Hugging Face's logo

AdaptiveChunking
/
pilot_A_s2

PyTorch
Safetensors
dynamic-chunking
tokenization
hnet
byte-level
beetle-hnet
legacy-trainer
Model card Files
xet
 Community
suchirsalhan commited on
Commit
49ca281
·
verified ·
1 Parent(s): 905a25c

run records (split from bundled legacy repo)

Browse files
Files changed (2) hide show
  1. final.json +114 -0
  2. log.jsonl +0 -0
final.json ADDED
@@ -0,0 +1,114 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "event": "final",
3
+ "run": "pilot_A_s2",
4
+ "condition": "A",
5
+ "step": 6000,
6
+ "params": 22516992,
7
+ "per_lang": {
8
+ "en": {
9
+ "cpc": 0.1836464977302764,
10
+ "cps": 24.96875,
11
+ "bpb": 2.0366536574759873,
12
+ "bytes_per_chunk": 5.642116137919159
13
+ },
14
+ "de": {
15
+ "cpc": 0.1884770856669751,
16
+ "cps": 30.234375,
17
+ "bpb": 2.2085573468042186,
18
+ "bytes_per_chunk": 5.45974090890157
19
+ },
20
+ "vi": {
21
+ "cpc": 0.2385679414157852,
22
+ "cps": 34.359375,
23
+ "bpb": 1.9038478483110506,
24
+ "bytes_per_chunk": 5.867275453792609
25
+ },
26
+ "tr": {
27
+ "cpc": 0.1954924598132906,
28
+ "cps": 27.6484375,
29
+ "bpb": 2.1745089955083285,
30
+ "bytes_per_chunk": 5.644971306136072
31
+ },
32
+ "fi": {
33
+ "cpc": 0.18829408324916272,
34
+ "cps": 27.671875,
35
+ "bpb": 2.3741402975011927,
36
+ "bytes_per_chunk": 5.352663120603579
37
+ },
38
+ "ru": {
39
+ "cpc": 0.32854881266490765,
40
+ "cps": 48.640625,
41
+ "bpb": 1.3188029527823422,
42
+ "bytes_per_chunk": 5.544969963617904
43
+ },
44
+ "el": {
45
+ "cpc": 0.3258295570822979,
46
+ "cps": 52.2421875,
47
+ "bpb": 1.4323273247556754,
48
+ "bytes_per_chunk": 5.576997947005074
49
+ },
50
+ "hi": {
51
+ "cpc": 0.4302526002971768,
52
+ "cps": 56.5546875,
53
+ "bpb": 1.104362611684646,
54
+ "bytes_per_chunk": 5.94869234696772
55
+ },
56
+ "ta": {
57
+ "cpc": 0.4440759561380048,
58
+ "cps": 64.859375,
59
+ "bpb": 0.9217435379199496,
60
+ "bytes_per_chunk": 6.291934186998212
61
+ },
62
+ "th": {
63
+ "cpc": 0.4862915760024868,
64
+ "cps": 61.109375,
65
+ "bpb": 1.0760360586894675,
66
+ "bytes_per_chunk": 6.057351480000924
67
+ },
68
+ "zh": {
69
+ "cpc": 0.513376227565188,
70
+ "cps": 23.6875,
71
+ "bpb": 3.080435111671317,
72
+ "bytes_per_chunk": 5.348949671995674
73
+ },
74
+ "my": {
75
+ "cpc": 0.5291851382731687,
76
+ "cps": 78.3359375,
77
+ "bpb": 1.1029825906473758,
78
+ "bytes_per_chunk": 5.396852193068237
79
+ }
80
+ },
81
+ "gini_cpc": 0.2186565677816319,
82
+ "gini_cps": 0.2229964643488509,
83
+ "bpic": 5.653815556865056,
84
+ "seed": 2,
85
+ "mean_boundary_entropy": 0.328884482383728,
86
+ "dechunk": "ema",
87
+ "router_noise": 0.0,
88
+ "transitions": [
89
+ {
90
+ "step": 0,
91
+ "phase": 1,
92
+ "objective": "global",
93
+ "router_trainable": true,
94
+ "encoder_trainable": true,
95
+ "frozen_router_params": 0
96
+ },
97
+ {
98
+ "step": 2400,
99
+ "phase": 2,
100
+ "objective": "global",
101
+ "router_trainable": true,
102
+ "encoder_trainable": true,
103
+ "frozen_router_params": 0
104
+ },
105
+ {
106
+ "step": 4200,
107
+ "phase": 3,
108
+ "objective": "global",
109
+ "router_trainable": true,
110
+ "encoder_trainable": true,
111
+ "frozen_router_params": 0
112
+ }
113
+ ]
114
+ }
log.jsonl ADDED
The diff for this file is too large to render. See raw diff