Google GitGoogle Git
Sign in
chromium/external/github.com/google-ai-edge/LiteRT-LM/HEAD
  1. e490e86 MiniCPM-V-4 multimodal support by Yi-Chun Kuo · 4 hours ago upstream/main
  2. 09dc239 Infer vision tokens per image from metadata in EmbeddingEngineImpl to selectively load matching vision signatures. by Matthew Chan · 6 hours ago
  3. 7b28b9f refactor: remove model_role and force_string_content from GenericModel by Wai Hon Law · 6 hours ago
  4. 4453b28 Update dependencies of litert_lm by Fengwu Yao · 6 hours ago
  5. 495bee7 Add activationDataType option to EngineConfig in Kotlin SDK. by Tenghui Zhu · 7 hours ago
  6. 58718e2 Support exact full-vocabulary token log-probabilities in TopPSampler on CPU. by Google AI Edge · 7 hours ago
  7. 3842dad Select text encoder signatures from bundle metadata when streaming by Google AI Edge · 7 hours ago
  8. 7cadb97 Add --output_embedding_path flag and audio preprocessor support to embedding_litert_lm_main CLI tool. by Matthew Chan · 7 hours ago
  9. d6fdcac Let external weights load when the .litertlm is opened with mmap by Google AI Edge · 7 hours ago
  10. 385965d Add activationDataType option to EngineConfig in Swift SDK. by Tenghui Zhu · 8 hours ago
  11. e3e0a1c Expose embedding model info in Kotlin. by Google AI Edge · 10 hours ago
  12. 5c2f444 Add max_input_length, activation_data_type, and audio/vision backend flags for LiteRT-LM embedding benchmarks. by Matthew Chan · 10 hours ago
  13. 9916532 Expose embedding model info in Kotlin. by Google AI Edge · 11 hours ago
  14. 9033956 chore: remove unused messages.name from chat template standard by Wai Hon Law · 12 hours ago
  15. 2193add Rename DetectIsSwa to DetectUsesRingbuffer and propagate ringbuffer naming. by Matt Kreileder · 12 hours ago
  16. 87926fe Expose embedding model info in Python API & CLI, Swift by Google AI Edge · 13 hours ago
  17. 7a01f22 Run Kokoro vocoder on GPU, derive speech capacity, and apply slice trimming. by Yi-Chun Kuo · 13 hours ago
  18. aa8f35a Compile the OSS stb implementations exactly once by Yi-Chun Kuo · 13 hours ago
  19. cf2406f Reinterpret the streamed weights scratch pointer as unsigned by Google AI Edge · 13 hours ago
  20. 91a926c fix(executor): clamp max_num_tokens against static model capacity by Tenghui Zhu · 14 hours ago
  21. 11ebc16 Enable GPU and NPU testing for litert_lm_api on Linux and Windows. by Google AI Edge · 14 hours ago
  22. 61d2ce3 Support multi-tokenizer extraction by ModelType in runtime by Google AI Edge · 15 hours ago
  23. 930a642 Record vision_executor timing marks on the raw image TensorBufferMap path as well. by Google AI Edge · 15 hours ago
  24. 0d0d009 feat: add chatTemplate to Swift ConversationConfig by Wai Hon Law · 29 hours ago
  25. 70663d5 refactor: clean up edit_template_for_minijinja in PromptTemplate by Wai Hon Law · 30 hours ago
  26. 70a4a60 Cleanup legacy model buffer workarounds by Matt Kreileder · 31 hours ago
  27. 322b6e3 Kokoro TTS text chunking, phonemizer, and token slicing & DSP join trimming utilities. by Yi-Chun Kuo · 33 hours ago
  28. a467324 docs: remove eos_token from chat template standard by Wai Hon Law · 34 hours ago
  29. d8e6ae8 Support Embedding Gemma and LiteRT-LM embedding model info capabilities. by Google AI Edge · 35 hours ago
  30. 7bdd839 Enable benchmark mode for LiteRT-LM Embedding Engine. by Google AI Edge · 35 hours ago
  31. 5464a0f feat(cli): use hard link for model import by Wai Hon Law · 36 hours ago
  32. d29494f Fix TSAN data races and constraint state leak in LLM sessions. by Google AI Edge · 36 hours ago
  33. ed85647 Support Embedding Gemma and LiteRT-LM embedding model info capabilities. by Google AI Edge · 2 days ago
  34. 966068f feat: support CPU thread count in C, Python, and Swift EmbeddingEngine by Wai Hon Law · 2 days ago
  35. 9764ebc Expose LiteRT Environment via Engine interface and Session. by Google AI Edge · 2 days ago
  36. f300c4f fix(tool_use): keep integers as integers in tool call arguments by Google AI Edge · 2 days ago
  37. 62a040a feat: add chatTemplate to Kotlin ConversationConfig by Wai Hon Law · 4 days ago
  38. 20b0dea Add min_input_length support to LiteRT-LM embedding runtime and metadata. by Matthew Chan · 4 days ago
  39. 72dd8c4 feat: add Prem-1B-SQL chat template and model metadata by Wai Hon Law · 4 days ago
  40. 561c94e Add WebGPU streamed loading integration tests for EmbeddingEngine by Google AI Edge · 4 days ago
  41. 801acde refactor: simplify chat templates and schemas to require multimodal parts for content by Wai Hon Law · 5 days ago
  42. 53385d7 feat: add Qwen 2.5 chat template and model metadata by Wai Hon Law · 5 days ago
  43. b69096d feat: support main CPU thread count in Kotlin EmbeddingEngine by Wai Hon Law · 5 days ago
  44. 1ce7234 [LiteRT-LM] Optimize llg_fc_tool_calls_test to prevent ASAN timeouts by Matthew Chan · 5 days ago
  45. d667433 Add `ActivationDataType` and `cache_dir` options for MediaPipe Universal Embedder by Google AI Edge · 5 days ago
  46. 1fe80ce No public description by Google AI Edge · 5 days ago
  47. 6c6b458 refactor: normalize message content to multimodal parts in model data processors by Wai Hon Law · 5 days ago
  48. 2518fba Support activation data type precedence waterfall in EmbeddingEngineSettings. by Matthew Chan · 5 days ago
  49. 4050343 Add WASM bindings and TypeScript SDK for EmbeddingEngine by Google AI Edge · 5 days ago
  50. 596b2bf Add default max_input_length field in EmbeddingMetadata and use it to default max_input_length in EmbeddingEngine. by Matthew Chan · 5 days ago
  51. c5264db Update LlguidanceConfig to support special tokens by Google AI Edge · 5 days ago
  52. dc558fd Add self-contained WAV audio decoding (16-bit PCM and 32-bit float, mono/stereo) to AudioPreprocessorPffft to support raw audio byte inputs in addition to pre-decoded PCM frames. by Google AI Edge · 5 days ago
  53. 97c091d Internal change by Mohammadreza Heydary · 5 days ago
  54. 8096ebb Add thread-local error reporting to the LiteRT LM C API. by Mohammadreza Heydary · 6 days ago
  55. 0d43b55 Add AudioSessionAdvanced and audio execution APIs to LiteRT-LM. by Google AI Edge · 6 days ago
  56. 2e99921 Suppress new Pyrefly findings before upgrading to v1.2.0 by Oleh Prypin · 6 days ago
  57. 12fae71 Embed model_metadata.json and provide reusable model_metadata config helper by Byungchul Kim · 6 days ago
  58. 8de38f5 Pick Skia decoder based on `SK_CODEC_DECODES_BMP_WITH_RUST`. by Lukasz Anforowicz · 6 days ago
  59. fdab3ec Fix LiteRT TTS async streaming lifecycle, scheduler state transitions, and JNI linkage by Yi-Chun Kuo · 6 days ago
  60. 8e3cdbd Add streamed model loading support to EmbeddingEngineImpl by Google AI Edge · 6 days ago
  61. 731af6c Rename InspectModel to GetModelInfo in LiteRT-LM model_info by Google AI Edge · 6 days ago
  62. 20ac077 feat: add support for lfm2 in models and runtime conversation data processor by Wai Hon Law · 6 days ago
  63. a3645fb Internal change by Matthew Chan · 6 days ago
  64. 526b394 Reverts 202da5129b17fded07fb69421d9bc306ef605418 by Google AI Edge · 6 days ago
  65. 8bb1e1c Add StreamedWeightsManager and streamed model support to Vision, Audio, and Embedding Executors by Google AI Edge · 6 days ago
  66. 1c5890e Add support for oneOf in constrained decoding. by Mohammadreza Heydary · 6 days ago
  67. 202da51 Rename InspectModel to GetModelInfo in LiteRT-LM model_info by Google AI Edge · 6 days ago
  68. bebd7dd Add ModelResourcesStreaming for streamed model loading in LiteRT-LM by Google AI Edge · 6 days ago
  69. 172d680 Disable Metal argument buffers. by Fengwu Yao · 7 days ago
  70. 945527d Factor out utils functions from litertlm jni. by Yi-Chun Kuo · 7 days ago
  71. 15b90e6 Apply ad-hoc model fixes for models missing model type. by Matt Kreileder · 7 days ago
  72. 12addd2 Make ad-hoc model fixes dependent on model type by Matt Kreileder · 7 days ago
  73. 6311aaa Support ingestion of pre-computed audio embeddings in LiteRT-LM prefill. by Google AI Edge · 7 days ago
  74. 8ef5385 Update LiteRT-LM Swift package to v0.17.0. by Google AI Edge · 7 days ago
  75. 9c2341d Fix tinygemma-asr transcription quality across streaming audio chunks. by Byungchul Kim · 8 days ago
  76. 46b747f NPU: Detect valid_mask and allocate buffer if needed by Matt Kreileder · 8 days ago
  77. 60de4ed NPU: Support re-sizing of KV cache during backend creation by Matt Kreileder · 8 days ago
  78. a0f1e97 Register Gated Delta Net custom op in LiteRT-LM runtime executor. by Google AI Edge · 9 days ago
  79. a345f8b Automated Code Change by Google AI Edge · 9 days ago
  80. 007e7c7 Support multimodal audio language models in omni ASR pipeline. by Byungchul Kim · 11 days ago
  81. dd4898b Add experimental.h to LiteRT LM Swift build target. by Mohammadreza Heydary · 11 days ago
  82. 262c3af Refactor InspectModel implementation in LiteRT-LM model_info into modular helper functions. by Salil Tambe · 11 days ago
  83. 26e88ef chore: version bump to 0.18.0 by Mohammadreza Heydary · 11 days ago
  84. 38fb04e Remove the Releases section from the README. by Mohammadreza Heydary · 11 days ago
  85. c1cf637 Refactor Capabilities API and CLI to ModelInfo across all languages. by Google AI Edge · 11 days ago
  86. 49cf347 docs: fix Gemma4-E4B Hugging Face repo in README by Wai Hon Law · 11 days ago
  87. cbbb015 ci: remove redundant Android API 33 and 34 from CI workflows by Wai Hon Law · 11 days ago
  88. d7c0ac3 Refactor Capabilities API and CLI to ModelInfo across all languages. by Google AI Edge · 11 days ago
  89. c895001 build(kotlin): version bump to Kotlin 2.4 and maven dependencies by Wai Hon Law · 12 days ago
  90. b5253ed Refactor Capabilities API and CLI to ModelInfo across all languages. by Google AI Edge · 12 days ago
  91. d8a8235 Rollback: refactor: remove unused bos_token and eos_token from the Chat Template Standard by Wai Hon Law · 12 days ago
  92. b41b3c3 [LiteRT-LM] Add LiteRT compiled model metrics collection by Google AI Edge · 12 days ago
  93. 669f521 [LiteRT-LM] Add latency stats tracking to LiteRT LM embedding pipeline by Google AI Edge · 12 days ago
  94. c9b1ba5 Support LiteRT-LM decoder and .litertlm bundles in ASR runner. by Byungchul Kim · 12 days ago
  95. 5215d2d More robust handling in the miniaudio preprocessor by resizing PCM frames when fewer frames are read than requested. by Google AI Edge · 12 days ago
  96. d2f8394 Reset the audio executor after generating embeddings. by Yi-Chun Kuo · 12 days ago
  97. b41b12a Internal code update by Tenghui Zhu · 12 days ago
  98. 49fab24 docs: remove now from chat template standard input by Wai Hon Law · 12 days ago
  99. 65c710d refactor: remove bos_token and eos_token from chat template inputs by Wai Hon Law · 13 days ago
  100. fd3999e refactor: remove unused bos_token and eos_token from the Chat Template Standard by Wai Hon Law · 13 days ago