Commit graph

  • 75bbe787b6 Merge remote-tracking branch 'origin/main' into RFC-0001-The-Llama-Stack Ashwin Bharambe 2024-08-20 18:58:47 -07:00
  • 57881c08c1 Bump version to 0.0.8 Ashwin Bharambe 2024-08-19 20:12:01 -07:00
  • e08e963f86 Add --manifest-file option to argparser Ashwin Bharambe 2024-08-19 18:26:30 -07:00
  • b3da6b8afb Bump version to 0.0.7 Ashwin Bharambe 2024-08-19 16:27:36 -07:00
  • 23de941424 Bump version to 0.0.6 Ashwin Bharambe 2024-08-19 14:12:18 -07:00
  • 38244c3161 llama_models.llama3_1 -> llama_models.llama3 Ashwin Bharambe 2024-08-19 10:55:37 -07:00
  • f502716cf7 Fix ShieldType Union equality bug dltn 2024-08-18 19:13:15 -07:00
  • 63ae942f2f
    Update cli_reference.md with conda init and activate Anisha Nainani 2024-08-17 16:18:14 -05:00
  • 5e072d0780 Add a --manifest-file option to llama download Ashwin Bharambe 2024-08-17 10:08:00 -07:00
  • 124b2c1854 First cut at an observability API Ashwin Bharambe 2024-08-15 16:53:53 -07:00
  • 1f5eb9ff96 Add OpenAPI generation utility, update SPEC to reflect latest types Ashwin Bharambe 2024-08-15 13:45:45 -07:00
  • b8fc4d4dee
    Updates to prompt for tool calls (#29) Hardik Shah 2024-08-15 13:23:51 -07:00
  • f8623459fe added docstrings for ToolPromptFormat Hardik Shah 2024-08-15 13:03:39 -07:00
  • 00053b5bb0 function_tag system prompt is also added as a user message Hardik Shah 2024-08-15 12:11:35 -07:00
  • 417ba2aea0 RFC-0001-The-Llama-Stack Raghotham Murthy 2024-07-23 08:45:38 -07:00
  • 0d933ac4c5 No need for unnecessary $(conda run ...) to get python interpreter Ashwin Bharambe 2024-08-14 20:48:35 -07:00
  • 00f0e6d92b
    Avoid using nearly double the memory needed (#30) Ashwin Bharambe 2024-08-14 17:44:36 -07:00
  • 82667f460e Avoid using nearly double the memory needed Ashwin Bharambe 2024-08-14 17:08:54 -07:00
  • 8e526b9536 JSON in caps Hardik Shah 2024-08-14 15:26:23 -07:00
  • 86df597a83 support json format Hardik Shah 2024-08-14 12:43:43 -07:00
  • 48b78430eb Add tool prompt formats Hardik Shah 2024-08-13 16:00:47 -07:00
  • 0c3e754453 update system prompts to drop new line Hardik Shah 2024-08-13 13:01:05 -07:00
  • b311dcd143 formatting Dalton Flanagan 2024-08-14 17:03:43 -04:00
  • 069d877210 Typo bugfix (rename variable x -> prompt) Ashwin Bharambe 2024-08-14 13:47:27 -07:00
  • b6ccaf1778 formatting Dalton Flanagan 2024-08-14 14:22:25 -04:00
  • 94dfa293a6 Bump version to 0.0.5 Hardik Shah 2024-08-13 15:23:57 -07:00
  • 432957d6b6 fix typo dltn 2024-08-13 11:39:57 -07:00
  • 7f13853e5e
    Update README.md Hardik Shah 2024-08-12 17:10:02 -07:00
  • e459670a9c
    Update README.md Hamid Shojanazeri 2024-08-12 16:10:08 -07:00
  • 37da47ef8e upgrade pydantic to latest Hardik Shah 2024-08-12 15:14:21 -07:00
  • 2cd8b2ff5b Add simple validation for RemoteProviderConfig Ashwin Bharambe 2024-08-09 15:15:20 -07:00
  • 898cd5b352 Bump version to 0.0.4 dltn 2024-08-08 15:24:45 -07:00
  • 416097a9ea
    Rename inline -> local (#24) Dalton Flanagan 2024-08-08 17:39:03 -04:00
  • 6c6c75e716 further rename dltn 2024-08-08 14:29:13 -07:00
  • 640741c639 Rename the "inline" distribution to "local" Ashwin Bharambe 2024-08-08 14:05:49 -07:00
  • dd15671f7f Bump version to 0.0.3 Ashwin Bharambe 2024-08-08 13:40:03 -07:00
  • e830814399
    Introduce Llama stack distributions (#22) Ashwin Bharambe 2024-08-08 13:38:41 -07:00
  • 50775b7120 fix PIP_ARGS Ashwin Bharambe 2024-08-08 13:34:59 -07:00
  • 8ff9523747 Must install llama-models at that same version first Ashwin Bharambe 2024-08-08 13:11:14 -07:00
  • 6441b78aac Fix #2 to help with testing packages Ashwin Bharambe 2024-08-08 13:05:47 -07:00
  • 83adaae09b allow installing from test.pypi.org Ashwin Bharambe 2024-08-08 12:55:03 -07:00
  • 7b37ed5dae Add scripts to MANIFEST Ashwin Bharambe 2024-08-08 12:28:37 -07:00
  • ad9eaf13bb for inline make 8b model the default Hardik Shah 2024-08-08 10:45:32 -07:00
  • 8d7ecf0c47 Remove additional_pip_packages; move deps to providers Ashwin Bharambe 2024-08-08 10:19:46 -07:00
  • 6de36b6a15 make optional subclasses default to yes for configuration Hardik Shah 2024-08-08 10:23:54 -07:00
  • 55cf1f0642 resort imports Hardik Shah 2024-08-08 10:13:26 -07:00
  • 9e3182216d add validation for configuration input Ashwin Bharambe 2024-08-08 10:04:39 -07:00
  • ab856c174c update dependencies and rely on LLAMA_TOOLCHAIN_DIR for dev purposes Hardik Shah 2024-08-08 08:22:13 -07:00
  • 72b20b6f5a Avoid using conda run since it buffers everything Ashwin Bharambe 2024-08-07 22:21:38 -07:00
  • e33f402046 Merge remote-tracking branch 'origin/main' into distros Ashwin Bharambe 2024-08-07 21:55:29 -07:00
  • f27d629fe8 Reduce a bunch of dependencies from toolchain Ashwin Bharambe 2024-08-07 18:02:35 -07:00
  • da4645a27a
    hide non-featured (older) models from model list command without show-all flag (#23) Dalton Flanagan 2024-08-07 23:31:30 -04:00
  • f5629cc131 hide non-featured (older) models from model list command without show-all flag dltn 2024-08-07 19:52:22 -07:00
  • 171a178783 get ollama working Hardik Shah 2024-08-07 17:52:49 -07:00
  • ea50086190 Simpler intro statements Ashwin Bharambe 2024-08-07 16:29:22 -07:00
  • 9ec46d718d Show message about checksum file so users can check themselves Ashwin Bharambe 2024-08-07 16:17:51 -07:00
  • 57402c1a19 Add llama model download alias for llama download Ashwin Bharambe 2024-08-07 16:10:26 -07:00
  • fddaf5c929 Refactor download functionality out of the Command so can be reused Ashwin Bharambe 2024-08-07 15:27:00 -07:00
  • 68654460f8 Update README, add newline between API surface configurations Ashwin Bharambe 2024-08-07 15:14:59 -07:00
  • 66412b932b Nuke fp8_requirements, fold fbgemm into common requirements Ashwin Bharambe 2024-08-07 13:58:13 -07:00
  • cc697c59e5 Update CLI_reference Ashwin Bharambe 2024-08-06 22:18:02 -07:00
  • e1a7aa4773 Make install + start scripts do proper configuration automatically Ashwin Bharambe 2024-08-06 21:34:09 -07:00
  • 9e1ca4eeb1 add DistributionConfig, fix a bug in model download Ashwin Bharambe 2024-08-06 19:24:52 -07:00
  • ade574a0ef minor fixes Hardik Shah 2024-08-06 18:56:22 -07:00
  • f83b97992c
    Update cli_reference.md Hardik Shah 2024-08-06 18:55:31 -07:00
  • 2a9bdb208b update safety to use model sku ids and not model dirs Hardik Shah 2024-08-06 17:10:01 -07:00
  • a0e61a3c7a Fix passthrough streaming, send headers properly not part of body :facepalm Ashwin Bharambe 2024-08-06 16:39:38 -07:00
  • 039861f1c7 update inference config to take model and not model_dir Hardik Shah 2024-08-06 15:02:41 -07:00
  • 08c3802f45 dict key instead of attr Ashwin Bharambe 2024-08-06 14:30:57 -07:00
  • 7cc0445517 Rename Distribution -> DistributionSpec, simplify RemoteProviders Ashwin Bharambe 2024-08-06 10:45:06 -07:00
  • 0a67f3d3e6 installation fixes Hardik Shah 2024-08-05 18:04:36 -07:00
  • 0de5a807c7 Make each inference provider into its own subdirectory Ashwin Bharambe 2024-08-05 15:13:52 -07:00
  • f64668319c Merge remote-tracking branch 'origin/main' into distros Ashwin Bharambe 2024-08-05 14:31:06 -07:00
  • 65a9e40174 Adapter -> Provider Ashwin Bharambe 2024-08-05 13:26:29 -07:00
  • db3e6dda07 refactor a method out Ashwin Bharambe 2024-08-05 13:14:15 -07:00
  • 125fdb1b2a ApiSurface -> Api Ashwin Bharambe 2024-08-05 12:44:56 -07:00
  • 7664d5701d update tests and formatting Hardik Shah 2024-08-05 12:34:16 -07:00
  • 7890921e5c move straggler files and fix some important existing bugs Ashwin Bharambe 2024-08-05 09:24:45 -07:00
  • 5e972ece13 refactor to reduce size of agentic_system Ashwin Bharambe 2024-08-04 18:17:56 -07:00
  • be19b22391 Bring agentic system api to toolchain Ashwin Bharambe 2024-08-04 10:53:38 -07:00
  • b0e5340645 fixes Ashwin Bharambe 2024-08-03 22:16:24 -07:00
  • 750202ddd5 Add a Path() wrapper at the earliest place Ashwin Bharambe 2024-08-03 21:25:48 -07:00
  • 803976df26 cleanup, moving stuff to common, nuke utils Ashwin Bharambe 2024-08-03 20:32:57 -07:00
  • fe582a739d add safety adapters, configuration handling, server + clients Ashwin Bharambe 2024-08-03 19:46:59 -07:00
  • 9dafa6ad94 implement full-passthrough in the server Ashwin Bharambe 2024-08-03 14:15:20 -07:00
  • 38fd76f85c undo a typo, add a passthrough distribution Ashwin Bharambe 2024-08-02 20:48:53 -07:00
  • 67229f23a4 local imports for faster cli Hardik Shah 2024-08-02 16:34:29 -07:00
  • af4710c959 Improved exception handling Ashwin Bharambe 2024-08-02 14:54:06 -07:00
  • 493f0d99b2 updated dependency and client model name Hardik Shah 2024-08-02 15:37:40 -07:00
  • d7a4cdd70d added options to ollama inference Hardik Shah 2024-08-02 14:44:22 -07:00
  • d3e269fcf2 Remove inference uvicorn server entrypoint and llama inference CLI command Ashwin Bharambe 2024-08-02 14:18:25 -07:00
  • 3bc827cd5f read existing configuration, save enums properly Ashwin Bharambe 2024-08-02 13:55:29 -07:00
  • 2cf9915806 Distribution server now functioning Ashwin Bharambe 2024-08-02 13:37:40 -07:00
  • 041cafbee3 getting closer to a distro definition, distro install + configure works Ashwin Bharambe 2024-08-01 22:59:11 -07:00
  • dac2b5a1ed More progress towards llama distribution install Ashwin Bharambe 2024-08-01 16:40:43 -07:00
  • 5a583cf16e Add distribution CLI scaffolding Ashwin Bharambe 2024-08-01 14:44:57 -07:00
  • 09cf3fe78b Use new definitions of Model / SKU Ashwin Bharambe 2024-07-31 11:36:16 -07:00
  • 156bfa0e15
    Added Ollama as an inference impl (#20) Hardik Shah 2024-07-31 22:08:37 -07:00
  • fd8adc1e50 addressing comments Hardik Shah 2024-07-31 22:07:45 -07:00
  • c253c1c9ad Begin adding a /safety/run_shield API Ashwin Bharambe 2024-07-31 21:57:10 -07:00