A high-throughput and memory-efficient inference and serving engine for LLMs
Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git
Synced Nov 16, 05:45 UTC
[Doc] Fix typos in docs (#10636)
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
Cyrus Leung committed
Nov 26, 2024 at 02:15 UTC
1b583cfefad4ffa030bda1c1265aec6e7755a6d2
2 files changed
+2
-2
docs/source/models/supported_models.rst
+1
-1
@@ -365,7 +365,7 @@ Text Embedding
365
366
.. note::
367
Unlike base Qwen2, :code:`Alibaba-NLP/gte-Qwen2-7B-instruct` uses bi-directional attention.
368
- You can set `--hf-overrides '{"is_causal": false}'` to change the attention mask accordingly.
368
+ You can set :code:`--hf-overrides '{"is_causal": false}'` to change the attention mask accordingly.
369
370
On the other hand, its 1.5B variant (:code:`Alibaba-NLP/gte-Qwen2-1.5B-instruct`) uses causal attention
371
despite being described otherwise on its model card.
docs/source/serving/compatibility_matrix.rst
+1
-1
@@ -393,7 +393,7 @@ Feature x Hardware
393
- ✅
394
- ✅
395
- ✅
396
- - ✗
396
+ - ?
397
* - :abbr:`enc-dec (Encoder-Decoder Models)`
398
- ✅
399
- ✅