality - prefer using Q4_K_M |
Qwen1.5-14B-Chat-Q5_K_M.gguf | Q5_K_M | 5 | 10.5 GB | large, very low quality loss - recommended |
Qwen1.5-14B-Chat-Q5_K_S.gguf | Q5_K_S | 5 | 10.0 GB | large, low quality loss - recommended |
Qwen1.5-14B-Chat-Q6_K.gguf | Q6_K | 6 | 12.3 GB | very large, extremely low quality loss |
Qwen1.5-14B-Chat-Q8_0.gguf | Q8_0 | 8 | 15.1 GB | very large, extremely low quality loss - not recommended |