mirror of
https://gh.wpcy.net/https://github.com/discourse/discourse.git
synced 2026-05-12 18:43:30 +08:00
00c98781c1
Adds support for: https://blog.google/products/gemini/updated-image-editing-model/ This mode generates images natively, we updated the Gemini endpoint to support handling image output. Additionally, there is a refinement to how we provide image context to models, we avoid very short descriptions and simply inline images in the context, prior to this fix image generating models could get confused and start generating markdown blobs. This takes care to properly omit images from context in model responses for anthropic and open ai in responses api. |
||
|---|---|---|
| .. | ||
| dialects | ||
| endpoints | ||
| anthropic_message_processor.rb | ||
| cancel_manager.rb | ||
| json_stream_decoder.rb | ||
| json_streaming_parser.rb | ||
| json_streaming_tracker.rb | ||
| llm.rb | ||
| nova_message_processor.rb | ||
| open_ai_message_processor.rb | ||
| open_ai_responses_message_processor.rb | ||
| prompt.rb | ||
| prompt_messages_builder.rb | ||
| report.rb | ||
| structured_output.rb | ||
| thinking.rb | ||
| tool_call.rb | ||
| tool_definition.rb | ||
| upload_encoder.rb | ||
| xml_tag_stripper.rb | ||
| xml_tool_processor.rb | ||
