Releases: instill-ai/model-backend
Releases · instill-ai/model-backend
v0.25.1-alpha
v0.25.0-alpha
v0.24.0-alpha
0.24.0-alpha (2024-06-06)
⚠ BREAKING CHANGES
- model: adopt containerized model serving (#542)
Features
- handler: implement get latest operation (#589) (33d2395)
- handler: support listing available regions for model deployment (#561) (52c2172)
- handler: support model profile image (#566) (0c8dbba)
- model: add permission field in model object (#576) (2d36a58)
- model: add task schema in model struct (#578) (647069d)
- model: adopt containerized model serving (#542) (3c80f39)
- model: embed sample input/output in model proto message (#558) (5fba538)
- model: support latest model version trigger (#580) (47cb36c)
- model: support resource spec in model definition (#557) (fee6e4b)
- model: support search/filter with list endpoints (#559) (7b17393)
- model: support watch latest model and
order_by
for list endpoints (#586) (1a5e48c) - prediction: implement sync/async prediction records (#555) (8d58eda)
- ray: support containerized model deployment (#529) (4dcab05)
- ray: support custom accelerator type (#547) (f0cc0d7)
Bug Fixes
- acl: fix wrong type name (#560) (89d09a5)
- dockerfile: update deploy config yaml path (#590) (ee369e0)
- model: fix missing package in test models (#552) (a28a21b)
- ray: check CDI availability for model container (#538) (28bad42)
- server: add missing message size option (#597) (d0a0aac)
- service: fix list model version pagination (#569) (d8fb04a)
- service: fix list model version return list size (#556) (9b69f9c)
v0.23.0-alpha
v0.22.0-alpha
0.22.0-alpha (2024-02-20)
⚠ BREAKING CHANGES
- triton: deprecate triton inference server (#512)
Features
Bug Fixes
- cmd,pkg: refactor codebase to align with
golanci-linter
checks (#506) (b213812) - handler: fix multipart request (352a4ae)
- pkg: fix isError and set maxBatchSize to 0 (2adfe5b)
- pkg: fix org model namespace (#510) (f4be09c)
- service: fix workflow retry when deleting (adcbde5)
- service: remove org subscription check (76cd66f)
- usage: add missing org usage collection (239d3f4)
- worker: fix temporal cloud namespace init (#513) (17c5d68)