Back up iclr-debug-medium-speed-full-decay-20260918-v1 decoder-mstok step 26000 and five-seed evaluation
Browse files- iclr-debug-medium-speed-full-decay-20260918-v1/decoder-mstok/evaluation/step-26000/EVAL_DONE.json +7 -0
- iclr-debug-medium-speed-full-decay-20260918-v1/decoder-mstok/evaluation/step-26000/random-seed-0.json +0 -0
- iclr-debug-medium-speed-full-decay-20260918-v1/decoder-mstok/evaluation/step-26000/random-seed-0.samples.json +0 -0
- iclr-debug-medium-speed-full-decay-20260918-v1/decoder-mstok/evaluation/step-26000/random-seed-1.json +0 -0
- iclr-debug-medium-speed-full-decay-20260918-v1/decoder-mstok/evaluation/step-26000/random-seed-1.samples.json +0 -0
- iclr-debug-medium-speed-full-decay-20260918-v1/decoder-mstok/evaluation/step-26000/random-seed-2.json +0 -0
- iclr-debug-medium-speed-full-decay-20260918-v1/decoder-mstok/evaluation/step-26000/random-seed-2.samples.json +0 -0
- iclr-debug-medium-speed-full-decay-20260918-v1/decoder-mstok/evaluation/step-26000/random-seed-3.json +0 -0
- iclr-debug-medium-speed-full-decay-20260918-v1/decoder-mstok/evaluation/step-26000/random-seed-3.samples.json +0 -0
- iclr-debug-medium-speed-full-decay-20260918-v1/decoder-mstok/evaluation/step-26000/random-seed-4.json +0 -0
- iclr-debug-medium-speed-full-decay-20260918-v1/decoder-mstok/evaluation/step-26000/random-seed-4.samples.json +0 -0
- iclr-debug-medium-speed-full-decay-20260918-v1/decoder-mstok/evaluation/step-26000/resource-provenance.json +14 -0
- iclr-debug-medium-speed-full-decay-20260918-v1/decoder-mstok/evaluation/step-26000/summary.json +35 -0
- iclr-debug-medium-speed-full-decay-20260918-v1/decoder-mstok/exports/step-26000/ncp.pt +3 -0
- iclr-debug-medium-speed-full-decay-20260918-v1/decoder-mstok/exports/step-26000/vqvae-teacher.pt +3 -0
- iclr-debug-medium-speed-full-decay-20260918-v1/decoder-mstok/exports/step-26000/vqvae.pt +3 -0
- iclr-debug-medium-speed-full-decay-20260918-v1/decoder-mstok/milestone-iter-26000.pt +3 -0
iclr-debug-medium-speed-full-decay-20260918-v1/decoder-mstok/evaluation/step-26000/EVAL_DONE.json
ADDED
|
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"step": 26000,
|
| 3 |
+
"reference_dtype": "float32",
|
| 4 |
+
"sampling": "supplied-level0-untruncated",
|
| 5 |
+
"concurrent_training": true,
|
| 6 |
+
"resource_provenance": "/home/ishank/mstok-results/medium-step26000-eval-20260919-v1/provenance-retry-7g.json"
|
| 7 |
+
}
|
iclr-debug-medium-speed-full-decay-20260918-v1/decoder-mstok/evaluation/step-26000/random-seed-0.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
iclr-debug-medium-speed-full-decay-20260918-v1/decoder-mstok/evaluation/step-26000/random-seed-0.samples.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
iclr-debug-medium-speed-full-decay-20260918-v1/decoder-mstok/evaluation/step-26000/random-seed-1.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
iclr-debug-medium-speed-full-decay-20260918-v1/decoder-mstok/evaluation/step-26000/random-seed-1.samples.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
iclr-debug-medium-speed-full-decay-20260918-v1/decoder-mstok/evaluation/step-26000/random-seed-2.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
iclr-debug-medium-speed-full-decay-20260918-v1/decoder-mstok/evaluation/step-26000/random-seed-2.samples.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
iclr-debug-medium-speed-full-decay-20260918-v1/decoder-mstok/evaluation/step-26000/random-seed-3.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
iclr-debug-medium-speed-full-decay-20260918-v1/decoder-mstok/evaluation/step-26000/random-seed-3.samples.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
iclr-debug-medium-speed-full-decay-20260918-v1/decoder-mstok/evaluation/step-26000/random-seed-4.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
iclr-debug-medium-speed-full-decay-20260918-v1/decoder-mstok/evaluation/step-26000/random-seed-4.samples.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
iclr-debug-medium-speed-full-decay-20260918-v1/decoder-mstok/evaluation/step-26000/resource-provenance.json
ADDED
|
@@ -0,0 +1,14 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"source_checkpoint": "/home/ishank/mstok-results/iclr-debug-medium-speed-full-decay-20260918-v1/decoder-mstok/checkpoint-iter-26000.pt",
|
| 3 |
+
"retained_checkpoint": "/home/ishank/mstok-results/medium-step26000-eval-20260919-v1/checkpoint-iter-26000.pt",
|
| 4 |
+
"bytes": 7538347441,
|
| 5 |
+
"inode": 8301001,
|
| 6 |
+
"source_revision": "4e22b4740368661b414f176673c2f099079047eb",
|
| 7 |
+
"gpu": 1,
|
| 8 |
+
"allocator_limit_gib": 7,
|
| 9 |
+
"minimum_driver_free_gib": 8,
|
| 10 |
+
"concurrent_training": true,
|
| 11 |
+
"original_training_child": 191800,
|
| 12 |
+
"evaluation_script": "evaluation/eval_random_gen_ppl_checkpoint.py",
|
| 13 |
+
"protocol": "unchanged: five seeds x 128 samples, batch 4, FP32 GPT-2 Large, no TF32, untruncated"
|
| 14 |
+
}
|
iclr-debug-medium-speed-full-decay-20260918-v1/decoder-mstok/evaluation/step-26000/summary.json
ADDED
|
@@ -0,0 +1,35 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"protocol": {
|
| 3 |
+
"sampling": "random",
|
| 4 |
+
"temperature": 1.0,
|
| 5 |
+
"top_k": 0,
|
| 6 |
+
"top_p": 1.0,
|
| 7 |
+
"n_provided_levels": 1,
|
| 8 |
+
"n_levels": 16,
|
| 9 |
+
"document_aware_validation_sampling": true
|
| 10 |
+
},
|
| 11 |
+
"reference_model": "gpt2-large",
|
| 12 |
+
"checkpoint": "/home/ishank/mstok-results/iclr-debug-medium-speed-full-decay-20260918-v1/decoder-mstok/exports/step-26000/ncp.pt",
|
| 13 |
+
"checkpoint_step": 26000,
|
| 14 |
+
"seeds": [
|
| 15 |
+
0,
|
| 16 |
+
1,
|
| 17 |
+
2,
|
| 18 |
+
3,
|
| 19 |
+
4
|
| 20 |
+
],
|
| 21 |
+
"samples_per_seed": 128,
|
| 22 |
+
"total_scored_samples": 640,
|
| 23 |
+
"total_scored_tokens": 161224,
|
| 24 |
+
"per_seed_mean_ppl": {
|
| 25 |
+
"0": 93.34529710608669,
|
| 26 |
+
"1": 84.15464632323328,
|
| 27 |
+
"2": 86.59177427778688,
|
| 28 |
+
"3": 89.69413191725255,
|
| 29 |
+
"4": 81.73470120711265
|
| 30 |
+
},
|
| 31 |
+
"mean_gen_ppl": 87.1041101662944,
|
| 32 |
+
"gen_ppl_se": 2.0427065952185726,
|
| 33 |
+
"mean_of_seed_median_ppl": 83.26203308105468,
|
| 34 |
+
"mean_entropy_nats": 4.407021080025564
|
| 35 |
+
}
|
iclr-debug-medium-speed-full-decay-20260918-v1/decoder-mstok/exports/step-26000/ncp.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1ca352a29d3f46f545d30411c9b83e61bfe05a74bbee83137d1c067316d89c73
|
| 3 |
+
size 1313738678
|
iclr-debug-medium-speed-full-decay-20260918-v1/decoder-mstok/exports/step-26000/vqvae-teacher.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c02b426902e3544eafa32f3b863115e19d7836fdc5a7a0603bda24a16a3ebfe3
|
| 3 |
+
size 1300865317
|
iclr-debug-medium-speed-full-decay-20260918-v1/decoder-mstok/exports/step-26000/vqvae.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6b94ec62ede21b32b74c6a363dd537e94812956a14152e232e88a8731fe2acb3
|
| 3 |
+
size 1300864277
|
iclr-debug-medium-speed-full-decay-20260918-v1/decoder-mstok/milestone-iter-26000.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d475edb413721ffa054d91c97d550ec9c9ba995b2af805c8584d3f7ad6bd4197
|
| 3 |
+
size 7538347441
|