Update README.md

README.md

@@ -2,11 +2,18 @@
 license: other
 license_name: tongyi-qianwen
 license_link: https://huggingface.co/Qwen/Qwen1.5-72B-Chat/blob/main/LICENSE
-library_name: gguf
 pipeline_tag: text-generation
+library_name: gguf
+base_model: Qwen/Qwen1.5-72B-Chat
 ---
-GGUF importance matrix (imatrix) quants for https://huggingface.co/Qwen/Qwen1.5-72B-Chat
+<u>**NOTE**</u>: You will need a recent build of llama.cpp to run these quants (i.e. at least commit `494c870`).
+
+**2023-03-07**: Updating quants using latest build as things seem to have stabilized a bit now.
+
+GGUF importance matrix (imatrix) quants for https://huggingface.co/Qwen/Qwen1.5-72B-Chat  
+* 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).
+* The [imatrix is being used on the K-quants](https://github.com/ggerganov/llama.cpp/pull/4930) as well.
 
-| Layers | Context | Template |
+| Layers | Context | [Template](https://huggingface.co/Qwen/Qwen1.5-72B-Chat/blob/main/tokenizer_config.json#L31) |
 | --- | --- | --- |
-| <pre>80</pre> | <pre>32768</pre> | <pre><\|im_start\|>system<br>{instructions}<\|im_end\|><br><\|im_start\|>user<br>{prompt}<\|im_end\|><br><\|im_start\|>assistant<br>{response}</pre> |
+| <pre>80</pre> | <pre>32768</pre> | <pre><\|im_start\|>system<br>{instructions}<\|im_end\|><br><\|im_start\|>user<br>{prompt}<\|im_end\|><br><\|im_start\|>assistant<br>{response}</pre> |