Document the 640x384 detector size

README.md

@@ -32,6 +32,7 @@ an SL2619.
 | File | Purpose |
 |---|---|
 | `ppocr_det_800x608.vmfb` | Detection (DBNet), static 800×608 bf16 input |
+| `ppocr_det_640x384.vmfb` | Detection (DBNet), static 640×384 bf16 input, for wide/16:9 sources |
 | `rec_buckets/rec_w320.vmfb` | Recognition, 48×320 lines |
 | `rec_buckets/rec_w640.vmfb` | Recognition, 48×640 lines |
 | `rec_buckets/rec_w1280.vmfb` | Recognition, 48×1280 lines |
@@ -41,6 +42,14 @@ an SL2619.
 | `ppocr_rec_dynamic.onnx` | fp32 recognition, CPU reference for accuracy checks |
 | `samples/sample.jpg` | Sample café menu card, 10 text lines |
 
+### Two detector sizes
+
+DBNet's stride-32 backbone requires input dims that are multiples of 32.
+800×608 is the default, matched to portrait documents; 640×384 serves wide/16:9
+sources — it is 640×360 rounded *up* to the next multiple of 32, so letterboxed
+16:9 content is padded rather than cropped. Select it in the demo with
+`--det-hw 640 384`.
+
 ### Why four recognition models
 
 Recognition input width is static per vmfb. Each detected line is routed to the