appvoid commited on
Commit
37b36bc
·
verified ·
1 Parent(s): 2191f5c

Qwen90M step 285000; complete resumable checkpoint

Browse files
complete.json CHANGED
@@ -1,18 +1,18 @@
1
  {
2
- "step": 280248,
3
  "files": {
4
  "config.json": "cbc72137da3dae65085a2a8c39ad7634d16bf92d785ca00d283ee82fe32b37e7",
5
  "generation_config.json": "54ed0d19f6226888c91157964ede8a892bd5ae175bfad85336152b3d9b6cd7c4",
6
- "model.safetensors": "ab3d534f71afc536f11eacb5cac49b41552e56c521fecc7b301c5b0f3ba211e7",
7
  "tokenization_byte.py": "279b9990f77b1eaedc7c1184fe13d60273a713ead856200f1fb9ae323c7b1e03",
8
  "tokenizer_config.json": "0ecbd4fd39e35125800d806af981ba0cacd8a7ba5f02c9e729c4ecc54af6d300",
9
  "special_tokens_map.json": "20d4b94d0e657c3f44d9123522c106e7315b6c1a1b6d259478f17bd2918b1263",
10
  "byte_vocab.json": "57a1cd2e70dd5b7cc6fd2eaa12ebfcd173c1620fa5d45c2c6282511c29b6a5f6",
11
- "trainer.pt": "93080e3d08484f8e03c44bc9ba9acf6621eb9dc2cb2663dd5f9becbb62b890ee",
12
- "training_config.json": "95ba9e002790d11959d1191b36ed994bfa10a6da51c22c772e143bec517a9783",
13
- "performance.json": "7ab839f3b8798329051e29b39e08bd08f8678f49ed79d09e3b70b9de78a4765c",
14
- "shard_scheduler.json": "1a27a42e8cdc3d5af3957968177c27079c751995153e489c49bae7322e8344e1",
15
- "coverage.json": "c8401a996db8071c49afdc637c6460b684b49b8ba1e5678dd9298a5b147168b1",
16
  "README.md": "9ce0f9efacbc46177a7e818b6aed274d84e7145ff944b5c7d6ecd5d33b9d4419"
17
  },
18
  "signature": "9fee47918905b0750bee87876aa424f021ccc4507ee8db22a11d01fbc6109063"
 
1
  {
2
+ "step": 285000,
3
  "files": {
4
  "config.json": "cbc72137da3dae65085a2a8c39ad7634d16bf92d785ca00d283ee82fe32b37e7",
5
  "generation_config.json": "54ed0d19f6226888c91157964ede8a892bd5ae175bfad85336152b3d9b6cd7c4",
6
+ "model.safetensors": "8ac057f2166a1e4db0ce4a9415d89891b9b8ddcc52297129f86c08f788b2758b",
7
  "tokenization_byte.py": "279b9990f77b1eaedc7c1184fe13d60273a713ead856200f1fb9ae323c7b1e03",
8
  "tokenizer_config.json": "0ecbd4fd39e35125800d806af981ba0cacd8a7ba5f02c9e729c4ecc54af6d300",
9
  "special_tokens_map.json": "20d4b94d0e657c3f44d9123522c106e7315b6c1a1b6d259478f17bd2918b1263",
10
  "byte_vocab.json": "57a1cd2e70dd5b7cc6fd2eaa12ebfcd173c1620fa5d45c2c6282511c29b6a5f6",
11
+ "trainer.pt": "f77cbd7043301ba03a27d0fc3fb559b972721c8fb60167219ee0788286252c59",
12
+ "training_config.json": "2122ad6f8105987f9b87dd55d13b632084853908566b081c829ffe801259e2bf",
13
+ "performance.json": "f89f82c3f4c44a52fcd532346ea71580902d757e59e1961e9b2bed0030259dba",
14
+ "shard_scheduler.json": "5316864976580fabdbfc5b4749a2a377f1a44f787f1e27a21ee310ec16436feb",
15
+ "coverage.json": "e35c09a84496ab9947c7e21ef4a25e40c3c17be24bb510256b35b90dd258081c",
16
  "README.md": "9ce0f9efacbc46177a7e818b6aed274d84e7145ff944b5c7d6ecd5d33b9d4419"
17
  },
18
  "signature": "9fee47918905b0750bee87876aa424f021ccc4507ee8db22a11d01fbc6109063"
coverage.json CHANGED
@@ -1,107 +1,107 @@
1
  {
2
- "step": 280248,
3
  "stats": {
4
- "climbmix/blocks": 10088928,
5
- "climbmix/targets": 16865599592,
6
- "climbmix/input_tokens": 16866668333,
7
- "ultra_qa/blocks": 3783348,
8
- "ultra_qa/targets": 6172241217,
9
- "ultra_qa/input_tokens": 6172253060,
10
- "cortex/blocks": 3783348,
11
- "cortex/targets": 143908822,
12
- "cortex/input_tokens": 6013958320,
13
- "ultra_style/blocks": 3783348,
14
- "ultra_style/targets": 5984022238,
15
- "ultra_style/input_tokens": 5984336893,
16
- "openmath/blocks": 2522232,
17
- "openmath/targets": 3897349251,
18
- "openmath/input_tokens": 3898648602,
19
- "rewrite6/blocks": 1261116,
20
- "rewrite6/targets": 2067382557,
21
- "rewrite6/input_tokens": 2069649291
22
  },
23
  "sources": {
24
  "climbmix": {
25
- "page": 1381,
26
- "row": 2390,
27
  "start": {
28
- "file": 66,
29
  "reader": {
30
  "format": "parquet-row-v1",
31
- "row_group": 70,
32
  "row": 1024
33
  }
34
  },
35
  "epoch": 0,
36
- "offset": 2048,
37
- "documents": 5658966
38
  },
39
  "rewrite6": {
40
- "page": 36,
41
- "row": 797,
42
  "start": {
43
  "file": 0,
44
  "reader": {
45
  "format": "jsonl-row-v1",
46
- "row": 147456
47
  }
48
  },
49
  "epoch": 15,
50
- "offset": 0,
51
- "documents": 3229313
52
  },
53
  "openmath": {
54
- "page": 863,
55
- "row": 3211,
56
  "start": {
57
  "file": 8,
58
  "reader": {
59
  "format": "parquet-row-v1",
60
- "row_group": 41,
61
- "row": 648
62
  }
63
  },
64
  "epoch": 0,
65
  "offset": 2048,
66
- "documents": 3538059
67
  },
68
  "ultra_qa": {
69
- "page": 373,
70
- "row": 1465,
71
  "start": {
72
  "file": 2,
73
  "reader": {
74
  "format": "parquet-row-v1",
75
  "row_group": 7,
76
- "row": 36665
77
  }
78
  },
79
  "epoch": 0,
80
  "offset": 2048,
81
- "documents": 1529273
82
  },
83
  "ultra_style": {
84
- "page": 570,
85
- "row": 1617,
86
  "start": {
87
  "file": 3,
88
  "reader": {
89
  "format": "parquet-row-v1",
90
  "row_group": 3,
91
- "row": 22523
92
  }
93
  },
94
  "epoch": 0,
95
  "offset": 2048,
96
- "documents": 2336337
97
  },
98
  "cortex": {
99
- "page": 73374,
100
- "row": 54,
101
  "start": null,
102
  "epoch": 0,
103
  "offset": 0,
104
- "documents": 9391926
105
  }
106
  }
107
  }
 
1
  {
2
+ "step": 285000,
3
  "stats": {
4
+ "climbmix/blocks": 10260000,
5
+ "climbmix/targets": 17151447815,
6
+ "climbmix/input_tokens": 17152534864,
7
+ "ultra_qa/blocks": 3847500,
8
+ "ultra_qa/targets": 6276806703,
9
+ "ultra_qa/input_tokens": 6276818734,
10
+ "cortex/blocks": 3847500,
11
+ "cortex/targets": 146351608,
12
+ "cortex/input_tokens": 6115948834,
13
+ "ultra_style/blocks": 3847500,
14
+ "ultra_style/targets": 6085396148,
15
+ "ultra_style/input_tokens": 6085715990,
16
+ "openmath/blocks": 2565000,
17
+ "openmath/targets": 3963414207,
18
+ "openmath/input_tokens": 3964735574,
19
+ "rewrite6/blocks": 1282500,
20
+ "rewrite6/targets": 2102512928,
21
+ "rewrite6/input_tokens": 2104817984
22
  },
23
  "sources": {
24
  "climbmix": {
25
+ "page": 1405,
26
+ "row": 588,
27
  "start": {
28
+ "file": 68,
29
  "reader": {
30
  "format": "parquet-row-v1",
31
+ "row_group": 1,
32
  "row": 1024
33
  }
34
  },
35
  "epoch": 0,
36
+ "offset": 0,
37
+ "documents": 5755468
38
  },
39
  "rewrite6": {
40
+ "page": 49,
41
+ "row": 2128,
42
  "start": {
43
  "file": 0,
44
  "reader": {
45
  "format": "jsonl-row-v1",
46
+ "row": 200704
47
  }
48
  },
49
  "epoch": 15,
50
+ "offset": 2048,
51
+ "documents": 3283892
52
  },
53
  "openmath": {
54
+ "page": 878,
55
+ "row": 1736,
56
  "start": {
57
  "file": 8,
58
  "reader": {
59
  "format": "parquet-row-v1",
60
+ "row_group": 103,
61
+ "row": 88
62
  }
63
  },
64
  "epoch": 0,
65
  "offset": 2048,
66
+ "documents": 3598024
67
  },
68
  "ultra_qa": {
69
+ "page": 379,
70
+ "row": 2700,
71
  "start": {
72
  "file": 2,
73
  "reader": {
74
  "format": "parquet-row-v1",
75
  "row_group": 7,
76
+ "row": 61241
77
  }
78
  },
79
  "epoch": 0,
80
  "offset": 2048,
81
+ "documents": 1555084
82
  },
83
  "ultra_style": {
84
+ "page": 579,
85
+ "row": 4081,
86
  "start": {
87
  "file": 3,
88
  "reader": {
89
  "format": "parquet-row-v1",
90
  "row_group": 3,
91
+ "row": 59387
92
  }
93
  },
94
  "epoch": 0,
95
  "offset": 2048,
96
+ "documents": 2375665
97
  },
98
  "cortex": {
99
+ "page": 74620,
100
+ "row": 66,
101
  "start": null,
102
  "epoch": 0,
103
  "offset": 0,
104
+ "documents": 9551426
105
  }
106
  }
107
  }
latest.json CHANGED
@@ -1 +1 @@
1
- {"step": 280248, "files": {"config.json": "cbc72137da3dae65085a2a8c39ad7634d16bf92d785ca00d283ee82fe32b37e7", "generation_config.json": "54ed0d19f6226888c91157964ede8a892bd5ae175bfad85336152b3d9b6cd7c4", "model.safetensors": "ab3d534f71afc536f11eacb5cac49b41552e56c521fecc7b301c5b0f3ba211e7", "tokenization_byte.py": "279b9990f77b1eaedc7c1184fe13d60273a713ead856200f1fb9ae323c7b1e03", "tokenizer_config.json": "0ecbd4fd39e35125800d806af981ba0cacd8a7ba5f02c9e729c4ecc54af6d300", "special_tokens_map.json": "20d4b94d0e657c3f44d9123522c106e7315b6c1a1b6d259478f17bd2918b1263", "byte_vocab.json": "57a1cd2e70dd5b7cc6fd2eaa12ebfcd173c1620fa5d45c2c6282511c29b6a5f6", "trainer.pt": "93080e3d08484f8e03c44bc9ba9acf6621eb9dc2cb2663dd5f9becbb62b890ee", "training_config.json": "95ba9e002790d11959d1191b36ed994bfa10a6da51c22c772e143bec517a9783", "performance.json": "7ab839f3b8798329051e29b39e08bd08f8678f49ed79d09e3b70b9de78a4765c", "shard_scheduler.json": "1a27a42e8cdc3d5af3957968177c27079c751995153e489c49bae7322e8344e1", "coverage.json": "c8401a996db8071c49afdc637c6460b684b49b8ba1e5678dd9298a5b147168b1", "README.md": "9ce0f9efacbc46177a7e818b6aed274d84e7145ff944b5c7d6ecd5d33b9d4419"}, "signature": "9fee47918905b0750bee87876aa424f021ccc4507ee8db22a11d01fbc6109063", "folder": "step-000000280248"}
 
1
+ {"step": 285000, "files": {"config.json": "cbc72137da3dae65085a2a8c39ad7634d16bf92d785ca00d283ee82fe32b37e7", "generation_config.json": "54ed0d19f6226888c91157964ede8a892bd5ae175bfad85336152b3d9b6cd7c4", "model.safetensors": "8ac057f2166a1e4db0ce4a9415d89891b9b8ddcc52297129f86c08f788b2758b", "tokenization_byte.py": "279b9990f77b1eaedc7c1184fe13d60273a713ead856200f1fb9ae323c7b1e03", "tokenizer_config.json": "0ecbd4fd39e35125800d806af981ba0cacd8a7ba5f02c9e729c4ecc54af6d300", "special_tokens_map.json": "20d4b94d0e657c3f44d9123522c106e7315b6c1a1b6d259478f17bd2918b1263", "byte_vocab.json": "57a1cd2e70dd5b7cc6fd2eaa12ebfcd173c1620fa5d45c2c6282511c29b6a5f6", "trainer.pt": "f77cbd7043301ba03a27d0fc3fb559b972721c8fb60167219ee0788286252c59", "training_config.json": "2122ad6f8105987f9b87dd55d13b632084853908566b081c829ffe801259e2bf", "performance.json": "f89f82c3f4c44a52fcd532346ea71580902d757e59e1961e9b2bed0030259dba", "shard_scheduler.json": "5316864976580fabdbfc5b4749a2a377f1a44f787f1e27a21ee310ec16436feb", "coverage.json": "e35c09a84496ab9947c7e21ef4a25e40c3c17be24bb510256b35b90dd258081c", "README.md": "9ce0f9efacbc46177a7e818b6aed274d84e7145ff944b5c7d6ecd5d33b9d4419"}, "signature": "9fee47918905b0750bee87876aa424f021ccc4507ee8db22a11d01fbc6109063", "folder": "step-000000285000"}
model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:ab3d534f71afc536f11eacb5cac49b41552e56c521fecc7b301c5b0f3ba211e7
3
  size 360622744
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8ac057f2166a1e4db0ce4a9415d89891b9b8ddcc52297129f86c08f788b2758b
3
  size 360622744
performance.json CHANGED
@@ -49,11 +49,11 @@
49
  "selected": {
50
  "microbatch": 86,
51
  "mode": "compiled",
52
- "microbatch_seconds": 0.8973423759999832,
53
- "peak_gb": 85.39913082122803,
54
  "fits_budget": true,
55
- "estimated_update_seconds": 0.9528963840000415,
56
- "input_tokens_per_second": 196277.36827175462
57
  },
58
  "measurements": [
59
  {
 
49
  "selected": {
50
  "microbatch": 86,
51
  "mode": "compiled",
52
+ "microbatch_seconds": 0.8979115860001912,
53
+ "peak_gb": 85.39889430999756,
54
  "fits_budget": true,
55
+ "estimated_update_seconds": 0.953845084999557,
56
+ "input_tokens_per_second": 196152.9428354681
57
  },
58
  "measurements": [
59
  {
shard_scheduler.json CHANGED
@@ -22,8 +22,8 @@
22
  "tokens_per_second": 1062967.6168812835,
23
  "note": "Bounded fresh sample; full-page latency is an extrapolation.",
24
  "estimated_pages_for_buffer": 1,
25
- "required_token_positions_per_second": 75609.08177184862,
26
- "required_pages_per_second_upper_estimate": 0.006553603578518049
27
  },
28
  {
29
  "source": "rewrite6",
@@ -36,8 +36,8 @@
36
  "tokens_per_second": 1206672.263377472,
37
  "note": "Bounded fresh sample; full-page latency is an extrapolation.",
38
  "estimated_pages_for_buffer": 1,
39
- "required_token_positions_per_second": 9451.135221481078,
40
- "required_pages_per_second_upper_estimate": 0.0037701991469128284
41
  },
42
  {
43
  "source": "openmath",
@@ -50,8 +50,8 @@
50
  "tokens_per_second": 410042.7581134458,
51
  "note": "Bounded fresh sample; full-page latency is an extrapolation.",
52
  "estimated_pages_for_buffer": 1,
53
- "required_token_positions_per_second": 18902.270442962155,
54
- "required_pages_per_second_upper_estimate": 0.004372707598381542
55
  },
56
  {
57
  "source": "ultra_qa",
@@ -64,8 +64,8 @@
64
  "tokens_per_second": 2163486.5624068812,
65
  "note": "Bounded fresh sample; full-page latency is an extrapolation.",
66
  "estimated_pages_for_buffer": 1,
67
- "required_token_positions_per_second": 28353.405664443235,
68
- "required_pages_per_second_upper_estimate": 0.0017265736461417594
69
  },
70
  {
71
  "source": "ultra_style",
@@ -78,8 +78,8 @@
78
  "tokens_per_second": 1546694.2737633225,
79
  "note": "Bounded fresh sample; full-page latency is an extrapolation.",
80
  "estimated_pages_for_buffer": 1,
81
- "required_token_positions_per_second": 28353.405664443235,
82
- "required_pages_per_second_upper_estimate": 0.0021827236552991267
83
  },
84
  {
85
  "source": "cortex",
@@ -91,9 +91,9 @@
91
  "read_seconds": 0.042780299000014566,
92
  "tokens_per_second": 1015177.8395898147,
93
  "note": "Bounded fresh sample; full-page latency is an extrapolation.",
94
- "estimated_pages_for_buffer": 22,
95
- "required_token_positions_per_second": 28353.405664443235,
96
- "required_pages_per_second_upper_estimate": 0.3520544056078974
97
  }
98
  ],
99
  "gpu_measurement": "Selected forward/backward estimate; replaced by live successful update timings.",
@@ -155,59 +155,263 @@
155
  "shard_benchmark_scope": "Local packing/read measurements; upload capacity measured only on real data commits. Targets larger than available pages are sample-limited.",
156
  "estimated_fresh_generation_batch_seconds": 0.1434574083060457,
157
  "selected": {
158
- "prefetch_batches": 62,
159
  "prefetch_capacity_limit": 512,
160
  "batch_memory_estimate_bytes": 690670,
161
- "producer_batch_seconds": 0.07287284228125657,
162
- "producer_p95_seconds": 0.1761082670000178,
163
- "gpu_update_seconds": 0.9751209546820736,
164
- "producer_batches_per_second": 13.722533233168638,
165
- "gpu_batches_per_second": 1.0255138044141794,
166
- "producer_headroom": 13.381129706983886,
167
  "sustainable": true,
168
- "buffer_seconds": 60.457499190288566,
169
  "buffer_drain_seconds": null,
170
  "shard_target_mib": 16,
171
- "upload_pages_per_commit": 79,
172
  "max_pending_pages": 4096,
173
- "generation_bytes_per_second": 382834.16564839584,
174
- "upload_bytes_per_second": 4501863.534696111,
175
  "upload_sustainable": true,
176
  "logical_page_rows": 4096,
177
  "logical_cortex_page_rows": 128,
178
- "estimated_full_shards_per_hour": 82.14730002488047,
179
  "selection_policy": "Smallest shard within 90% of measured local peak, enlarged to amortize 30 seconds of measured page production; bounded to 256 MiB."
180
  },
181
  "upload_measurements": [
182
  {
183
- "seconds": 0.8291290090000984,
184
- "bytes": 27860,
185
  "pages": 1,
186
  "shards": 1
187
  },
188
  {
189
- "seconds": 0.7566531879992908,
190
- "bytes": 935968,
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
191
  "pages": 27,
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
192
  "shards": 1
193
  },
194
  {
195
- "seconds": 1.5188891550005792,
196
- "bytes": 5974366,
197
- "pages": 29,
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
198
  "shards": 2
199
  },
200
  {
201
- "seconds": 1.1771950970005491,
202
- "bytes": 7732243,
203
- "pages": 17,
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
204
  "shards": 3
205
  },
206
  {
207
- "seconds": 1.3709215139997468,
208
- "bytes": 10777643,
209
- "pages": 39,
210
  "shards": 3
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
211
  }
212
  ],
213
  "cache_key": {
@@ -252,7 +456,7 @@
252
  },
253
  "cache_reused": true,
254
  "resume_probe_seconds": [
255
- 3.215140269000017,
256
- 0.014588057999958437
257
  ]
258
  }
 
22
  "tokens_per_second": 1062967.6168812835,
23
  "note": "Bounded fresh sample; full-page latency is an extrapolation.",
24
  "estimated_pages_for_buffer": 1,
25
+ "required_token_positions_per_second": 74771.75244308176,
26
+ "required_pages_per_second_upper_estimate": 0.00648102599449232
27
  },
28
  {
29
  "source": "rewrite6",
 
36
  "tokens_per_second": 1206672.263377472,
37
  "note": "Bounded fresh sample; full-page latency is an extrapolation.",
38
  "estimated_pages_for_buffer": 1,
39
+ "required_token_positions_per_second": 9346.46905538522,
40
+ "required_pages_per_second_upper_estimate": 0.0037284462483585528
41
  },
42
  {
43
  "source": "openmath",
 
50
  "tokens_per_second": 410042.7581134458,
51
  "note": "Bounded fresh sample; full-page latency is an extrapolation.",
52
  "estimated_pages_for_buffer": 1,
53
+ "required_token_positions_per_second": 18692.93811077044,
54
+ "required_pages_per_second_upper_estimate": 0.004324282247452207
55
  },
56
  {
57
  "source": "ultra_qa",
 
64
  "tokens_per_second": 2163486.5624068812,
65
  "note": "Bounded fresh sample; full-page latency is an extrapolation.",
66
  "estimated_pages_for_buffer": 1,
67
+ "required_token_positions_per_second": 28039.40716615566,
68
+ "required_pages_per_second_upper_estimate": 0.001707452785018847
69
  },
70
  {
71
  "source": "ultra_style",
 
78
  "tokens_per_second": 1546694.2737633225,
79
  "note": "Bounded fresh sample; full-page latency is an extrapolation.",
80
  "estimated_pages_for_buffer": 1,
81
+ "required_token_positions_per_second": 28039.40716615566,
82
+ "required_pages_per_second_upper_estimate": 0.0021585511816974746
83
  },
84
  {
85
  "source": "cortex",
 
91
  "read_seconds": 0.042780299000014566,
92
  "tokens_per_second": 1015177.8395898147,
93
  "note": "Bounded fresh sample; full-page latency is an extrapolation.",
94
+ "estimated_pages_for_buffer": 21,
95
+ "required_token_positions_per_second": 28039.40716615566,
96
+ "required_pages_per_second_upper_estimate": 0.34815559514453803
97
  }
98
  ],
99
  "gpu_measurement": "Selected forward/backward estimate; replaced by live successful update timings.",
 
155
  "shard_benchmark_scope": "Local packing/read measurements; upload capacity measured only on real data commits. Targets larger than available pages are sample-limited.",
156
  "estimated_fresh_generation_batch_seconds": 0.1434574083060457,
157
  "selected": {
158
+ "prefetch_batches": 61,
159
  "prefetch_capacity_limit": 512,
160
  "batch_memory_estimate_bytes": 690670,
161
+ "producer_batch_seconds": 0.0779331768125644,
162
+ "producer_p95_seconds": 0.2094176940008765,
163
+ "gpu_update_seconds": 0.9860408187720853,
164
+ "producer_batches_per_second": 12.831505668055609,
165
+ "gpu_batches_per_second": 1.014156798544403,
166
+ "producer_headroom": 12.652388355008206,
167
  "sustainable": true,
168
+ "buffer_seconds": 60.148489945097204,
169
  "buffer_drain_seconds": null,
170
  "shard_target_mib": 16,
171
+ "upload_pages_per_commit": 83,
172
  "max_pending_pages": 4096,
173
+ "generation_bytes_per_second": 313527.5392295587,
174
+ "upload_bytes_per_second": 5440988.382144982,
175
  "upload_sustainable": true,
176
  "logical_page_rows": 4096,
177
  "logical_cortex_page_rows": 128,
178
+ "estimated_full_shards_per_hour": 67.2757113710887,
179
  "selection_policy": "Smallest shard within 90% of measured local peak, enlarged to amortize 30 seconds of measured page production; bounded to 256 MiB."
180
  },
181
  "upload_measurements": [
182
  {
183
+ "seconds": 0.8854921419988386,
184
+ "bytes": 43628,
185
  "pages": 1,
186
  "shards": 1
187
  },
188
  {
189
+ "seconds": 1.4819931290003296,
190
+ "bytes": 7791997,
191
+ "pages": 45,
192
+ "shards": 3
193
+ },
194
+ {
195
+ "seconds": 0.999563274999673,
196
+ "bytes": 6224997,
197
+ "pages": 32,
198
+ "shards": 2
199
+ },
200
+ {
201
+ "seconds": 1.0198160460004146,
202
+ "bytes": 6464312,
203
+ "pages": 37,
204
+ "shards": 3
205
+ },
206
+ {
207
+ "seconds": 1.407603981999273,
208
+ "bytes": 11506972,
209
+ "pages": 28,
210
+ "shards": 3
211
+ },
212
+ {
213
+ "seconds": 1.4075674270006857,
214
+ "bytes": 6094729,
215
+ "pages": 35,
216
+ "shards": 2
217
+ },
218
+ {
219
+ "seconds": 1.2288260940003966,
220
+ "bytes": 2371624,
221
+ "pages": 28,
222
+ "shards": 2
223
+ },
224
+ {
225
+ "seconds": 1.1107631630002288,
226
+ "bytes": 10173347,
227
+ "pages": 40,
228
+ "shards": 3
229
+ },
230
+ {
231
+ "seconds": 0.9085161899984087,
232
+ "bytes": 2572379,
233
+ "pages": 35,
234
+ "shards": 2
235
+ },
236
+ {
237
+ "seconds": 1.1832448179993662,
238
+ "bytes": 6068567,
239
+ "pages": 35,
240
+ "shards": 2
241
+ },
242
+ {
243
+ "seconds": 1.2676941070003522,
244
+ "bytes": 16626023,
245
+ "pages": 38,
246
+ "shards": 4
247
+ },
248
+ {
249
+ "seconds": 0.9997899369991501,
250
+ "bytes": 2248595,
251
  "pages": 27,
252
+ "shards": 2
253
+ },
254
+ {
255
+ "seconds": 0.9953625339985592,
256
+ "bytes": 6050082,
257
+ "pages": 36,
258
+ "shards": 2
259
+ },
260
+ {
261
+ "seconds": 1.2188318490007077,
262
+ "bytes": 7592648,
263
+ "pages": 38,
264
+ "shards": 3
265
+ },
266
+ {
267
+ "seconds": 0.8659798669996235,
268
+ "bytes": 5051955,
269
+ "pages": 32,
270
+ "shards": 2
271
+ },
272
+ {
273
+ "seconds": 1.0979891799997858,
274
+ "bytes": 6035254,
275
+ "pages": 36,
276
+ "shards": 2
277
+ },
278
+ {
279
+ "seconds": 1.2733040839993919,
280
+ "bytes": 13322526,
281
+ "pages": 35,
282
+ "shards": 4
283
+ },
284
+ {
285
+ "seconds": 0.8405508110008668,
286
+ "bytes": 1090490,
287
+ "pages": 33,
288
  "shards": 1
289
  },
290
  {
291
+ "seconds": 1.1666756240010727,
292
+ "bytes": 10146533,
293
+ "pages": 27,
294
+ "shards": 3
295
+ },
296
+ {
297
+ "seconds": 1.1112560749988916,
298
+ "bytes": 2663601,
299
+ "pages": 36,
300
+ "shards": 3
301
+ },
302
+ {
303
+ "seconds": 1.182102421998934,
304
+ "bytes": 5962190,
305
+ "pages": 36,
306
+ "shards": 2
307
+ },
308
+ {
309
+ "seconds": 0.9607173089989374,
310
+ "bytes": 7750938,
311
+ "pages": 37,
312
+ "shards": 3
313
+ },
314
+ {
315
+ "seconds": 1.1050905230003991,
316
+ "bytes": 6187776,
317
+ "pages": 15,
318
+ "shards": 2
319
+ },
320
+ {
321
+ "seconds": 1.2981188600006135,
322
+ "bytes": 10611352,
323
+ "pages": 45,
324
+ "shards": 3
325
+ },
326
+ {
327
+ "seconds": 1.1970796210007393,
328
+ "bytes": 2312155,
329
+ "pages": 27,
330
  "shards": 2
331
  },
332
  {
333
+ "seconds": 1.6818293220003397,
334
+ "bytes": 5841398,
335
+ "pages": 35,
336
+ "shards": 2
337
+ },
338
+ {
339
+ "seconds": 1.1767334710002615,
340
+ "bytes": 1036234,
341
+ "pages": 30,
342
+ "shards": 1
343
+ },
344
+ {
345
+ "seconds": 1.7388112589997036,
346
+ "bytes": 12464285,
347
+ "pages": 43,
348
+ "shards": 4
349
+ },
350
+ {
351
+ "seconds": 0.8487006530012877,
352
+ "bytes": 592565,
353
+ "pages": 19,
354
+ "shards": 1
355
+ },
356
+ {
357
+ "seconds": 1.3723258429999987,
358
+ "bytes": 11921886,
359
+ "pages": 40,
360
  "shards": 3
361
  },
362
  {
363
+ "seconds": 1.34920726099881,
364
+ "bytes": 7816018,
365
+ "pages": 35,
366
  "shards": 3
367
+ },
368
+ {
369
+ "seconds": 1.4124847809998755,
370
+ "bytes": 4874448,
371
+ "pages": 37,
372
+ "shards": 2
373
+ },
374
+ {
375
+ "seconds": 1.2419382850002876,
376
+ "bytes": 7186580,
377
+ "pages": 29,
378
+ "shards": 3
379
+ },
380
+ {
381
+ "seconds": 1.1265525479993812,
382
+ "bytes": 6298017,
383
+ "pages": 35,
384
+ "shards": 2
385
+ },
386
+ {
387
+ "seconds": 0.9624441889991431,
388
+ "bytes": 1027701,
389
+ "pages": 30,
390
+ "shards": 1
391
+ },
392
+ {
393
+ "seconds": 1.7364580140001635,
394
+ "bytes": 17104134,
395
+ "pages": 27,
396
+ "shards": 5
397
+ },
398
+ {
399
+ "seconds": 0.9077872060006484,
400
+ "bytes": 916446,
401
+ "pages": 29,
402
+ "shards": 1
403
+ },
404
+ {
405
+ "seconds": 1.0004914799992548,
406
+ "bytes": 6063816,
407
+ "pages": 36,
408
+ "shards": 2
409
+ },
410
+ {
411
+ "seconds": 0.9483974380000291,
412
+ "bytes": 2643403,
413
+ "pages": 36,
414
+ "shards": 2
415
  }
416
  ],
417
  "cache_key": {
 
456
  },
457
  "cache_reused": true,
458
  "resume_probe_seconds": [
459
+ 1.1575575899996693,
460
+ 0.014994120000665134
461
  ]
462
  }
trainer.pt CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:93080e3d08484f8e03c44bc9ba9acf6621eb9dc2cb2663dd5f9becbb62b890ee
3
- size 721763455
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f77cbd7043301ba03a27d0fc3fb559b972721c8fb60167219ee0788286252c59
3
+ size 721776383
training_config.json CHANGED
@@ -46,8 +46,8 @@
46
  "cpu_threads": 4,
47
  "page_rows": 4096,
48
  "cortex_page_rows": 128,
49
- "prefetch_batches": 62,
50
- "prefetch_low_water": 31,
51
  "memory_pages": 12,
52
  "cache_gb": 20,
53
  "force_rebenchmark": false,
 
46
  "cpu_threads": 4,
47
  "page_rows": 4096,
48
  "cortex_page_rows": 128,
49
+ "prefetch_batches": 61,
50
+ "prefetch_low_water": 30,
51
  "memory_pages": 12,
52
  "cache_gb": 20,
53
  "force_rebenchmark": false,