aboutsummaryrefslogtreecommitdiff
path: root/README.md
diff options
context:
space:
mode:
authorOnur Gündüz <ogunduz@gmail.com>2026-06-14 16:06:18 +0200
committerOnur Gündüz <ogunduz@gmail.com>2026-06-14 16:41:53 +0200
commitcb1f049e710cf8704ae29811b3058142c7f64840 (patch)
tree85df308a574fbc3408cd88cd82471896337c8669 /README.md
parentbb41b40bf881a1c59ee0d6153bcb0bb2f7a0044b (diff)
downloadffwd-cb1f049e710cf8704ae29811b3058142c7f64840.tar.gz
ffwd-cb1f049e710cf8704ae29811b3058142c7f64840.zip
all: add Qwen3-Embedding-4B + fix terminal token
Diffstat (limited to 'README.md')
-rw-r--r--README.md6
1 files changed, 4 insertions, 2 deletions
diff --git a/README.md b/README.md
index 59ea6b9..f056d40 100644
--- a/README.md
+++ b/README.md
@@ -81,8 +81,10 @@ Late-interaction reranking endpoint:
- `POST /v1/rerank` — late-interaction MaxSim reranking
-For `Qwen3-Embedding-0.6B`, `/v1/embeddings` accepts `dimensions` from 32
-to 1024. Truncated Matryoshka embeddings are re-normalized before encoding.
+For Qwen3-Embedding models, `/v1/embeddings` accepts `dimensions` from 32 up to
+the model's output size (1024 for `Qwen3-Embedding-0.6B`, 2560 for
+`Qwen3-Embedding-4B`). Truncated Matryoshka embeddings are re-normalized before
+encoding.
Example: