|
--- |
|
license: other |
|
license_name: tongyi-qianwen-license-agreement |
|
license_link: >- |
|
https://github.com/QwenLM/Qwen/blob/main/Tongyi%20Qianwen%20LICENSE%20AGREEMENT |
|
pipeline_tag: text-generation |
|
--- |
|
GGUF importance matrix (imatrix) quants for https://huggingface.co/abacusai/Smaug-72B-v0.1 |
|
The importance matrix was trained for 100K tokens (200 batches of 512 tokens) using wiki.train.raw. |
|
|
|
**Update 2024-03-14**: |
|
* New quant IQ1_S using latest commit `4755afd1`. |
|
|
|
**Update 2024-03-02**: |
|
* New quants IQ2_S/IQ2_M, requires commit [a33e6a0d](https://github.com/ggerganov/llama.cpp/commit/a33e6a0d2a66104ea9a906bdbf8a94d050189d91) or later. |
|
* The importance matrix was trained for ~50K tokens (105 batches of 512 tokens) using a [general purpose imatrix calibration dataset](https://github.com/ggerganov/llama.cpp/discussions/5263#discussioncomment-8395384). |
|
* This is a different calibration dataset than the previous quants I posted so we can compare the quality |
|
|
|
Llama-2 conversation template and system prompt set to the [Qwen system prompt](https://github.com/QwenLM/Qwen/blob/main/examples/system_prompt.md). |
|
|
|
| Layers | Context | Template | |
|
| --- | --- | --- | |
|
| <pre>80</pre> | <pre>32768</pre> | <pre>[INST] \<\<SYS\>\><br>{instructions}<br>\<\</SYS\>\><br><br>{prompt} [/INST]<br>{response}</pre> | |
