infra: prepare for release 0.2.17 [generated] [skip ci]

Signed-off-by: Aaron Pham <[email protected]>
bentoml · Aug 8, 2023 · 126491f · 126491f
1 parent cb6f3aa
commit 126491f
Show file tree

Hide file tree

Showing 5 changed files with 21 additions and 9 deletions.
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -18,6 +18,25 @@ This changelog is managed by towncrier and is compiled at release time.
 
 <!-- towncrier release notes start -->
 
+## [0.2.17](https://github.com/bentoml/openllm/tree/v0.2.17)
+
+### Changes
+
+- Updated loading logics for PyTorch and vLLM where it will check for initialized parameters after placing to correct devices
+
+  Added xformers to base container for requirements on vLLM-based container
+  [#185](https://github.com/bentoml/openllm/issues/185)
+
+
+### Features
+
+- Importing models now won't load into memory if it is a remote ID. Note that for GPTQ and local model the behaviour is unchanged.
+
+  Fixes that when there is one GPU, we ensure to call `to('cuda')` to place the model onto the memory. Note that the GPU must have
+  enough VRAM to offload this model onto the GPU.
+  [#183](https://github.com/bentoml/openllm/issues/183)
+
+
 ## [0.2.16](https://github.com/bentoml/openllm/tree/v0.2.16)
 No significant changes.
 

diff --git a/changelog.d/183.feature.md b/changelog.d/183.feature.md
diff --git a/changelog.d/185.change.md b/changelog.d/185.change.md
diff --git a/package.json b/package.json
@@ -1,6 +1,6 @@
 {
   "name": "openllm",
-  "version": "0.2.17.dev0",
+  "version": "0.2.17",
   "description": "OpenLLM: Operating LLMs in production",
   "repository": "[email protected]:bentoml/OpenLLM.git",
   "author": "Aaron Pham <[email protected]>",

diff --git a/src/openllm_js/package.json b/src/openllm_js/package.json
@@ -1,6 +1,6 @@
 {
   "name": "openllm",
-  "version": "0.2.17.dev0",
+  "version": "0.2.17",
   "description": "TS/JS binding for OpenLLM",
   "type": "module",
   "repository": "[email protected]:bentoml/OpenLLM.git",