vovaRL commited on
Commit
b10a6a0
·
verified ·
1 Parent(s): 0db507b

step 4000 | vovaRL | 1048576000 tok

Browse files
Files changed (3) hide show
  1. model.safetensors +1 -1
  2. optimizer.pt +1 -1
  3. training_state.json +53 -48
model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:1a63f537b8c0d95c8af14b80281b4d5e3b0ab3ce0c478662a8e41c0a0327d1ec
3
  size 5167420536
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0348f54f73dcae8d807a4f7656992c791b273db21f7d5ecbb630b00272b9df4e
3
  size 5167420536
optimizer.pt CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:2535d85eb5fb8bd639cc57d5c678e1819ba53fb0ce96690c2a238cee1dd18370
3
  size 10333957008
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b777984b56a298199f2db17c023428d6b1c9ea3d5aceae012b8a87acff32cd4b
3
  size 10333957008
training_state.json CHANGED
@@ -1,65 +1,70 @@
1
  {
2
- "step": 2000,
3
- "global_tokens": 524288000,
4
  "contributor": "vovaRL",
5
  "cloud": "home",
6
  "base_model": "scratch:colibri-nano",
7
- "wall_clock_s": 24170.038534879684,
8
  "history_tail": [
9
  {
10
- "step": 1960,
11
- "ntp": 3.183401107788086,
12
- "zloss": 2.059804916381836,
13
- "lr": 0.000294,
14
- "grad_norm": 0.3203125,
15
- "ppl": 24.128678391480943,
16
- "balance_max_imbalance": 0.2388916015625,
17
- "balance_dead_frac": 0.046875,
18
- "tokens": 513802240
 
19
  },
20
  {
21
- "step": 1970,
22
- "ntp": 3.9721052646636963,
23
- "zloss": 2.2449395656585693,
24
- "lr": 0.00029549999999999997,
25
- "grad_norm": 0.3828125,
26
- "ppl": 53.09619480002791,
27
- "balance_max_imbalance": 0.2203369140625,
28
- "balance_dead_frac": 0.046875,
29
- "tokens": 516423680
 
30
  },
31
  {
32
- "step": 1980,
33
- "ntp": 3.722830057144165,
34
- "zloss": 2.109726905822754,
35
- "lr": 0.00029699999999999996,
36
- "grad_norm": 0.51953125,
37
- "ppl": 41.38134010583109,
38
- "balance_max_imbalance": 0.246337890625,
39
- "balance_dead_frac": 0.0625,
40
- "tokens": 519045120
 
41
  },
42
  {
43
- "step": 1990,
44
- "ntp": 3.3724567890167236,
45
- "zloss": 2.113959550857544,
46
- "lr": 0.0002985,
47
- "grad_norm": 1.0703125,
48
- "ppl": 29.150054691870785,
49
- "balance_max_imbalance": 0.23040771484375,
50
- "balance_dead_frac": 0.0625,
51
- "tokens": 521666560
 
52
  },
53
  {
54
- "step": 2000,
55
- "ntp": 3.7987537384033203,
56
- "zloss": 2.129600763320923,
57
- "lr": 0.0003,
58
- "grad_norm": 0.263671875,
59
- "ppl": 44.645509823548146,
60
- "balance_max_imbalance": 0.20025634765625,
61
- "balance_dead_frac": 0.03125,
62
- "tokens": 524288000
 
63
  }
64
  ]
65
  }
 
1
  {
2
+ "step": 4000,
3
+ "global_tokens": 1048576000,
4
  "contributor": "vovaRL",
5
  "cloud": "home",
6
  "base_model": "scratch:colibri-nano",
7
+ "wall_clock_s": 24723.55901813507,
8
  "history_tail": [
9
  {
10
+ "step": 3960,
11
+ "t": 1785039126.503,
12
+ "ntp": 3.189427137374878,
13
+ "zloss": 0.6651961207389832,
14
+ "lr": 0.00029217775612628503,
15
+ "grad_norm": 0.2216796875,
16
+ "ppl": 24.274517495414393,
17
+ "balance_max_imbalance": 0.08148193359375,
18
+ "balance_dead_frac": 0.0,
19
+ "tokens": 1038090240
20
  },
21
  {
22
+ "step": 3970,
23
+ "t": 1785039249.712,
24
+ "ntp": 2.8921048641204834,
25
+ "zloss": 0.6262736916542053,
26
+ "lr": 0.00029209852354962506,
27
+ "grad_norm": 0.18359375,
28
+ "ppl": 18.03122296057614,
29
+ "balance_max_imbalance": 0.0859375,
30
+ "balance_dead_frac": 0.0,
31
+ "tokens": 1040711680
32
  },
33
  {
34
+ "step": 3980,
35
+ "t": 1785039373.033,
36
+ "ntp": 2.9871861934661865,
37
+ "zloss": 0.5937920808792114,
38
+ "lr": 0.00029201890380882044,
39
+ "grad_norm": 0.224609375,
40
+ "ppl": 19.82980667697314,
41
+ "balance_max_imbalance": 0.05596923828125,
42
+ "balance_dead_frac": 0.0,
43
+ "tokens": 1043333120
44
  },
45
  {
46
+ "step": 3990,
47
+ "t": 1785039496.27,
48
+ "ntp": 2.8618695735931396,
49
+ "zloss": 0.5900548696517944,
50
+ "lr": 0.00029193889714640653,
51
+ "grad_norm": 0.212890625,
52
+ "ppl": 17.49420308194801,
53
+ "balance_max_imbalance": 0.1015625,
54
+ "balance_dead_frac": 0.0,
55
+ "tokens": 1045954560
56
  },
57
  {
58
+ "step": 4000,
59
+ "t": 1785039619.251,
60
+ "ntp": 2.899266481399536,
61
+ "zloss": 0.7581670880317688,
62
+ "lr": 0.00029185850380609757,
63
+ "grad_norm": 0.2138671875,
64
+ "ppl": 18.160819183867048,
65
+ "balance_max_imbalance": 0.14605712890625,
66
+ "balance_dead_frac": 0.0,
67
+ "tokens": 1048576000
68
  }
69
  ]
70
  }