Commit graph

  • f046899a1c datasets api Xi Yan 2024-10-14 13:16:39 -07:00
  • 209cd3d35e Bump version to 0.0.42 Xi Yan 2024-10-14 11:13:04 -07:00
  • 6f7318b243 readme sanjaychelliah 2024-10-14 14:21:58 +05:30
  • 18fe966e96 registry refactor Xi Yan 2024-10-14 00:12:46 -07:00
  • 78cb88c3c4 RunEvalTask / InferenceGenerator Xi Yan 2024-10-13 23:48:15 -07:00
  • 8890de7322 cleanup original BaseTask Xi Yan 2024-10-13 23:30:21 -07:00
  • a25aff290e generator + scorer Api for MMLU Xi Yan 2024-10-13 23:27:02 -07:00
  • 4b78f3dce6
    Add llama download cli reference for multiple models ABucket 2024-10-12 14:29:43 +08:00
  • 8fcded0004 Add llama download support for multiple models with comma-separated list ABucket 2024-10-12 14:28:36 +08:00
  • dfa570ca2e wait for tgi server to start before starting server Xi Yan 2024-10-11 18:54:48 -07:00
  • 2da8aa18bf path Xi Yan 2024-10-11 18:25:20 -07:00
  • 0325a767b6 tgi docker compose Xi Yan 2024-10-11 18:24:11 -07:00
  • fe9029f169
    Remove request arg from chat completion response processing Yuan Tang 2024-10-11 16:18:40 -04:00
  • a2b87ed0cb
    Switch to pre-commit/action (#239) Yuan Tang 2024-10-11 14:09:11 -04:00
  • 2dd4fc5ca5
    Switch to pre-commit/action Yuan Tang 2024-10-11 13:33:28 -04:00
  • 05282d1234
    Enable pre-commit on main branch (#237) Yuan Tang 2024-10-11 13:03:59 -04:00
  • fb565dfb06 eleuther eval fix Xi Yan 2024-10-11 09:30:10 -07:00
  • 2128e61da2
    Fix incorrect completion() signature for Databricks provider (#236) Yuan Tang 2024-10-11 11:47:57 -04:00
  • 29156780ff
    refactor: Qdrant with tests Anush008 2024-10-11 11:43:56 +05:30
  • 65b1f47d1a
    Merge branch 'meta-llama:main' into qdrant Anush 2024-10-11 10:49:53 +05:30
  • ad18dc94ac add data structure to tasks Xi Yan 2024-10-10 21:33:13 -07:00
  • 4ce9314fdd Fixes #183: Preload models during server initialization to prevent request timeouts Ezreal 2024-10-11 12:27:47 +08:00
  • 9fbe8852aa
    Add Swift Package Index badge Dalton Flanagan 2024-10-10 23:39:25 -04:00
  • ca29980c6b fix agents context retriever Xi Yan 2024-10-10 20:17:29 -07:00
  • 9816c9aae6 wip add datatypes Xi Yan 2024-10-10 19:56:19 -07:00
  • b9c54ab43f
    Enable pre-commit on main branch Yuan Tang 2024-10-10 21:14:33 -04:00
  • c18ca95619
    Fix precommit Yuan Tang 2024-10-10 21:10:50 -04:00
  • 98ed50bb0e
    Fix precommit Yuan Tang 2024-10-10 21:09:44 -04:00
  • d7c0d2c46f
    Fix precommit Yuan Tang 2024-10-10 21:08:19 -04:00
  • 59b4a4bbe5
    Fix incorrect completion() signature for Databricks provider Yuan Tang 2024-10-10 21:03:14 -04:00
  • 99ed1425fc add dataset datatypes Xi Yan 2024-10-10 17:19:18 -07:00
  • 1ff0476002 Split off meta-reference-quantized provider Ashwin Bharambe 2024-10-10 15:54:08 -07:00
  • 7ff5800dea generate openapi Xi Yan 2024-10-10 15:30:34 -07:00
  • a3e65d58a9
    Add logo Dalton Flanagan 2024-10-10 15:04:21 -04:00
  • c8de439d9f clean Xi Yan 2024-10-10 11:38:37 -07:00
  • 31c046dcdf evals new rebase Xi Yan 2024-10-10 11:35:26 -07:00
  • eba9d1ea14
    ci: Run pre-commit checks in CI (#176) Russell Bryant 2024-10-10 14:21:59 -04:00
  • 89d24a07f0 Bump version to 0.0.41 Ashwin Bharambe 2024-10-10 10:27:03 -07:00
  • 6bb57e72a7
    Remove "routing_table" and "routing_key" concepts for the user (#201) Ashwin Bharambe 2024-10-10 10:24:13 -07:00
  • a33cafc2fe Update bedrock Ashwin Bharambe 2024-10-10 10:19:06 -07:00
  • fe0dabe596 Fix pgvector, store source of truth in Chroma Ashwin Bharambe 2024-10-10 10:12:14 -07:00
  • 8c3010553f
    Fix agents path in generate.py Dalton Flanagan 2024-10-10 11:41:03 -04:00
  • 09d7573739 README: add deepinfra in the comparison table Iskren Chernev 2024-10-10 14:20:59 +03:00
  • dd9d34cf7d serve routing table API unconditionally Ashwin Bharambe 2024-10-09 22:48:31 -07:00
  • 8be385c994 Make sure we always serve the routing table APIs if the corresponding router APIs are being served Ashwin Bharambe 2024-10-09 22:47:18 -07:00
  • a5d7caf21b typedef fix Ashwin Bharambe 2024-10-09 22:31:26 -07:00
  • 94dc0782bf Update test Ashwin Bharambe 2024-10-09 22:29:51 -07:00
  • 8f397d7f62 bug fix for llama stack configure Ashwin Bharambe 2024-10-09 22:27:47 -07:00
  • 8a175129fc fix weaviate, update run.yamls Ashwin Bharambe 2024-10-09 22:15:28 -07:00
  • 238e658cdf Kill irrelevant (now) method Ashwin Bharambe 2024-10-09 22:11:18 -07:00
  • 77a486f176 added tool calling test Ashwin Bharambe 2024-10-09 22:01:28 -07:00
  • ef4b74c935 Add a simple agents test case Ashwin Bharambe 2024-10-09 21:52:49 -07:00
  • 2d94ca71a9 Pass memory bank API to agent impl Ashwin Bharambe 2024-10-09 21:16:57 -07:00
  • 6788173ffc re-gen openapi spec Ashwin Bharambe 2024-10-09 21:10:12 -07:00
  • fcd22b6baa Make Safety test work, other cleanup Ashwin Bharambe 2024-10-09 21:09:50 -07:00
  • ba1f294cc6 Safety test placeholder Ashwin Bharambe 2024-10-09 19:35:48 -07:00
  • b55034c0de Another round of simplification and clarity for models/shields/memory_banks stuff Ashwin Bharambe 2024-10-09 19:19:26 -07:00
  • 38e7740d0b ifevals eleuther task Xi Yan 2024-10-09 16:08:07 -07:00
  • 58992dacc2 refactor Xi Yan 2024-10-09 14:47:36 -07:00
  • 7a8aa775e5
    JSON serialization for parallel processing queue (#232) Dalton Flanagan 2024-10-09 17:24:12 -04:00
  • adb768f827 async call in separate thread Xi Yan 2024-10-09 13:18:15 -07:00
  • da47273dd1 second round of feedback dltn 2024-10-09 11:56:02 -07:00
  • ae43044a57 rename tasks Xi Yan 2024-10-09 11:30:35 -07:00
  • fd41ddf733 bidirectional wrapper dltn 2024-10-09 11:28:11 -07:00
  • 3adf1dc20e address feedback dltn 2024-10-09 11:14:02 -07:00
  • d6924327ce fixup dltn 2024-10-09 09:33:25 -07:00
  • 0deeb4abec send/recv pydantic json over socket dltn 2024-10-09 09:20:13 -07:00
  • 4a336ca8ba add deepinfra inference provider Iskren Chernev 2024-10-09 17:07:52 +03:00
  • 0f66ae0f61
    Add function for stopping inference (#224) kebbbnnn 2024-10-09 22:50:19 +08:00
  • 63eb188239
    Add --disable-ipv6 when starting container to avoid httpx.ReadError Yuan Tang 2024-10-09 10:29:19 -04:00
  • 9c38d9ae13 eleuther generate until Xi Yan 2024-10-08 23:57:22 -07:00
  • 6abef716dd rebase on top of registry Xi Yan 2024-10-08 23:41:03 -07:00
  • 0919072a33 eleuther custom tasks Xi Yan 2024-10-08 23:22:50 -07:00
  • ec4ece8523 Add function for stopping inference Kevin Ladan 2024-10-09 12:33:19 +08:00
  • 73a0a34e39 Kill non-llama guard shields Ashwin Bharambe 2024-10-08 17:47:03 -07:00
  • 24c61403b7 Fixes Ashwin Bharambe 2024-10-08 17:43:25 -07:00
  • a86f3ae07d Update run.yaml Ashwin Bharambe 2024-10-08 17:41:06 -07:00
  • 924b1fba09 minor Ashwin Bharambe 2024-10-08 17:26:26 -07:00
  • f40cd62306 Test fixes Ashwin Bharambe 2024-10-08 16:57:52 -07:00
  • 8eee5b9adc Fix server conditional awaiting on coroutines Ashwin Bharambe 2024-10-08 17:03:31 -07:00
  • 216e7eb4d5 Move async with SEMAPHORE inside the async methods Ashwin Bharambe 2024-10-08 16:53:05 -07:00
  • 4540d8bd87 move codeshield into an independent safety provider Ashwin Bharambe 2024-10-08 15:45:50 -07:00
  • 380b9dab90 regen openapi specs Ashwin Bharambe 2024-10-08 15:19:19 -07:00
  • 7f1160296c Updates to server.py to clean up streaming vs non-streaming stuff Ashwin Bharambe 2024-10-08 14:28:50 -07:00
  • 640c5c54f7 rename augment_messages Ashwin Bharambe 2024-10-08 13:48:44 -07:00
  • 336cf7a674 update vllm; not quite tested yet Ashwin Bharambe 2024-10-08 13:38:32 -07:00
  • ed899a5dec Convert TGI to work with openai_compat Ashwin Bharambe 2024-10-08 12:57:34 -07:00
  • 05e73d12b3 introduce openai_compat with the completions (not chat-completions) API Ashwin Bharambe 2024-10-08 12:15:55 -07:00
  • 0c9eb3341c Separate chat_completion stream and non-stream implementations Ashwin Bharambe 2024-10-08 10:52:16 -07:00
  • f8752ab8dc weaviate fixes, test now passes Ashwin Bharambe 2024-10-08 09:54:00 -07:00
  • f21ad1173e improve memory test, but it fails on chromadb :/ Ashwin Bharambe 2024-10-07 23:09:31 -07:00
  • 4ab6e1b81a Add really basic testing for memory API Ashwin Bharambe 2024-10-07 22:34:53 -07:00
  • dba7caf1d0 Fix fireworks and update the test Ashwin Bharambe 2024-10-07 17:43:47 -07:00
  • bbd3a02615 Make Together inference work using the raw completions API Ashwin Bharambe 2024-10-07 17:28:19 -07:00
  • 3ae2b712e8 Add inference test Ashwin Bharambe 2024-10-07 15:46:16 -07:00
  • 4fa467731e Fix a bug in meta-reference inference when stream=False Ashwin Bharambe 2024-10-07 14:35:50 -07:00
  • 353c7dc82a A few bug fixes for covering corner cases Ashwin Bharambe 2024-10-07 13:55:01 -07:00
  • a05599c67a Weaviate "should" work (i.e., is code-complete) but not tested Ashwin Bharambe 2024-10-06 22:50:34 -07:00
  • 118c0ef105 Partial cleanup of weaviate Zain Hasan 2024-10-07 01:21:50 -04:00
  • 862f8ddb8d more memory related fixes; memory.client now works Ashwin Bharambe 2024-10-06 22:10:24 -07:00