appvoid commited on
Commit
1b49619
·
verified ·
1 Parent(s): ebc3a45

Qwen90M step 125000; complete resumable checkpoint

Browse files
Files changed (7) hide show
  1. complete.json +6 -6
  2. coverage.json +49 -55
  3. latest.json +1 -1
  4. model.safetensors +1 -1
  5. shard_scheduler.json +270 -24
  6. trainer.pt +2 -2
  7. training_config.json +2 -2
complete.json CHANGED
@@ -1,18 +1,18 @@
1
  {
2
- "step": 120000,
3
  "files": {
4
  "config.json": "cbc72137da3dae65085a2a8c39ad7634d16bf92d785ca00d283ee82fe32b37e7",
5
  "generation_config.json": "54ed0d19f6226888c91157964ede8a892bd5ae175bfad85336152b3d9b6cd7c4",
6
- "model.safetensors": "665c262180b366c0f7d7c0ccdbe415e3e0c4393bd3ac12b2f5d7868b4231e070",
7
  "tokenization_byte.py": "279b9990f77b1eaedc7c1184fe13d60273a713ead856200f1fb9ae323c7b1e03",
8
  "tokenizer_config.json": "0ecbd4fd39e35125800d806af981ba0cacd8a7ba5f02c9e729c4ecc54af6d300",
9
  "special_tokens_map.json": "20d4b94d0e657c3f44d9123522c106e7315b6c1a1b6d259478f17bd2918b1263",
10
  "byte_vocab.json": "57a1cd2e70dd5b7cc6fd2eaa12ebfcd173c1620fa5d45c2c6282511c29b6a5f6",
11
- "trainer.pt": "bdb878e6fce84b23404d1ef62d5e22e387ecfe6564176b5d4c1419f25bad1f91",
12
- "training_config.json": "2122ad6f8105987f9b87dd55d13b632084853908566b081c829ffe801259e2bf",
13
  "performance.json": "10fa55d76d1a4ec4f9132ef494d950eba8aa9e6cecd5fa1ceb70ba6de35537e4",
14
- "shard_scheduler.json": "bae87ec96511284350c98b6a73334b71b1f36afb5494e3adbee78945a9424227",
15
- "coverage.json": "63e769b4b6d328325a99ba526c0df24cca09fa2e738841235713342ebf77ee84",
16
  "README.md": "9ce0f9efacbc46177a7e818b6aed274d84e7145ff944b5c7d6ecd5d33b9d4419"
17
  },
18
  "signature": "9fee47918905b0750bee87876aa424f021ccc4507ee8db22a11d01fbc6109063"
 
1
  {
2
+ "step": 125000,
3
  "files": {
4
  "config.json": "cbc72137da3dae65085a2a8c39ad7634d16bf92d785ca00d283ee82fe32b37e7",
5
  "generation_config.json": "54ed0d19f6226888c91157964ede8a892bd5ae175bfad85336152b3d9b6cd7c4",
6
+ "model.safetensors": "f0616fc9490755c1e55484204bd32a0a2d067fdbf3f53d5cd1eef115bc2c5201",
7
  "tokenization_byte.py": "279b9990f77b1eaedc7c1184fe13d60273a713ead856200f1fb9ae323c7b1e03",
8
  "tokenizer_config.json": "0ecbd4fd39e35125800d806af981ba0cacd8a7ba5f02c9e729c4ecc54af6d300",
9
  "special_tokens_map.json": "20d4b94d0e657c3f44d9123522c106e7315b6c1a1b6d259478f17bd2918b1263",
10
  "byte_vocab.json": "57a1cd2e70dd5b7cc6fd2eaa12ebfcd173c1620fa5d45c2c6282511c29b6a5f6",
11
+ "trainer.pt": "d623373e105ee371e8cd3e233d58364ce3b44c87e5d4ec425bc868f1ad80c99f",
12
+ "training_config.json": "95ba9e002790d11959d1191b36ed994bfa10a6da51c22c772e143bec517a9783",
13
  "performance.json": "10fa55d76d1a4ec4f9132ef494d950eba8aa9e6cecd5fa1ceb70ba6de35537e4",
14
+ "shard_scheduler.json": "35169e3854bb2f8709178543fd2af9faeefbd503738dc898df29818c62aff861",
15
+ "coverage.json": "81eba1ac2ae272a1f49252c853da80c2647c2cfb73db9e4286115d001cacf6ff",
16
  "README.md": "9ce0f9efacbc46177a7e818b6aed274d84e7145ff944b5c7d6ecd5d33b9d4419"
17
  },
18
  "signature": "9fee47918905b0750bee87876aa424f021ccc4507ee8db22a11d01fbc6109063"
coverage.json CHANGED
@@ -1,107 +1,101 @@
1
  {
2
- "step": 120000,
3
  "stats": {
4
- "climbmix/blocks": 4320000,
5
- "climbmix/targets": 7218731219,
6
- "climbmix/input_tokens": 7219189067,
7
- "ultra_qa/blocks": 1620000,
8
- "ultra_qa/targets": 2642548151,
9
- "ultra_qa/input_tokens": 2642553234,
10
- "cortex/blocks": 1620000,
11
- "cortex/targets": 61617547,
12
- "cortex/input_tokens": 2575003385,
13
- "ultra_style/blocks": 1620000,
14
- "ultra_style/targets": 2561942397,
15
- "ultra_style/input_tokens": 2562077219,
16
- "openmath/blocks": 1080000,
17
- "openmath/targets": 1668565915,
18
- "openmath/input_tokens": 1669122279,
19
- "rewrite6/blocks": 540000,
20
- "rewrite6/targets": 885200928,
21
- "rewrite6/input_tokens": 886171529
22
  },
23
  "sources": {
24
  "climbmix": {
25
- "page": 592,
26
- "row": 792,
27
  "start": {
28
- "file": 28,
29
  "reader": {
30
  "format": "parquet-row-v1",
31
- "row_group": 48,
32
  "row": 1024
33
  }
34
  },
35
  "epoch": 0,
36
- "offset": 6144,
37
- "documents": 2425624
38
  },
39
  "rewrite6": {
40
- "page": 36,
41
- "row": 2922,
42
- "start": {
43
- "file": 0,
44
- "reader": {
45
- "format": "jsonl-row-v1",
46
- "row": 147456
47
- }
48
- },
49
- "epoch": 6,
50
- "offset": 4096,
51
- "documents": 1382802
52
  },
53
  "openmath": {
54
- "page": 369,
55
- "row": 3575,
56
  "start": {
57
  "file": 3,
58
  "reader": {
59
  "format": "parquet-row-v1",
60
- "row_group": 201,
61
- "row": 474
62
  }
63
  },
64
  "epoch": 0,
65
  "offset": 0,
66
- "documents": 1514999
67
  },
68
  "ultra_qa": {
69
- "page": 159,
70
- "row": 3808,
71
  "start": {
72
  "file": 1,
73
  "reader": {
74
  "format": "parquet-row-v1",
75
  "row_group": 2,
76
- "row": 2444
77
  }
78
  },
79
  "epoch": 0,
80
  "offset": 2048,
81
- "documents": 655072
82
  },
83
  "ultra_style": {
84
- "page": 244,
85
- "row": 1607,
86
  "start": {
87
  "file": 1,
88
  "reader": {
89
  "format": "parquet-row-v1",
90
- "row_group": 3,
91
- "row": 58215
92
  }
93
  },
94
  "epoch": 0,
95
- "offset": 2048,
96
- "documents": 1001031
97
  },
98
  "cortex": {
99
- "page": 31415,
100
- "row": 67,
101
  "start": null,
102
  "epoch": 0,
103
  "offset": 0,
104
- "documents": 4021187
105
  }
106
  }
107
  }
 
1
  {
2
+ "step": 125000,
3
  "stats": {
4
+ "climbmix/blocks": 4500000,
5
+ "climbmix/targets": 7519797808,
6
+ "climbmix/input_tokens": 7520274931,
7
+ "ultra_qa/blocks": 1687500,
8
+ "ultra_qa/targets": 2752448828,
9
+ "ultra_qa/input_tokens": 2752454113,
10
+ "cortex/blocks": 1687500,
11
+ "cortex/targets": 64164981,
12
+ "cortex/input_tokens": 2682291180,
13
+ "ultra_style/blocks": 1687500,
14
+ "ultra_style/targets": 2668606742,
15
+ "ultra_style/input_tokens": 2668746996,
16
+ "openmath/blocks": 1125000,
17
+ "openmath/targets": 1738258233,
18
+ "openmath/input_tokens": 1738837895,
19
+ "rewrite6/blocks": 562500,
20
+ "rewrite6/targets": 922143403,
21
+ "rewrite6/input_tokens": 923154365
22
  },
23
  "sources": {
24
  "climbmix": {
25
+ "page": 616,
26
+ "row": 3708,
27
  "start": {
28
+ "file": 29,
29
  "reader": {
30
  "format": "parquet-row-v1",
31
+ "row_group": 61,
32
  "row": 1024
33
  }
34
  },
35
  "epoch": 0,
36
+ "offset": 0,
37
+ "documents": 2526844
38
  },
39
  "rewrite6": {
40
+ "page": 0,
41
+ "row": 2445,
42
+ "start": null,
43
+ "epoch": 7,
44
+ "offset": 2048,
45
+ "documents": 1440273
 
 
 
 
 
 
46
  },
47
  "openmath": {
48
+ "page": 385,
49
+ "row": 1190,
50
  "start": {
51
  "file": 3,
52
  "reader": {
53
  "format": "parquet-row-v1",
54
+ "row_group": 267,
55
+ "row": 10
56
  }
57
  },
58
  "epoch": 0,
59
  "offset": 0,
60
+ "documents": 1578150
61
  },
62
  "ultra_qa": {
63
+ "page": 166,
64
+ "row": 2396,
65
  "start": {
66
  "file": 1,
67
  "reader": {
68
  "format": "parquet-row-v1",
69
  "row_group": 2,
70
+ "row": 31116
71
  }
72
  },
73
  "epoch": 0,
74
  "offset": 2048,
75
+ "documents": 682332
76
  },
77
  "ultra_style": {
78
+ "page": 254,
79
+ "row": 1950,
80
  "start": {
81
  "file": 1,
82
  "reader": {
83
  "format": "parquet-row-v1",
84
+ "row_group": 4,
85
+ "row": 14310
86
  }
87
  },
88
  "epoch": 0,
89
+ "offset": 0,
90
+ "documents": 1042334
91
  },
92
  "cortex": {
93
+ "page": 32727,
94
+ "row": 12,
95
  "start": null,
96
  "epoch": 0,
97
  "offset": 0,
98
+ "documents": 4189068
99
  }
100
  }
101
  }
latest.json CHANGED
@@ -1 +1 @@
1
- {"step": 120000, "files": {"config.json": "cbc72137da3dae65085a2a8c39ad7634d16bf92d785ca00d283ee82fe32b37e7", "generation_config.json": "54ed0d19f6226888c91157964ede8a892bd5ae175bfad85336152b3d9b6cd7c4", "model.safetensors": "665c262180b366c0f7d7c0ccdbe415e3e0c4393bd3ac12b2f5d7868b4231e070", "tokenization_byte.py": "279b9990f77b1eaedc7c1184fe13d60273a713ead856200f1fb9ae323c7b1e03", "tokenizer_config.json": "0ecbd4fd39e35125800d806af981ba0cacd8a7ba5f02c9e729c4ecc54af6d300", "special_tokens_map.json": "20d4b94d0e657c3f44d9123522c106e7315b6c1a1b6d259478f17bd2918b1263", "byte_vocab.json": "57a1cd2e70dd5b7cc6fd2eaa12ebfcd173c1620fa5d45c2c6282511c29b6a5f6", "trainer.pt": "bdb878e6fce84b23404d1ef62d5e22e387ecfe6564176b5d4c1419f25bad1f91", "training_config.json": "2122ad6f8105987f9b87dd55d13b632084853908566b081c829ffe801259e2bf", "performance.json": "10fa55d76d1a4ec4f9132ef494d950eba8aa9e6cecd5fa1ceb70ba6de35537e4", "shard_scheduler.json": "bae87ec96511284350c98b6a73334b71b1f36afb5494e3adbee78945a9424227", "coverage.json": "63e769b4b6d328325a99ba526c0df24cca09fa2e738841235713342ebf77ee84", "README.md": "9ce0f9efacbc46177a7e818b6aed274d84e7145ff944b5c7d6ecd5d33b9d4419"}, "signature": "9fee47918905b0750bee87876aa424f021ccc4507ee8db22a11d01fbc6109063", "folder": "step-000000120000"}
 
1
+ {"step": 125000, "files": {"config.json": "cbc72137da3dae65085a2a8c39ad7634d16bf92d785ca00d283ee82fe32b37e7", "generation_config.json": "54ed0d19f6226888c91157964ede8a892bd5ae175bfad85336152b3d9b6cd7c4", "model.safetensors": "f0616fc9490755c1e55484204bd32a0a2d067fdbf3f53d5cd1eef115bc2c5201", "tokenization_byte.py": "279b9990f77b1eaedc7c1184fe13d60273a713ead856200f1fb9ae323c7b1e03", "tokenizer_config.json": "0ecbd4fd39e35125800d806af981ba0cacd8a7ba5f02c9e729c4ecc54af6d300", "special_tokens_map.json": "20d4b94d0e657c3f44d9123522c106e7315b6c1a1b6d259478f17bd2918b1263", "byte_vocab.json": "57a1cd2e70dd5b7cc6fd2eaa12ebfcd173c1620fa5d45c2c6282511c29b6a5f6", "trainer.pt": "d623373e105ee371e8cd3e233d58364ce3b44c87e5d4ec425bc868f1ad80c99f", "training_config.json": "95ba9e002790d11959d1191b36ed994bfa10a6da51c22c772e143bec517a9783", "performance.json": "10fa55d76d1a4ec4f9132ef494d950eba8aa9e6cecd5fa1ceb70ba6de35537e4", "shard_scheduler.json": "35169e3854bb2f8709178543fd2af9faeefbd503738dc898df29818c62aff861", "coverage.json": "81eba1ac2ae272a1f49252c853da80c2647c2cfb73db9e4286115d001cacf6ff", "README.md": "9ce0f9efacbc46177a7e818b6aed274d84e7145ff944b5c7d6ecd5d33b9d4419"}, "signature": "9fee47918905b0750bee87876aa424f021ccc4507ee8db22a11d01fbc6109063", "folder": "step-000000125000"}
model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:665c262180b366c0f7d7c0ccdbe415e3e0c4393bd3ac12b2f5d7868b4231e070
3
  size 360622744
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f0616fc9490755c1e55484204bd32a0a2d067fdbf3f53d5cd1eef115bc2c5201
3
  size 360622744
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": 74847.18384015403,
26
- "required_pages_per_second_upper_estimate": 0.006487564196811416
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": 9355.897980019254,
40
- "required_pages_per_second_upper_estimate": 0.003732207587370055
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": 18711.795960038507,
54
- "required_pages_per_second_upper_estimate": 0.004328644678993562
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": 28067.693940057758,
68
- "required_pages_per_second_upper_estimate": 0.0017091753011402517
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": 28067.693940057758,
82
- "required_pages_per_second_upper_estimate": 0.00216072877585526
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": 21,
95
- "required_token_positions_per_second": 28067.693940057758,
96
- "required_pages_per_second_upper_estimate": 0.34850682220665974
97
  }
98
  ],
99
  "gpu_measurement": "Selected forward/backward estimate; replaced by live successful update timings.",
@@ -155,27 +155,27 @@
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": 688946,
161
- "producer_batch_seconds": 0.05070418975005708,
162
- "producer_p95_seconds": 0.17277111900057207,
163
- "gpu_update_seconds": 0.9850470814968244,
164
- "producer_batches_per_second": 19.72223607022286,
165
- "gpu_batches_per_second": 1.015179902345839,
166
- "producer_headroom": 19.42733108156443,
167
  "sustainable": true,
168
- "buffer_seconds": 60.08787197130629,
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": 151662.29245108145,
174
- "upload_bytes_per_second": 3580018.611091431,
175
  "upload_sustainable": true,
176
  "logical_page_rows": 4096,
177
  "logical_cortex_page_rows": 128,
178
- "estimated_full_shards_per_hour": 32.54319744252522,
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": [
@@ -184,6 +184,252 @@
184
  "bytes": 4555121,
185
  "pages": 1,
186
  "shards": 1
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
187
  }
188
  ],
189
  "cache_key": {
 
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": 76000.48524185155,
26
+ "required_pages_per_second_upper_estimate": 0.006587529439294878
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": 9500.060655231444,
40
+ "required_pages_per_second_upper_estimate": 0.003789716233936271
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": 19000.121310462888,
54
+ "required_pages_per_second_upper_estimate": 0.0043953436744613855
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": 28500.181965694333,
68
+ "required_pages_per_second_upper_estimate": 0.001735511552812213
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": 28500.181965694333,
82
+ "required_pages_per_second_upper_estimate": 0.002194022901272243
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": 22,
95
+ "required_token_positions_per_second": 28500.181965694333,
96
+ "required_pages_per_second_upper_estimate": 0.3538768760407556
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": 62,
159
  "prefetch_capacity_limit": 512,
160
  "batch_memory_estimate_bytes": 688946,
161
+ "producer_batch_seconds": 0.08882844120317657,
162
+ "producer_p95_seconds": 0.2087490030007757,
163
+ "gpu_update_seconds": 0.9700990693069923,
164
+ "producer_batches_per_second": 11.257655616321221,
165
+ "gpu_batches_per_second": 1.030822553736051,
166
+ "producer_headroom": 10.92104123597185,
167
  "sustainable": true,
168
+ "buffer_seconds": 60.14614229703352,
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": 538524.4832146277,
174
+ "upload_bytes_per_second": 6058163.920097129,
175
  "upload_sustainable": true,
176
  "logical_page_rows": 4096,
177
  "logical_cortex_page_rows": 128,
178
+ "estimated_full_shards_per_hour": 115.5548178894913,
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": [
 
184
  "bytes": 4555121,
185
  "pages": 1,
186
  "shards": 1
187
+ },
188
+ {
189
+ "seconds": 0.8187937080001575,
190
+ "bytes": 1193417,
191
+ "pages": 39,
192
+ "shards": 1
193
+ },
194
+ {
195
+ "seconds": 1.121016316999885,
196
+ "bytes": 13167246,
197
+ "pages": 32,
198
+ "shards": 4
199
+ },
200
+ {
201
+ "seconds": 0.7197847669995099,
202
+ "bytes": 852888,
203
+ "pages": 25,
204
+ "shards": 1
205
+ },
206
+ {
207
+ "seconds": 0.9669863799999803,
208
+ "bytes": 6140852,
209
+ "pages": 35,
210
+ "shards": 2
211
+ },
212
+ {
213
+ "seconds": 1.2269071499995334,
214
+ "bytes": 11690462,
215
+ "pages": 40,
216
+ "shards": 4
217
+ },
218
+ {
219
+ "seconds": 0.9768262770003275,
220
+ "bytes": 1050575,
221
+ "pages": 32,
222
+ "shards": 1
223
+ },
224
+ {
225
+ "seconds": 0.9254432580000866,
226
+ "bytes": 6140767,
227
+ "pages": 27,
228
+ "shards": 2
229
+ },
230
+ {
231
+ "seconds": 1.0060759970001527,
232
+ "bytes": 2526019,
233
+ "pages": 35,
234
+ "shards": 2
235
+ },
236
+ {
237
+ "seconds": 1.1933680550000645,
238
+ "bytes": 16188168,
239
+ "pages": 34,
240
+ "shards": 4
241
+ },
242
+ {
243
+ "seconds": 1.1706134349997228,
244
+ "bytes": 7624358,
245
+ "pages": 37,
246
+ "shards": 3
247
+ },
248
+ {
249
+ "seconds": 0.8122863409998899,
250
+ "bytes": 1264426,
251
+ "pages": 34,
252
+ "shards": 1
253
+ },
254
+ {
255
+ "seconds": 1.096932616000231,
256
+ "bytes": 5773373,
257
+ "pages": 27,
258
+ "shards": 2
259
+ },
260
+ {
261
+ "seconds": 1.2743502019993684,
262
+ "bytes": 11373993,
263
+ "pages": 42,
264
+ "shards": 4
265
+ },
266
+ {
267
+ "seconds": 0.9631920529991476,
268
+ "bytes": 874408,
269
+ "pages": 26,
270
+ "shards": 1
271
+ },
272
+ {
273
+ "seconds": 1.358759662999546,
274
+ "bytes": 11963324,
275
+ "pages": 41,
276
+ "shards": 3
277
+ },
278
+ {
279
+ "seconds": 0.99124130600012,
280
+ "bytes": 2283760,
281
+ "pages": 27,
282
+ "shards": 2
283
+ },
284
+ {
285
+ "seconds": 1.149097012999846,
286
+ "bytes": 11248733,
287
+ "pages": 40,
288
+ "shards": 3
289
+ },
290
+ {
291
+ "seconds": 1.4425641930001802,
292
+ "bytes": 7147715,
293
+ "pages": 28,
294
+ "shards": 3
295
+ },
296
+ {
297
+ "seconds": 1.1025827090006715,
298
+ "bytes": 1108574,
299
+ "pages": 33,
300
+ "shards": 1
301
+ },
302
+ {
303
+ "seconds": 1.2109743800001525,
304
+ "bytes": 6387518,
305
+ "pages": 36,
306
+ "shards": 3
307
+ },
308
+ {
309
+ "seconds": 1.4837992330003544,
310
+ "bytes": 17296274,
311
+ "pages": 40,
312
+ "shards": 5
313
+ },
314
+ {
315
+ "seconds": 0.8423324309997042,
316
+ "bytes": 1212988,
317
+ "pages": 34,
318
+ "shards": 1
319
+ },
320
+ {
321
+ "seconds": 1.0933355659999506,
322
+ "bytes": 5690232,
323
+ "pages": 27,
324
+ "shards": 2
325
+ },
326
+ {
327
+ "seconds": 0.9681096090007486,
328
+ "bytes": 2564629,
329
+ "pages": 35,
330
+ "shards": 2
331
+ },
332
+ {
333
+ "seconds": 1.0954296749987407,
334
+ "bytes": 10402230,
335
+ "pages": 33,
336
+ "shards": 3
337
+ },
338
+ {
339
+ "seconds": 0.9827494309993199,
340
+ "bytes": 7704375,
341
+ "pages": 37,
342
+ "shards": 3
343
+ },
344
+ {
345
+ "seconds": 0.9620766659991205,
346
+ "bytes": 6846709,
347
+ "pages": 33,
348
+ "shards": 2
349
+ },
350
+ {
351
+ "seconds": 1.0866875020001316,
352
+ "bytes": 5928612,
353
+ "pages": 35,
354
+ "shards": 2
355
+ },
356
+ {
357
+ "seconds": 1.1460560750001605,
358
+ "bytes": 6695792,
359
+ "pages": 30,
360
+ "shards": 3
361
+ },
362
+ {
363
+ "seconds": 1.0918549559992243,
364
+ "bytes": 5922050,
365
+ "pages": 34,
366
+ "shards": 2
367
+ },
368
+ {
369
+ "seconds": 1.1802996549995441,
370
+ "bytes": 7557182,
371
+ "pages": 38,
372
+ "shards": 3
373
+ },
374
+ {
375
+ "seconds": 0.960146198000075,
376
+ "bytes": 1081037,
377
+ "pages": 33,
378
+ "shards": 1
379
+ },
380
+ {
381
+ "seconds": 1.2600166319989512,
382
+ "bytes": 16069321,
383
+ "pages": 30,
384
+ "shards": 4
385
+ },
386
+ {
387
+ "seconds": 0.9229335809995973,
388
+ "bytes": 2649370,
389
+ "pages": 36,
390
+ "shards": 2
391
+ },
392
+ {
393
+ "seconds": 1.2004358359990874,
394
+ "bytes": 6185442,
395
+ "pages": 36,
396
+ "shards": 2
397
+ },
398
+ {
399
+ "seconds": 1.0889191619990015,
400
+ "bytes": 6163490,
401
+ "pages": 35,
402
+ "shards": 2
403
+ },
404
+ {
405
+ "seconds": 1.1912384690003819,
406
+ "bytes": 6120903,
407
+ "pages": 31,
408
+ "shards": 3
409
+ },
410
+ {
411
+ "seconds": 1.0073837069994624,
412
+ "bytes": 6037950,
413
+ "pages": 35,
414
+ "shards": 2
415
+ },
416
+ {
417
+ "seconds": 1.402298138000333,
418
+ "bytes": 13311379,
419
+ "pages": 39,
420
+ "shards": 4
421
+ },
422
+ {
423
+ "seconds": 0.9249057269989862,
424
+ "bytes": 914039,
425
+ "pages": 28,
426
+ "shards": 1
427
+ },
428
+ {
429
+ "seconds": 1.1632707989992923,
430
+ "bytes": 10884064,
431
+ "pages": 37,
432
+ "shards": 3
433
  }
434
  ],
435
  "cache_key": {
trainer.pt CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:bdb878e6fce84b23404d1ef62d5e22e387ecfe6564176b5d4c1419f25bad1f91
3
- size 721766847
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d623373e105ee371e8cd3e233d58364ce3b44c87e5d4ec425bc868f1ad80c99f
3
+ size 721783615
training_config.json CHANGED
@@ -46,8 +46,8 @@
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,
 
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,