mlboydaisuke commited on
Commit
cf7c86e
·
verified ·
1 Parent(s): af2e7a7

README: keep maxTokens at 1933 or below (the decoder's 2048-position KV cache)

Browse files
Files changed (1) hide show
  1. README.md +1 -0
README.md CHANGED
@@ -134,6 +134,7 @@ The exact, verified recipe is in `assets/recipe.json`. Reference implementations
134
  roughly square or portrait, with text still legible when fit to 640². Very dense small-text scans
135
  (newspaper) want the tiled `crop_mode` vision export (not included here; Base mode only).
136
  - Prompt is fixed to `document parsing` (layout + structured extraction).
 
137
  - License: **MIT** (inherited from `baidu/Unlimited-OCR`).
138
 
139
  *Community port — not affiliated with Apple or baidu.*
 
134
  roughly square or portrait, with text still legible when fit to 640². Very dense small-text scans
135
  (newspaper) want the tiled `crop_mode` vision export (not included here; Base mode only).
136
  - Prompt is fixed to `document parsing` (layout + structured extraction).
137
+ - Keep the `maxTokens` of `KitDocReader.read(imageAt:maxTokens:)` at 1933 or below; the default of 1024, which `CoreAI.read` uses, fits. The decoder has a fixed KV cache of 2048 positions, shared by the image prompt (115 positions) and the generated tokens (one each), and a page that needs more tokens than `maxTokens` is cut there. CoreAIKit 0.7.3 and earlier do not stop at the end of the cache, and what they output past it has not been checked.
138
  - License: **MIT** (inherited from `baidu/Unlimited-OCR`).
139
 
140
  *Community port — not affiliated with Apple or baidu.*