TheBloke commited on
Commit
25ccb44
1 Parent(s): 6eb1ec0

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +20 -1
README.md CHANGED
@@ -34,6 +34,25 @@ GGML files are for CPU + GPU inference using [llama.cpp](https://github.com/gger
34
  * [2, 3, 4, 5, 6 and 8-bit GGML models for CPU+GPU inference](https://huggingface.co/TheBloke/tulu-30B-GGML)
35
  * [Unquantised fp16 model in pytorch format, for GPU inference and for further conversions](https://huggingface.co/TheBloke/tulu-30B-fp16)
36
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
37
  <!-- compatibility_ggml start -->
38
  ## Compatibility
39
 
@@ -88,7 +107,7 @@ Refer to the Provided Files table below to see what files use which methods, and
88
  I use the following command line; adjust for your tastes and needs:
89
 
90
  ```
91
- ./main -t 10 -ngl 32 -m tulu-30b.ggmlv3.q5_0.bin --color -c 2048 --temp 0.7 --repeat_penalty 1.1 -n -1 -p "### Instruction: Write a story about llamas\n### Response:"
92
  ```
93
  Change `-t 10` to the number of physical CPU cores you have. For example if your system has 8 cores/16 threads, use `-t 8`.
94
 
 
34
  * [2, 3, 4, 5, 6 and 8-bit GGML models for CPU+GPU inference](https://huggingface.co/TheBloke/tulu-30B-GGML)
35
  * [Unquantised fp16 model in pytorch format, for GPU inference and for further conversions](https://huggingface.co/TheBloke/tulu-30B-fp16)
36
 
37
+ ## Prompt template
38
+
39
+ The following template should be used:
40
+
41
+ ```
42
+ <|user|>
43
+ prompt goes here
44
+ <|assistant|>
45
+
46
+ ```
47
+
48
+ **Note**: There should be a newline after `<|assistant|>`. This appears to be very important for getting this model to respond correctly.
49
+
50
+ In other words, the prompt is:
51
+
52
+ ```
53
+ <|user|>\nprompt goes here\n<|assistant|>\n
54
+ ```
55
+
56
  <!-- compatibility_ggml start -->
57
  ## Compatibility
58
 
 
107
  I use the following command line; adjust for your tastes and needs:
108
 
109
  ```
110
+ ./main -t 10 -ngl 32 -m tulu-30b.ggmlv3.q5_0.bin --color -c 2048 --temp 0.7 --repeat_penalty 1.1 -n -1 -p "<|user|>\nwrite a story about llamas\n<|assistant|>\n"
111
  ```
112
  Change `-t 10` to the number of physical CPU cores you have. For example if your system has 8 cores/16 threads, use `-t 8`.
113