bartowski commited on
Commit
708d712
·
verified ·
1 Parent(s): eef6c70

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +5 -1
README.md CHANGED
@@ -12,16 +12,20 @@ Using <a href="https://github.com/ggerganov/llama.cpp/">llama.cpp</a> release <a
12
 
13
  Original model: https://huggingface.co/deepseek-ai/DeepSeek-R1-0528
14
 
 
 
15
  All quants made using imatrix option with dataset from [here](https://gist.github.com/bartowski1182/eb213dccb3571f863da82e99418f81e8)
16
 
17
  Run them in [LM Studio](https://lmstudio.ai/)
18
 
19
  Run them directly with [llama.cpp](https://github.com/ggerganov/llama.cpp), or any other llama.cpp based project
20
 
 
 
21
  ## Prompt format
22
 
23
  ```
24
- <|begin▁of▁sentence|>{system_prompt}<|User|>{prompt}<|Assistant|><|end▁of▁sentence|><|Assistant|>
25
  ```
26
 
27
  ## Download a file (not the whole branch) from below:
 
12
 
13
  Original model: https://huggingface.co/deepseek-ai/DeepSeek-R1-0528
14
 
15
+ Huge shoutout to [Artus](https://huggingface.co/ArtusDev) who helped with both conversion and imatrix calculation!
16
+
17
  All quants made using imatrix option with dataset from [here](https://gist.github.com/bartowski1182/eb213dccb3571f863da82e99418f81e8)
18
 
19
  Run them in [LM Studio](https://lmstudio.ai/)
20
 
21
  Run them directly with [llama.cpp](https://github.com/ggerganov/llama.cpp), or any other llama.cpp based project
22
 
23
+ These quants were made with the changes from this PR for improved performance: https://github.com/ggml-org/llama.cpp/pull/12727
24
+
25
  ## Prompt format
26
 
27
  ```
28
+ <|begin▁of▁sentence|>{system_prompt}<|User|>{prompt}<|Assistant|>
29
  ```
30
 
31
  ## Download a file (not the whole branch) from below: