Shorten the checkpoint notes

README.md

@@ -15,7 +15,7 @@ assets/            mt5_tokenizer/ flan_t5_tokenizer/ yolo/yolo11n.onnx mediapipe
 unisign/<name>/    projector.onnx encoder.onnx decoder.onnx prefix.npy meta.json
 spamo/shared/      clip_vit_s2.onnx(+.data) videomae.onnx(+.data) hand_clip.onnx(+.data)
 spamo/<name>/      projector.onnx(+.data) encoder.onnx(+.data) decoder.onnx(+.data) prompt_embeds.npy meta.json
-checkpoints/spamo/ spamo_pjm_ms_dual_bleu5.36.ckpt, <model>/*.ckpt (PyTorch Lightning, not used by signsdk)
+checkpoints/spamo/<model>/*.ckpt (PyTorch Lightning, not used by signsdk)
 ```
 
 ## Models (test-set BLEU-4, ✓ = uploaded)
@@ -45,11 +45,5 @@ BLEU-4 = reported PJM test score. Decoding faithful to the original eval
 
 ## Training checkpoints
 
-PyTorch Lightning checkpoints of the SpaMo and Mod-SpaMo models above (not used by signsdk), for further training.
-`checkpoints/spamo/<model>/` holds the checkpoint behind each table row; the BLEU-4 in a file name is the
-validation score at that epoch, not the test score.
-
-`checkpoints/spamo/spamo_pjm_ms_dual_bleu5.36.ckpt` is a SpaMo checkpoint (PHOENIX-2014T init, PJM MS; epoch 43,
-dev BLEU-4 5.36). Its test run gave the 5.54 listed for `spamo-pjm-phoenix-ms`; which checkpoint the ONNX export
-was made from is not recorded. It is the starting point of the SpaMo-PJM `from_upstream` configs (v26 to v29),
-including the PHOENIX SI models. Moved here from the former `spamo-pjm-checkpoints` repo.
+`checkpoints/spamo/<model>/` holds the PyTorch Lightning checkpoint behind each SpaMo and Mod-SpaMo row above
+(not used by signsdk), for further training. The BLEU-4 in a file name is the validation score at that epoch.