eachadea
/

ggml-vicuna-13b-1.1

Document Question Answering

Model card Files Files and versions Community

eachadea commited on Jun 3, 2023

Commit

f8d9047

•

1 Parent(s): cd2811e

Update README.md

Files changed (1) hide show

README.md +1 -2

README.md CHANGED Viewed

@@ -27,9 +27,8 @@ New files inference up to 10% faster without any quality reduction.
 - Main files are based on v1.1 release
   - See changelog below
   - Use prompt template: ```HUMAN: <prompt> ASSISTANT: <response>```
-- Uncensored files are based on v0 release
   - Use prompt template: ```### User: <prompt> ### Assistant: <response>```
-- PR #896 was used for q4_0. Everything else is latest as of upload time.
 ### Quantization
 Several quantization methods are supported. They differ in the resulting model disk size and inference speed.

 - Main files are based on v1.1 release
   - See changelog below
   - Use prompt template: ```HUMAN: <prompt> ASSISTANT: <response>```
+- Uncensored files are based on a modified v0 release
   - Use prompt template: ```### User: <prompt> ### Assistant: <response>```
 ### Quantization
 Several quantization methods are supported. They differ in the resulting model disk size and inference speed.