You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Resolved merge conflicts by integrating both feature sets:
- Merged embedding model selection (bert/qwen3/gemma) with HNSW indexing
- Combined flash-attn feature with cuda as default in Cargo.toml
- Integrated both embedding_model and HNSW configuration options
- Merged test targets from both branches in rust.mk
- Updated Cargo.lock with merged dependencies
All features from both branches are now available:
- Semantic cache supports both embedding model selection and HNSW indexing
- Rust builds support both CUDA (default) and Flash Attention (optional)
- Test infrastructure includes both Rust unit tests and CI-friendly builds
Signed-off-by: Huamin Chen <[email protected]>
Copy file name to clipboardExpand all lines: README.md
+2-1Lines changed: 2 additions & 1 deletion
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -16,6 +16,7 @@
16
16
17
17
*Latest News* 🔥
18
18
19
+
-[2025/10/21] We announced the [2025 Q4 Roadmap: Journey to Iris](https://vllm-semantic-router.com/blog/q4-roadmap-iris) 📅.
19
20
-[2025/10/16] We established the [vLLM Semantic Router Youtube Channel](https://www.youtube.com/@vLLMSemanticRouter) ✨.
20
21
-[2025/10/15] We announced the [vLLM Semantic Router Dashboard](https://www.youtube.com/watch?v=E2IirN8PsFw) 🚀.
21
22
-[2025/10/12] Our paper [When to Reason: Semantic Router for vLLM](https://arxiv.org/abs/2510.08731) accepted by NeurIPS 2025 MLForSys 🧠.
@@ -75,7 +76,7 @@ Detect PII in the prompt, avoiding sending PII to the LLM so as to protect the p
75
76
76
77
#### Prompt guard
77
78
78
-
Detect if the prompt is a jailbreak prompt, avoiding sending jailbreak prompts to the LLM so as to prevent the LLM from misbehaving.
79
+
Detect if the prompt is a jailbreak prompt, avoiding sending jailbreak prompts to the LLM so as to prevent the LLM from misbehaving. Can be configured globally or at the category level for fine-grained security control.
0 commit comments