Hugging Face's logo

AdaptiveChunking
/
spec_C_s1

PyTorch
Safetensors
dynamic-chunking
tokenization
hnet
byte-level
beetle-hnet
legacy-trainer
Model card Files
xet
 Community
suchirsalhan commited on
Commit
13d87f1
·
verified ·
1 Parent(s): c922752

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": "spec_C_s1",
4
+ "condition": "C",
5
+ "step": 11000,
6
+ "params": 82470528,
7
+ "per_lang": {
8
+ "en": {
9
+ "cpc": 0.1829569614434293,
10
+ "cps": 24.875,
11
+ "bpb": 1.6195725652203883,
12
+ "bytes_per_chunk": 5.149443350406868
13
+ },
14
+ "de": {
15
+ "cpc": 0.17211318365557882,
16
+ "cps": 27.609375,
17
+ "bpb": 1.7585520463522628,
18
+ "bytes_per_chunk": 5.525451594545033
19
+ },
20
+ "vi": {
21
+ "cpc": 0.21442907512883103,
22
+ "cps": 30.8828125,
23
+ "bpb": 1.505241884835232,
24
+ "bytes_per_chunk": 6.172669154785312
25
+ },
26
+ "tr": {
27
+ "cpc": 0.18090924156217203,
28
+ "cps": 25.5859375,
29
+ "bpb": 1.6821185492534023,
30
+ "bytes_per_chunk": 5.745779336420176
31
+ },
32
+ "fi": {
33
+ "cpc": 0.169847429695391,
34
+ "cps": 24.9609375,
35
+ "bpb": 1.8778386325492604,
36
+ "bytes_per_chunk": 5.858588341779295
37
+ },
38
+ "ru": {
39
+ "cpc": 0.283377308707124,
40
+ "cps": 41.953125,
41
+ "bpb": 1.092111857507061,
42
+ "bytes_per_chunk": 6.115251153269687
43
+ },
44
+ "el": {
45
+ "cpc": 0.3028796959508844,
46
+ "cps": 48.5625,
47
+ "bpb": 1.0907878436949359,
48
+ "bytes_per_chunk": 5.791957578435705
49
+ },
50
+ "hi": {
51
+ "cpc": 0.4353046062407132,
52
+ "cps": 57.21875,
53
+ "bpb": 0.9181170633737359,
54
+ "bytes_per_chunk": 5.804590192974104
55
+ },
56
+ "ta": {
57
+ "cpc": 0.4493179994650976,
58
+ "cps": 65.625,
59
+ "bpb": 0.7916944870358134,
60
+ "bytes_per_chunk": 6.1309477869379645
61
+ },
62
+ "th": {
63
+ "cpc": 0.46950575069940936,
64
+ "cps": 59.0,
65
+ "bpb": 0.8850112952426465,
66
+ "bytes_per_chunk": 6.0923341506211
67
+ },
68
+ "zh": {
69
+ "cpc": 0.606501862512699,
70
+ "cps": 27.984375,
71
+ "bpb": 2.5927716945666806,
72
+ "bytes_per_chunk": 4.49993991932023
73
+ },
74
+ "my": {
75
+ "cpc": 0.47334811061853493,
76
+ "cps": 70.0703125,
77
+ "bpb": 0.8808864740932686,
78
+ "bytes_per_chunk": 5.883307430328398
79
+ }
80
+ },
81
+ "gini_cpc": 0.24422553299626815,
82
+ "gini_cps": 0.21776084105296856,
83
+ "bpic": 5.945115717074078,
84
+ "seed": 1,
85
+ "mean_boundary_entropy": 0.2494300752878189,
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": 4400,
99
+ "phase": 2,
100
+ "objective": "parity_a",
101
+ "router_trainable": true,
102
+ "encoder_trainable": true,
103
+ "frozen_router_params": 0
104
+ },
105
+ {
106
+ "step": 7700,
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