imone commited on
Commit
845e9e4
β€’
1 Parent(s): dc7d889

Update README

Browse files
Files changed (1) hide show
  1. README.md +17 -8
README.md CHANGED
@@ -7,24 +7,33 @@ tags:
7
 
8
  # OpenChat: Less is More for Open-source Models
9
 
10
- OpenChat is a series of open-source language models fine-tuned on very little diverse and high-quality multi-round conversations. The [dataset](https://huggingface.co/datasets/openchat/openchat_sharegpt4_dataset) contains only ~6K GPT-4 conversations filtered from the 90K ShareGPT conversations.
11
 
12
- Generic models:
13
 
14
  - OpenChat: based on LLaMA-13B (2048 context length)
15
- - **105.7%** of ChatGPT score on Vicuna GPT-4 evaluation
16
- - **80.87%** Win-rate on AlpacaEval
17
- - **πŸš€ Only used 6K data for finetuning!!!**
18
  - OpenChat-8192: based on LLaMA-13B (extended to 8192 context length)
19
  - **106.6%** of ChatGPT score on Vicuna GPT-4 evaluation
 
20
 
21
- Code models:
22
 
23
  - OpenCoderPlus: based on StarCoderPlus (native 8192 context length)
24
  - **102.5%** of ChatGPT score on Vicuna GPT-4 evaluation
25
- - **78.70%** Win-rate on AlpacaEval
26
 
27
- **NOTE:** Please load the pretrained models using *bfloat16*
 
 
 
 
 
 
 
 
28
 
29
  ## Conversation Template
30
 
 
7
 
8
  # OpenChat: Less is More for Open-source Models
9
 
10
+ OpenChat is a series of open-source language models fine-tuned on a diverse and high-quality dataset of multi-round conversations. With only ~6K GPT-4 conversations filtered from the ~90K ShareGPT conversations, OpenChat is designed to achieve high performance with limited data.
11
 
12
+ **Generic models:**
13
 
14
  - OpenChat: based on LLaMA-13B (2048 context length)
15
+ - **πŸš€ 105.7%** of ChatGPT score on Vicuna GPT-4 evaluation
16
+ - **πŸ”₯ 80.9%** Win-rate on AlpacaEval
17
+ - **πŸ€— Only used 6K data for finetuning!!!**
18
  - OpenChat-8192: based on LLaMA-13B (extended to 8192 context length)
19
  - **106.6%** of ChatGPT score on Vicuna GPT-4 evaluation
20
+ - **79.5%** of ChatGPT score on Vicuna GPT-4 evaluation
21
 
22
+ **Code models:**
23
 
24
  - OpenCoderPlus: based on StarCoderPlus (native 8192 context length)
25
  - **102.5%** of ChatGPT score on Vicuna GPT-4 evaluation
26
+ - **78.7%** Win-rate on AlpacaEval
27
 
28
+ *Note:* Please load the pretrained models using *bfloat16*
29
+
30
+ ## Code and Inference Server
31
+
32
+ We provide the full source code, including an inference server compatible with the "ChatCompletions" API, in the [OpenChat](https://github.com/imoneoi/openchat) GitHub repository.
33
+
34
+ ## Web UI
35
+
36
+ OpenChat also includes a web UI for a better user experience. See the GitHub repository for instructions.
37
 
38
  ## Conversation Template
39