fsal
commited on
Commit
·
b5ec36b
1
Parent(s):
60ad63d
Update AI Changelog
Browse files- AI_CHANGELOG.md +5 -0
AI_CHANGELOG.md
CHANGED
@@ -1,4 +1,9 @@
|
|
1 |
# AI CHANGELOG
|
|
|
|
|
|
|
|
|
|
|
2 |
## [Update Dockerfile, modify chatbot behavior and add new dependencies](https://github.com/fsal/streamlit-hf/commit/9032eeeefab2e2a3f2256743c3ed54aade53df5b)
|
3 |
Wed Feb 14 01:14:25 2024 +0100
|
4 |
- 1. Dockerfile: Changed the base image from python:3.11-slim-bookworm to python:3.11 and added gcc and tesseract-ocr packages.
|
|
|
1 |
# AI CHANGELOG
|
2 |
+
## [Added model parameter to get_texts_and_retriever_cacheable_wrapper and get_texts_and_multiretriever functions](https://github.com/fsal/streamlit-hf/commit/60ad63df0529a9f6a993084d9eeefdf68b04bdbc)
|
3 |
+
Wed Feb 14 11:45:08 2024 +0100
|
4 |
+
- The commit introduces a new 'model' parameter to the 'get_texts_and_retriever_cacheable_wrapper' and 'get_texts_and_multiretriever' functions in 'app.py' and 'llm_resources.py' respectively. This allows the caller to specify the model to be used.
|
5 |
+
- It also changes the default model from 'gpt-3.5-turbo' to 'gpt-4-turbo-preview' in 'defaults.py'. The new default model is fetched from the environment variable 'DEFAULT_MODEL', with 'gpt-4-turbo-preview' as the fallback.
|
6 |
+
- Finally, the 'model' parameter is used in the 'ChatOpenAI' instance in 'get_texts_and_multiretriever' function in 'llm_resources.py'.
|
7 |
## [Update Dockerfile, modify chatbot behavior and add new dependencies](https://github.com/fsal/streamlit-hf/commit/9032eeeefab2e2a3f2256743c3ed54aade53df5b)
|
8 |
Wed Feb 14 01:14:25 2024 +0100
|
9 |
- 1. Dockerfile: Changed the base image from python:3.11-slim-bookworm to python:3.11 and added gcc and tesseract-ocr packages.
|