joint-main-300k-20260922-v1-conditional: final raw adjacency-5 full test
Browse files- mstok-matched-xsum/joint-main-300k-20260922-v1-conditional/step-50002/final-test-raw-adjacency5/command.json +23 -0
- mstok-matched-xsum/joint-main-300k-20260922-v1-conditional/step-50002/final-test-raw-adjacency5/comparison.json +62 -0
- mstok-matched-xsum/joint-main-300k-20260922-v1-conditional/step-50002/final-test-raw-adjacency5/metrics.json +28 -0
- mstok-matched-xsum/joint-main-300k-20260922-v1-conditional/step-50002/final-test-raw-adjacency5/predictions.jsonl +0 -0
- mstok-matched-xsum/joint-main-300k-20260922-v1-conditional/step-50002/final-test-raw-adjacency5/source-hashes.json +5 -0
mstok-matched-xsum/joint-main-300k-20260922-v1-conditional/step-50002/final-test-raw-adjacency5/command.json
ADDED
|
@@ -0,0 +1,23 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
[
|
| 2 |
+
"/home/ishank/ncm-xsum-env-20260921/bin/python",
|
| 3 |
+
"-u",
|
| 4 |
+
"scripts/evaluate_matched_xsum.py",
|
| 5 |
+
"--checkpoint",
|
| 6 |
+
"/home/ishank/mstok/experiments/mstok-matched-xsum/joint-main-300k-20260922-v1/joint-main-300k-20260922-v1-conditional/training/checkpoint-50002.pt",
|
| 7 |
+
"--variant",
|
| 8 |
+
"raw",
|
| 9 |
+
"--split",
|
| 10 |
+
"test",
|
| 11 |
+
"--count",
|
| 12 |
+
"11233",
|
| 13 |
+
"--frequency-penalty",
|
| 14 |
+
"0",
|
| 15 |
+
"--adjacency-penalty",
|
| 16 |
+
"5",
|
| 17 |
+
"--adjacency-window",
|
| 18 |
+
"1",
|
| 19 |
+
"--adjacency-decay",
|
| 20 |
+
"1",
|
| 21 |
+
"--output",
|
| 22 |
+
"/home/ishank/mstok/experiments/mstok-matched-xsum/joint-main-300k-20260922-v1/joint-main-300k-20260922-v1-conditional/evaluations/final-test-raw-adjacency5"
|
| 23 |
+
]
|
mstok-matched-xsum/joint-main-300k-20260922-v1-conditional/step-50002/final-test-raw-adjacency5/comparison.json
ADDED
|
@@ -0,0 +1,62 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"step": 50002,
|
| 3 |
+
"split": "test",
|
| 4 |
+
"documents": 11233,
|
| 5 |
+
"variant": "raw",
|
| 6 |
+
"baseline": {
|
| 7 |
+
"step": 50002,
|
| 8 |
+
"variant": "raw",
|
| 9 |
+
"split": "test",
|
| 10 |
+
"documents": 11233,
|
| 11 |
+
"decoding": "greedy; predicted L0; 9 NFE",
|
| 12 |
+
"empty_predictions": 0,
|
| 13 |
+
"l0_code_accuracy": 0.013954420012463277,
|
| 14 |
+
"released_inference_parity": null,
|
| 15 |
+
"self": {
|
| 16 |
+
"rouge1": 29.70586990452856,
|
| 17 |
+
"rouge2": 9.071249240716856,
|
| 18 |
+
"rougeL": 23.364595951105045,
|
| 19 |
+
"gen_ppl": 198.5111993865347
|
| 20 |
+
},
|
| 21 |
+
"oracle": {
|
| 22 |
+
"rouge1": 38.075879837275494,
|
| 23 |
+
"rouge2": 15.59852443009699,
|
| 24 |
+
"rougeL": 33.02291335996958,
|
| 25 |
+
"gen_ppl": 128.99291858057666
|
| 26 |
+
}
|
| 27 |
+
},
|
| 28 |
+
"adjacency5": {
|
| 29 |
+
"step": 50002,
|
| 30 |
+
"variant": "raw",
|
| 31 |
+
"split": "test",
|
| 32 |
+
"documents": 11233,
|
| 33 |
+
"decoding": "greedy; predicted L0; 9 NFE",
|
| 34 |
+
"empty_predictions": 0,
|
| 35 |
+
"l0_code_accuracy": 0.013954420012463277,
|
| 36 |
+
"released_inference_parity": null,
|
| 37 |
+
"token_decoding_penalties": {
|
| 38 |
+
"frequency": 0.0,
|
| 39 |
+
"adjacency": 5.0,
|
| 40 |
+
"window": 1,
|
| 41 |
+
"decay": 1.0
|
| 42 |
+
},
|
| 43 |
+
"self": {
|
| 44 |
+
"rouge1": 30.006417843233375,
|
| 45 |
+
"rouge2": 9.179745116240863,
|
| 46 |
+
"rougeL": 23.587278302179897,
|
| 47 |
+
"gen_ppl": 196.46467730493563
|
| 48 |
+
},
|
| 49 |
+
"oracle": {
|
| 50 |
+
"rouge1": 38.29919657663228,
|
| 51 |
+
"rouge2": 15.680812260251884,
|
| 52 |
+
"rougeL": 33.19819722044507,
|
| 53 |
+
"gen_ppl": 126.74572046044578
|
| 54 |
+
}
|
| 55 |
+
},
|
| 56 |
+
"delta_self": {
|
| 57 |
+
"rouge1": 0.3005479387048169,
|
| 58 |
+
"rouge2": 0.10849587552400664,
|
| 59 |
+
"rougeL": 0.22268235107485168,
|
| 60 |
+
"gen_ppl": -2.046522081599079
|
| 61 |
+
}
|
| 62 |
+
}
|
mstok-matched-xsum/joint-main-300k-20260922-v1-conditional/step-50002/final-test-raw-adjacency5/metrics.json
ADDED
|
@@ -0,0 +1,28 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"step": 50002,
|
| 3 |
+
"variant": "raw",
|
| 4 |
+
"split": "test",
|
| 5 |
+
"documents": 11233,
|
| 6 |
+
"decoding": "greedy; predicted L0; 9 NFE",
|
| 7 |
+
"empty_predictions": 0,
|
| 8 |
+
"l0_code_accuracy": 0.013954420012463277,
|
| 9 |
+
"released_inference_parity": null,
|
| 10 |
+
"token_decoding_penalties": {
|
| 11 |
+
"frequency": 0.0,
|
| 12 |
+
"adjacency": 5.0,
|
| 13 |
+
"window": 1,
|
| 14 |
+
"decay": 1.0
|
| 15 |
+
},
|
| 16 |
+
"self": {
|
| 17 |
+
"rouge1": 30.006417843233375,
|
| 18 |
+
"rouge2": 9.179745116240863,
|
| 19 |
+
"rougeL": 23.587278302179897,
|
| 20 |
+
"gen_ppl": 196.46467730493563
|
| 21 |
+
},
|
| 22 |
+
"oracle": {
|
| 23 |
+
"rouge1": 38.29919657663228,
|
| 24 |
+
"rouge2": 15.680812260251884,
|
| 25 |
+
"rougeL": 33.19819722044507,
|
| 26 |
+
"gen_ppl": 126.74572046044578
|
| 27 |
+
}
|
| 28 |
+
}
|
mstok-matched-xsum/joint-main-300k-20260922-v1-conditional/step-50002/final-test-raw-adjacency5/predictions.jsonl
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
mstok-matched-xsum/joint-main-300k-20260922-v1-conditional/step-50002/final-test-raw-adjacency5/source-hashes.json
ADDED
|
@@ -0,0 +1,5 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"scripts/evaluate_matched_xsum.py": "d64a58d15f4c281e2a906c9485644f9fbc70b5dba594b76b108d05052e570de5",
|
| 3 |
+
"evaluation/decoding_penalties.py": "70193076934ae8bf8cf6937f4d0d9acd46356bf20d770d3a18fdac46e652ddf0",
|
| 4 |
+
"scripts/queue_matched_xsum_final_penalty.py": "773bcb289ff64a5caa8341637f65d020101dcefe55827644625afbe414dc3bb8"
|
| 5 |
+
}
|