f6a540f05b
Model: second-state/ELYZA-japanese-Llama-2-13b-fast-instruct-GGUF Source: Original Platform
77 lines
4.8 KiB
Markdown
77 lines
4.8 KiB
Markdown
---
|
|
base_model: elyza/ELYZA-japanese-Llama-2-13b-fast-instruct
|
|
inference: false
|
|
library_name: transformers
|
|
license: llama2
|
|
model_creator: elyza
|
|
model_name: ELYZA-japanese-Llama-2-13b-fast-instruct
|
|
pipeline_tag: text-generation
|
|
quantized_by: Second State Inc.
|
|
language:
|
|
- ja
|
|
- en
|
|
---
|
|
|
|
<!-- header start -->
|
|
<!-- 200823 -->
|
|
<div style="width: auto; margin-left: auto; margin-right: auto">
|
|
<img src="https://github.com/LlamaEdge/LlamaEdge/raw/dev/assets/logo.svg" style="width: 100%; min-width: 400px; display: block; margin: auto;">
|
|
</div>
|
|
<hr style="margin-top: 1.0em; margin-bottom: 1.0em;">
|
|
<!-- header end -->
|
|
|
|
# ELYZA-japanese-Llama-2-13b-fast-instruct-GGUF
|
|
|
|
## Original Model
|
|
|
|
[elyza/ELYZA-japanese-Llama-2-13b-fast-instruct](https://huggingface.co/elyza/ELYZA-japanese-Llama-2-13b-fast-instruct)
|
|
|
|
## Run with LlamaEdge
|
|
|
|
- LlamaEdge version: [v0.2.8](https://github.com/LlamaEdge/LlamaEdge/releases/tag/0.2.8) and above
|
|
|
|
- Prompt template
|
|
|
|
- Prompt type: `llama-2-chat`
|
|
|
|
- Prompt string
|
|
|
|
```text
|
|
<s>[INST] <<SYS>>
|
|
{{ system_prompt }}
|
|
<</SYS>>
|
|
|
|
{{ user_msg_1 }} [/INST] {{ model_answer_1 }} </s><s>[INST] {{ user_msg_2 }}
|
|
```
|
|
|
|
- Context size: `5120`
|
|
|
|
- Run as LlamaEdge service
|
|
|
|
```bash
|
|
wasmedge --dir .:. --nn-preload default:GGML:AUTO:ELYZA-japanese-Llama-2-13b-fast-instruct-Q5_K_M.gguf llama-api-server.wasm -p llama-2-chat
|
|
```
|
|
|
|
- Run as LlamaEdge command app
|
|
|
|
```bash
|
|
wasmedge --dir .:. --nn-preload default:GGML:AUTO:ELYZA-japanese-Llama-2-13b-fast-instruct-Q5_K_M.gguf llama-chat.wasm -p llama-2-chat
|
|
```
|
|
|
|
## Quantized GGUF Models
|
|
|
|
| Name | Quant method | Bits | Size | Use case |
|
|
| ---- | ---- | ---- | ---- | ----- |
|
|
| [ELYZA-japanese-Llama-2-13b-fast-instruct-Q2_K.gguf](https://huggingface.co/second-state/ELYZA-japanese-Llama-2-13b-fast-instruct-GGUF/blob/main/ELYZA-japanese-Llama-2-13b-fast-instruct-Q2_K.gguf) | Q2_K | 2 | 5.5 GB| smallest, significant quality loss - not recommended for most purposes |
|
|
| [ELYZA-japanese-Llama-2-13b-fast-instruct-Q3_K_L.gguf](https://huggingface.co/second-state/ELYZA-japanese-Llama-2-13b-fast-instruct-GGUF/blob/main/ELYZA-japanese-Llama-2-13b-fast-instruct-Q3_K_L.gguf) | Q3_K_L | 3 | 7.01 GB| small, substantial quality loss |
|
|
| [ELYZA-japanese-Llama-2-13b-fast-instruct-Q3_K_M.gguf](https://huggingface.co/second-state/ELYZA-japanese-Llama-2-13b-fast-instruct-GGUF/blob/main/ELYZA-japanese-Llama-2-13b-fast-instruct-Q3_K_M.gguf) | Q3_K_M | 3 | 6.42 GB| very small, high quality loss |
|
|
| [ELYZA-japanese-Llama-2-13b-fast-instruct-Q3_K_S.gguf](https://huggingface.co/second-state/ELYZA-japanese-Llama-2-13b-fast-instruct-GGUF/blob/main/ELYZA-japanese-Llama-2-13b-fast-instruct-Q3_K_S.gguf) | Q3_K_S | 3 | 5.74 GB| very small, high quality loss |
|
|
| [ELYZA-japanese-Llama-2-13b-fast-instruct-Q4_0.gguf](https://huggingface.co/second-state/ELYZA-japanese-Llama-2-13b-fast-instruct-GGUF/blob/main/ELYZA-japanese-Llama-2-13b-fast-instruct-Q4_0.gguf) | Q4_0 | 4 | 7.46 GB| legacy; small, very high quality loss - prefer using Q3_K_M |
|
|
| [ELYZA-japanese-Llama-2-13b-fast-instruct-Q4_K_M.gguf](https://huggingface.co/second-state/ELYZA-japanese-Llama-2-13b-fast-instruct-GGUF/blob/main/ELYZA-japanese-Llama-2-13b-fast-instruct-Q4_K_M.gguf) | Q4_K_M | 4 | 7.96 GB| medium, balanced quality - recommended |
|
|
| [ELYZA-japanese-Llama-2-13b-fast-instruct-Q4_K_S.gguf](https://huggingface.co/second-state/ELYZA-japanese-Llama-2-13b-fast-instruct-GGUF/blob/main/ELYZA-japanese-Llama-2-13b-fast-instruct-Q4_K_S.gguf) | Q4_K_S | 4 | 7.5 GB| small, greater quality loss |
|
|
| [ELYZA-japanese-Llama-2-13b-fast-instruct-Q5_0.gguf](https://huggingface.co/second-state/ELYZA-japanese-Llama-2-13b-fast-instruct-GGUF/blob/main/ELYZA-japanese-Llama-2-13b-fast-instruct-Q5_0.gguf) | Q5_0 | 5 | 9.07 GB| legacy; medium, balanced quality - prefer using Q4_K_M |
|
|
| [ELYZA-japanese-Llama-2-13b-fast-instruct-Q5_K_M.gguf](https://huggingface.co/second-state/ELYZA-japanese-Llama-2-13b-fast-instruct-GGUF/blob/main/ELYZA-japanese-Llama-2-13b-fast-instruct-Q5_K_M.gguf) | Q5_K_M | 5 | 9.33 GB| large, very low quality loss - recommended |
|
|
| [ELYZA-japanese-Llama-2-13b-fast-instruct-Q5_K_S.gguf](https://huggingface.co/second-state/ELYZA-japanese-Llama-2-13b-fast-instruct-GGUF/blob/main/ELYZA-japanese-Llama-2-13b-fast-instruct-Q5_K_S.gguf) | Q5_K_S | 5 | 9.07 GB| large, low quality loss - recommended |
|
|
| [ELYZA-japanese-Llama-2-13b-fast-instruct-Q6_K.gguf](https://huggingface.co/second-state/ELYZA-japanese-Llama-2-13b-fast-instruct-GGUF/blob/main/ELYZA-japanese-Llama-2-13b-fast-instruct-Q6_K.gguf) | Q6_K | 6 | 10.8 GB| very large, extremely low quality loss |
|
|
| [ELYZA-japanese-Llama-2-13b-fast-instruct-Q8_0.gguf](https://huggingface.co/second-state/ELYZA-japanese-Llama-2-13b-fast-instruct-GGUF/blob/main/ELYZA-japanese-Llama-2-13b-fast-instruct-Q8_0.gguf) | Q8_0 | 8 | 14 GB| very large, extremely low quality loss - not recommended |
|