Hugging Face's logo

AdaptiveChunking
/
pilot_A_s1

PyTorch
Safetensors
dynamic-chunking
tokenization
hnet
byte-level
beetle-hnet
legacy-trainer
Model card Files
xet
 Community
suchirsalhan commited on
Commit
e7737b8
·
verified ·
1 Parent(s): 580efb3

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_s1",
4
+ "condition": "A",
5
+ "step": 6000,
6
+ "params": 22516992,
7
+ "per_lang": {
8
+ "en": {
9
+ "cpc": 0.16594839970120095,
10
+ "cps": 22.5625,
11
+ "bpb": 2.022311981224875,
12
+ "bytes_per_chunk": 5.737196883480697
13
+ },
14
+ "de": {
15
+ "cpc": 0.1613987240052598,
16
+ "cps": 25.890625,
17
+ "bpb": 2.212188749728271,
18
+ "bytes_per_chunk": 6.239741026373417
19
+ },
20
+ "vi": {
21
+ "cpc": 0.2288039056143206,
22
+ "cps": 32.953125,
23
+ "bpb": 1.8670015220171892,
24
+ "bytes_per_chunk": 5.71686530220589
25
+ },
26
+ "tr": {
27
+ "cpc": 0.1896370767276142,
28
+ "cps": 26.8203125,
29
+ "bpb": 2.1998532618483893,
30
+ "bytes_per_chunk": 5.7227934595149215
31
+ },
32
+ "fi": {
33
+ "cpc": 0.1615544096539259,
34
+ "cps": 23.7421875,
35
+ "bpb": 2.395499962296907,
36
+ "bytes_per_chunk": 6.238033481266434
37
+ },
38
+ "ru": {
39
+ "cpc": 0.34649076517150396,
40
+ "cps": 51.296875,
41
+ "bpb": 1.2738760323866076,
42
+ "bytes_per_chunk": 5.32029022274088
43
+ },
44
+ "el": {
45
+ "cpc": 0.3327973493154022,
46
+ "cps": 53.359375,
47
+ "bpb": 1.3333722856930763,
48
+ "bytes_per_chunk": 5.615466181117121
49
+ },
50
+ "hi": {
51
+ "cpc": 0.42906389301634473,
52
+ "cps": 56.3984375,
53
+ "bpb": 1.140567442690789,
54
+ "bytes_per_chunk": 5.7856937915204485
55
+ },
56
+ "ta": {
57
+ "cpc": 0.4700722118213426,
58
+ "cps": 68.65625,
59
+ "bpb": 0.9174700377689008,
60
+ "bytes_per_chunk": 5.817278032976056
61
+ },
62
+ "th": {
63
+ "cpc": 0.47118433322971714,
64
+ "cps": 59.2109375,
65
+ "bpb": 1.0642222156874839,
66
+ "bytes_per_chunk": 6.21260560960292
67
+ },
68
+ "zh": {
69
+ "cpc": 0.502539790044023,
70
+ "cps": 23.1875,
71
+ "bpb": 3.109084046145538,
72
+ "bytes_per_chunk": 5.7431043925950265
73
+ },
74
+ "my": {
75
+ "cpc": 0.41803884314967277,
76
+ "cps": 61.8828125,
77
+ "bpb": 1.0778031266825046,
78
+ "bytes_per_chunk": 6.518967982592478
79
+ }
80
+ },
81
+ "gini_cpc": 0.22451192807644027,
82
+ "gini_cps": 0.22301957393367622,
83
+ "bpic": 5.925929929126198,
84
+ "seed": 1,
85
+ "mean_boundary_entropy": 0.24016138911247253,
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