|
|
--- |
|
|
base_model: openai/gpt-oss-20b |
|
|
license: apache-2.0 |
|
|
model_creator: openai |
|
|
model_name: gpt-oss-20b |
|
|
quantized_by: Second State Inc. |
|
|
pipeline_tag: text-generation |
|
|
library_name: transformers |
|
|
--- |
|
|
|
|
|
<!-- 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 --> |
|
|
|
|
|
# gpt-oss-20b-GGUF |
|
|
|
|
|
## Original Model |
|
|
|
|
|
[openai/gpt-oss-20b](https://huggingface.co/openai/gpt-oss-20b) |
|
|
|
|
|
## Run with LlamaEdge |
|
|
|
|
|
- LlamaEdge version: [v0.25.0](https://github.com/LlamaEdge/LlamaEdge/releases/tag/0.25.0) and above ([0.25.1](https://github.com/LlamaEdge/LlamaEdge/releases/tag/0.25.1)+ with tool call support) |
|
|
|
|
|
- Prompt template |
|
|
|
|
|
- Prompt type: `gpt-oss` |
|
|
|
|
|
- Prompt string |
|
|
|
|
|
```text |
|
|
<|start|>system<|message|> |
|
|
You are ChatGPT, a large language model trained by OpenAI. |
|
|
Knowledge cutoff: 2024-06 |
|
|
Current date: 2025-08-06 |
|
|
Reasoning: medium |
|
|
# Valid channels: analysis, commentary, final. Channel must be included for every message. |
|
|
<|end|> |
|
|
|
|
|
<|start|>user<|message|>Hello!<|end|> |
|
|
<|start|>assistant<|channel|>final<|message|>Hi there!<|end|> |
|
|
<|start|>user<|message|>What's your favorite color?<|end|> |
|
|
<|start|>assistant |
|
|
``` |
|
|
|
|
|
- Context size: `128000` |
|
|
|
|
|
- Run as LlamaEdge service |
|
|
|
|
|
```bash |
|
|
wasmedge --dir .:. --nn-preload default:GGML:AUTO:gpt-oss-20b-MXFP4_MOE.gguf \ |
|
|
llama-api-server.wasm \ |
|
|
--model-name gpt-oss-20b \ |
|
|
--prompt-template gpt-oss \ |
|
|
--ctx-size 128000 |
|
|
``` |
|
|
|
|
|
<!-- ## Quantized GGUF Models |
|
|
|
|
|
| Name | Quant method | Bits | Size | Use case | |
|
|
| ---- | ---- | ---- | ---- | ----- | |
|
|
| [gpt-oss-20b-Q2_K.gguf](https://huggingface.co/second-state/gpt-oss-20b-GGUF/blob/main/gpt-oss-20b-Q2_K.gguf) | Q2_K | 2 | 968 MB| smallest, significant quality loss - not recommended for most purposes | |
|
|
| [gpt-oss-20b-Q3_K_L.gguf](https://huggingface.co/second-state/gpt-oss-20b-GGUF/blob/main/gpt-oss-20b-Q3_K_L.gguf) | Q3_K_L | 3 | 971 MB| small, substantial quality loss | |
|
|
| [gpt-oss-20b-Q3_K_M.gguf](https://huggingface.co/second-state/gpt-oss-20b-GGUF/blob/main/gpt-oss-20b-Q3_K_M.gguf) | Q3_K_M | 3 | 970 MB| very small, high quality loss | |
|
|
| [gpt-oss-20b-Q3_K_S.gguf](https://huggingface.co/second-state/gpt-oss-20b-GGUF/blob/main/gpt-oss-20b-Q3_K_S.gguf) | Q3_K_S | 3 | 968 MB| very small, high quality loss | |
|
|
| [gpt-oss-20b-Q4_0.gguf](https://huggingface.co/second-state/gpt-oss-20b-GGUF/blob/main/gpt-oss-20b-Q4_0.gguf) | Q4_0 | 4 | 969 MB| legacy; small, very high quality loss - prefer using Q3_K_M | |
|
|
| [gpt-oss-20b-Q4_K_M.gguf](https://huggingface.co/second-state/gpt-oss-20b-GGUF/blob/main/gpt-oss-20b-Q4_K_M.gguf) | Q4_K_M | 4 | 1.04 GB| medium, balanced quality - recommended | |
|
|
| [gpt-oss-20b-Q4_K_S.gguf](https://huggingface.co/second-state/gpt-oss-20b-GGUF/blob/main/gpt-oss-20b-Q4_K_S.gguf) | Q4_K_S | 4 | 1.04 GB| small, greater quality loss | |
|
|
| [gpt-oss-20b-Q5_0.gguf](https://huggingface.co/second-state/gpt-oss-20b-GGUF/blob/main/gpt-oss-20b-Q5_0.gguf) | Q5_0 | 5 | 1.05 GB| legacy; medium, balanced quality - prefer using Q4_K_M | |
|
|
| [gpt-oss-20b-Q5_K_M.gguf](https://huggingface.co/second-state/gpt-oss-20b-GGUF/blob/main/gpt-oss-20b-Q5_K_M.gguf) | Q5_K_M | 5 | 1.08 GB| large, very low quality loss - recommended | |
|
|
| [gpt-oss-20b-Q5_K_S.gguf](https://huggingface.co/second-state/gpt-oss-20b-GGUF/blob/main/gpt-oss-20b-Q5_K_S.gguf) | Q5_K_S | 5 | 1.08 GB| large, low quality loss - recommended | |
|
|
| [gpt-oss-20b-Q6_K.gguf](https://huggingface.co/second-state/gpt-oss-20b-GGUF/blob/main/gpt-oss-20b-Q6_K.gguf) | Q6_K | 6 | 1.27 GB| very large, extremely low quality loss | |
|
|
| [gpt-oss-20b-Q8_0.gguf](https://huggingface.co/second-state/gpt-oss-20b-GGUF/blob/main/gpt-oss-20b-Q8_0.gguf) | Q8_0 | 8 | 1.27 GB| very large, extremely low quality loss - not recommended | |
|
|
| [gpt-oss-20b-f16.gguf](https://huggingface.co/second-state/gpt-oss-20b-GGUF/blob/main/gpt-oss-20b-f16.gguf) | f16 | 16 | 13.8 GB| | --> |
|
|
|
|
|
*Quantized with llama.cpp b6115* |
|
|
|
