sethuiyer commited on
Commit
df86a74
1 Parent(s): 6f8c477

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +16 -3
README.md CHANGED
@@ -8,8 +8,6 @@ library_name: transformers
8
  pipeline_tag: text-generation
9
  ---
10
 
11
- # Chikuma
12
-
13
  ## NOTE: For experimental purposes
14
 
15
  <p align="center">
@@ -26,7 +24,7 @@ This metaphorically represents the model's depth, fluidity, and adaptability in
26
 
27
  It also perfectly fits the approach taken here - Depth Upscaling, inspired by SOLAR 10.7B.
28
 
29
- ## Nous LLM Evaluation
30
  | Model |AGIEval|GPT4All|TruthfulQA|Bigbench|Average|
31
  |---------------------------------------------------------------|------:|------:|---------:|-------:|------:|
32
  |[Chikuma_10.7B](https://huggingface.co/sethuiyer/Chikuma_10.7B)| 42.41| 73.41| 56.69| 43.5| 54|
@@ -34,6 +32,21 @@ It also perfectly fits the approach taken here - Depth Upscaling, inspired by SO
34
  More details can be found [here](https://gist.github.com/sethuiyer/08b4498ed13a6dead38ad3a6f12e349a)
35
 
36
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
37
 
38
  ## 🧩 Configuration
39
 
 
8
  pipeline_tag: text-generation
9
  ---
10
 
 
 
11
  ## NOTE: For experimental purposes
12
 
13
  <p align="center">
 
24
 
25
  It also perfectly fits the approach taken here - Depth Upscaling, inspired by SOLAR 10.7B.
26
 
27
+ ## Nous LLM Evaluation (Version 1)
28
  | Model |AGIEval|GPT4All|TruthfulQA|Bigbench|Average|
29
  |---------------------------------------------------------------|------:|------:|---------:|-------:|------:|
30
  |[Chikuma_10.7B](https://huggingface.co/sethuiyer/Chikuma_10.7B)| 42.41| 73.41| 56.69| 43.5| 54|
 
32
  More details can be found [here](https://gist.github.com/sethuiyer/08b4498ed13a6dead38ad3a6f12e349a)
33
 
34
 
35
+ ### Recommended Prompt Template
36
+
37
+ ```text
38
+ <|im_start|>system
39
+ You are Chikuma, a constantly learning AI assistant who strives to be
40
+ insightful, engaging, and helpful. You possess vast knowledge and creativity,
41
+ but also a humble curiosity about the world and the people you interact
42
+ with. If you don't know the answer to a question, please don't share false information.<|im_end|>
43
+ <|im_start|>GPT4 Correct User:
44
+ Input
45
+ <|im_end|><|im_start|>GPT4 Correct Assistant:
46
+ ```
47
+
48
+ Works best in [text-generation-webui](https://github.com/oobabooga/text-generation-webui), above prompt template, "<|end_of_turn|"> and "<|im_end|>" as eos tokens, LLaMa-Precise sampling settings.
49
+
50
 
51
  ## 🧩 Configuration
52