yiyexy commited on
Commit
2e2dd7b
·
verified ·
1 Parent(s): ffd8845

Model card: drop benchmark count from codec highlight

Browse files
Files changed (2) hide show
  1. README.md +1 -1
  2. README_CN.md +1 -1
README.md CHANGED
@@ -29,7 +29,7 @@ LLaVA-OneVision-2.5-4B-Instruct is a 4B vision-language model for single-image,
29
  ## Highlights
30
 
31
  - **Offline and streaming in one model.** `apply_chat_template(..., streaming=True)` switches mode; eight protocol tokens carry the per-frame decision to stay silent, respond or caption. Nothing else about the input changes.
32
- - **Codec-aligned video.** Besides uniform frame sampling, *codec mode* uses the encoder's bit-cost to pack the informative regions of a video into a handful of canvases. It beats frame sampling on **17 of 21** video benchmarks.
33
  - **9× spatial compression.** A patch-16 vision tower with 3×3 merging gives one visual token per 48×48 pixels, so long videos fit comfortably in the 262k context.
34
 
35
  ## Performance
 
29
  ## Highlights
30
 
31
  - **Offline and streaming in one model.** `apply_chat_template(..., streaming=True)` switches mode; eight protocol tokens carry the per-frame decision to stay silent, respond or caption. Nothing else about the input changes.
32
+ - **Codec-aligned video.** Besides uniform frame sampling, *codec mode* uses the encoder's bit-cost to pack the informative regions of a video into a handful of canvases.
33
  - **9× spatial compression.** A patch-16 vision tower with 3×3 merging gives one visual token per 48×48 pixels, so long videos fit comfortably in the 262k context.
34
 
35
  ## Performance
README_CN.md CHANGED
@@ -29,7 +29,7 @@ LLaVA-OneVision-2.5-4B-Instruct 是一个 4B 视觉语言模型,支持单图
29
  ## 亮点
30
 
31
  - **离线与流式合一。** `apply_chat_template(..., streaming=True)` 即可切换模式;八个协议 token 承载每一帧"沉默 / 回答 / caption"的决策,输入格式的其余部分完全相同。
32
- - **Codec 对齐的视频输入。** 除均匀抽帧外,**codec 模式**利用编码器的 bit-cost,把视频中信息量大的区域打包进少量画布。在 21 个视频 bench 中,它有 **17 个**优于抽帧。
33
  - **9 倍空间压缩。** patch 16 的视觉塔加 3×3 合并,每 48×48 像素对应一个视觉 token,长视频可以从容放进 262k 上下文。
34
 
35
  ## 评测结果
 
29
  ## 亮点
30
 
31
  - **离线与流式合一。** `apply_chat_template(..., streaming=True)` 即可切换模式;八个协议 token 承载每一帧"沉默 / 回答 / caption"的决策,输入格式的其余部分完全相同。
32
+ - **Codec 对齐的视频输入。** 除均匀抽帧外,**codec 模式**利用编码器的 bit-cost,把视频中信息量大的区域打包进少量画布。
33
  - **9 倍空间压缩。** patch 16 的视觉塔加 3×3 合并,每 48×48 像素对应一个视觉 token,长视频可以从容放进 262k 上下文。
34
 
35
  ## 评测结果