Hugging Face's logo

AdaptiveChunking
/
spec_E_s2

PyTorch
Safetensors
dynamic-chunking
tokenization
hnet
byte-level
beetle-hnet
legacy-trainer
Model card Files
xet
 Community
suchirsalhan commited on
Commit
7032b51
·
verified ·
1 Parent(s): 1d17dda

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_E_s2",
4
+ "condition": "E",
5
+ "step": 11000,
6
+ "params": 82470528,
7
+ "per_lang": {
8
+ "en": {
9
+ "cpc": 0.18381888180198816,
10
+ "cps": 24.9921875,
11
+ "bpb": 1.6438765837477893,
12
+ "bytes_per_chunk": 5.193541357107479
13
+ },
14
+ "de": {
15
+ "cpc": 0.1784931573564506,
16
+ "cps": 28.6328125,
17
+ "bpb": 1.7824091142168186,
18
+ "bytes_per_chunk": 5.5738851708722486
19
+ },
20
+ "vi": {
21
+ "cpc": 0.18014646053702196,
22
+ "cps": 25.9453125,
23
+ "bpb": 1.4739747294799386,
24
+ "bytes_per_chunk": 7.134480482809768
25
+ },
26
+ "tr": {
27
+ "cpc": 0.1762138872010164,
28
+ "cps": 24.921875,
29
+ "bpb": 1.6789048384437173,
30
+ "bytes_per_chunk": 5.853976619287413
31
+ },
32
+ "fi": {
33
+ "cpc": 0.1817021955238956,
34
+ "cps": 26.703125,
35
+ "bpb": 1.8691479514912825,
36
+ "bytes_per_chunk": 5.733808700977712
37
+ },
38
+ "ru": {
39
+ "cpc": 0.19593667546174143,
40
+ "cps": 29.0078125,
41
+ "bpb": 1.0907239198933933,
42
+ "bytes_per_chunk": 9.00026608300073
43
+ },
44
+ "el": {
45
+ "cpc": 0.1705890951615261,
46
+ "cps": 27.3515625,
47
+ "bpb": 1.095154461702511,
48
+ "bytes_per_chunk": 10.631410321403225
49
+ },
50
+ "hi": {
51
+ "cpc": 0.2234769687964339,
52
+ "cps": 29.375,
53
+ "bpb": 0.907511446864184,
54
+ "bytes_per_chunk": 11.288362579394983
55
+ },
56
+ "ta": {
57
+ "cpc": 0.17865739502540787,
58
+ "cps": 26.09375,
59
+ "bpb": 0.7897428494116536,
60
+ "bytes_per_chunk": 14.990150248030764
61
+ },
62
+ "th": {
63
+ "cpc": 0.2431457880012434,
64
+ "cps": 30.5546875,
65
+ "bpb": 0.8658079972935708,
66
+ "bytes_per_chunk": 11.621403555437336
67
+ },
68
+ "zh": {
69
+ "cpc": 0.5130375888926515,
70
+ "cps": 23.671875,
71
+ "bpb": 2.479610840172693,
72
+ "bytes_per_chunk": 5.378995480637533
73
+ },
74
+ "my": {
75
+ "cpc": 0.08982478361832383,
76
+ "cps": 13.296875,
77
+ "bpb": 0.8824134811779903,
78
+ "bytes_per_chunk": 27.510126980795466
79
+ }
80
+ },
81
+ "gini_cpc": 0.19030326488159188,
82
+ "gini_cps": 0.07857023060796653,
83
+ "bpic": 9.654867924528302,
84
+ "seed": 2,
85
+ "mean_boundary_entropy": 0.10425905883312225,
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_b",
101
+ "router_trainable": true,
102
+ "encoder_trainable": true,
103
+ "frozen_router_params": 0
104
+ },
105
+ {
106
+ "step": 7700,
107
+ "phase": 3,
108
+ "objective": "none",
109
+ "router_trainable": false,
110
+ "encoder_trainable": true,
111
+ "frozen_router_params": 295680
112
+ }
113
+ ]
114
+ }
log.jsonl ADDED
The diff for this file is too large to render. See raw diff