v2.2.2

Fixed

• fix(chatwithtools): Expect a list of tool messages
• fix(main): Don't drop background task when app is disabled
• fix: Only run background thread once
• fix(summarize): Improve prompt
• fix(chat): Expect json-stringified messages in history {role, content}

v2.2.1

Fixed

• fixed failed import

v2.2.0

New

enh(summarize): Try to make it use bulleted lists for longer texts
enh: Implement chatwithtools task type
feat: Use chunking if the text doesn't fit the context

Fixed

fix(summarize): Use a better algorithm for chunked summaries
fix(summarize): Always summarize at least once
fix(ci): app_api is pre-installed from NC 31 (#37) Anupam Kumar* 03.10.24, 14:13

v2.1.4

[2.1.4] - 2024-09-12

Fix

• update docker image version

2.1.3

[2.1.3] - 2024-09-11

Fix

• update context size for llama 3.1

v2.1.2

[2.1.2] - 2024-08-26

Fix

• filename of the llama3.1 model in config
• catch JSONDecodeError for when server is in maintenance mode

Change

• better app_enabled handling

v2.1.1

[2.1.1] - 2024-08-23

Fix

• compare uppercase COMPUTE_DEVICE value (#27)

v2.1.0

[2.1.0] - 2024-08-22

Fix

• Catch network exceptions and keep the loop going

Change

• Migrate default config to llama compatible config
• Use COMPUTE_DEVICE to determine gpu offloading
• Use TaskProcessingProvider class for registration
• Better handling of app enabled state
• Download models on /init

v2.0.1

fix: Loading multiple models was broken

Signed-off-by: Marcel Klehr <[email protected]>

v2.0.0

Signed-off-by: Marcel Klehr <[email protected]>