0
代码介绍
代码
Issues
Pull Requests
流水线
Actions
讨论
Wiki
项目成员
分析
项目设置
0
  1. octopus
  2. /
  3. internal
  4. /
  5. transformer
  6. /
  7. model
GGitHubfeat(transformer): add support for DeepSeek thinking mode and Nvidia chat template kwargs (#269)
4c4c79be创建于 18 天前历史提交
文件最后提交记录最后更新时间
embedding_test.go
feat: :sparkles: implement OpenAI Embedding API support (#123) Add full support for OpenAI Embedding API: - Add EmbeddingInbound and EmbeddingOutbound transformers with field name mapping - Internal format uses embedding_input/embedding_dimensions/embedding_encoding_format - External format maintains OpenAI standard (input/dimensions/encoding_format/data) - Add EmbeddingInput and related model structures - Add embedding endpoint handler (/embeddings) - Add channel type validation (embedding vs chat requests) - Add EmbeddingChannelTypes and ChatChannelTypes for type filtering - Update frontend UI with OpenAI Embedding channel option Co-authored-by: Claude <noreply@anthropic.com>4 个月前
gemini.go
fix: :bug: Gemini function call #67 4 个月前
interface.go
feat: :sparkles: 日志记录 5 个月前
model.go
feat(transformer): add support for DeepSeek thinking mode and Nvidia chat template kwargs (#269) Add new fields to InternalLLMRequest struct: - Thinking: controls DeepSeek thinking mode with type and budget_tokens configuration - ChatTemplateKwargs: enables passing chat template parameters for Nvidia models18 天前