chironYu commited on
Commit
25b6967
·
verified ·
1 Parent(s): d48bb1e

Card: the fixed split is read by fixed_split_eval.py of the code repository

Browse files
Files changed (1) hide show
  1. README.md +19 -5
README.md CHANGED
@@ -163,11 +163,23 @@ python contrast_experiment/w9/eval/k10_eval.py table scores.json
163
  ```
164
 
165
  `cells` takes the 162 five-fold checkpoints of `rank_ckpt_plan.json` and `i2bce_sel_pool.json`;
166
- `--only <regex>` keeps a subset, for example `--only "_d512$"` for the 512-sentence runs. The
167
- script does not score the fixed-split towers. In its table the five families of the main
168
- comparison go by the paper's names (`I-CE`, `CE`, `SimCLR`, `VICReg`, `BYOL`) and every other
169
- family by its plan key without the fold; a family with one fold, the DINO-style control, is
170
- printed with its spread over the ten draws.
 
 
 
 
 
 
 
 
 
 
 
 
171
 
172
  ## The 512-sentence packs and the `_d512` runs
173
 
@@ -214,6 +226,8 @@ dataset keeps the first packs as `wscan_gal_rev_rawprefix.npz`; the workers read
214
 
215
  ## Changes
216
 
 
 
217
  - 2026-11-11, later that day. 62 checkpoints were added, so that every tower the paper reads is
218
  here: the DINO-style control (fold 0), eight fixed-split runs at their selected epochs (CE and
219
  I-CE trained at 2,048 sentences, the SimCLR-style tower, I + SimCLR, a+u @ views and the three
 
163
  ```
164
 
165
  `cells` takes the 162 five-fold checkpoints of `rank_ckpt_plan.json` and `i2bce_sel_pool.json`;
166
+ `--only <regex>` keeps a subset, for example `--only "_d512$"` for the 512-sentence runs. In its
167
+ table the five families of the main comparison go by the paper's names (`I-CE`, `CE`, `SimCLR`,
168
+ `VICReg`, `BYOL`) and every other family by its plan key without the fold; a family with one
169
+ fold, the DINO-style control, is printed with its spread over the ten draws.
170
+
171
+ The fixed split has one command of its own:
172
+
173
+ ```bash
174
+ python contrast_experiment/w9/eval/fixed_split_eval.py --ckpt-dir checkpoints
175
+ ```
176
+
177
+ It reads every run of `fixed_split_ckpt_plan.json` at its selected checkpoint and prints each
178
+ reading beside the number the paper prints: Table 3, the gallery-aimed settings of Section 6.2
179
+ and Table A7 on the one 4,096-sentence anchor draw of the fixed split, the SimCLR-style test
180
+ curve of Appendix L, the latent queries of Figure A1, and, through `k10_eval.py`, the towers of
181
+ Figure 9 against the ten draws. A near tie can move a cell by one query of 204 (0.005), and the
182
+ collapsed a+u @ gallery row by about 0.01.
183
 
184
  ## The 512-sentence packs and the `_d512` runs
185
 
 
226
 
227
  ## Changes
228
 
229
+ - 2026-11-11, at night. The code repository gained `fixed_split_eval.py`, which reads the
230
+ fixed-split checkpoints as the paper does.
231
  - 2026-11-11, later that day. 62 checkpoints were added, so that every tower the paper reads is
232
  here: the DINO-style control (fold 0), eight fixed-split runs at their selected epochs (CE and
233
  I-CE trained at 2,048 sentences, the SimCLR-style tower, I + SimCLR, a+u @ views and the three