OpenGVLab
/

InternVideo2_5_Chat_8B

Video-Text-to-Text

feature-extraction

Model card Files Files and versions Community

ynhe commited on 22 days ago

Commit

7dca794

·

verified ·

1 Parent(s): d1a6d08

Update README.md

Files changed (1) hide show

README.md +4 -4

README.md CHANGED Viewed

@@ -97,10 +97,10 @@ model-index:
 We measured the average inference speed (tokens/s) of generating 1024 new tokens and 5198 (8192-2998) tokens with the context of an video (which takes 2998 tokens) under BF16 precision.
-|Quantization |	Speed (3022 tokens)	| Speed (8192 tokens)|
-|--- |--- |---|
-|BF16 |	33.40 |	31.91 |
-|INT4 | -     | 31.95 |
 ## 🚀 How to use the model

 We measured the average inference speed (tokens/s) of generating 1024 new tokens and 5198 (8192-2998) tokens with the context of an video (which takes 2998 tokens) under BF16 precision.
+|Quantization |	Speed (3022 tokens)	| Speed (8192 tokens) w/o vision| Speed(8192 tokens) w/ vision|
+|--- |--- |---| ---|
+|BF16 |	33.40 |	31.91 | 21.33|
+|INT4 | -     | 31.95 | - |
 ## 🚀 How to use the model