Hugging Face's logo

AdaptiveChunking
/
specfull_A_s1

PyTorch
Safetensors
dynamic-chunking
tokenization
hnet
byte-level
beetle-hnet
legacy-trainer
Model card Files
xet
 Community
suchirsalhan commited on
Commit
773fc41
·
verified ·
1 Parent(s): ae4e049

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": "specfull_A_s1",
4
+ "condition": "A",
5
+ "step": 76300,
6
+ "params": 82470528,
7
+ "per_lang": {
8
+ "en": {
9
+ "cpc": 0.17491237143021318,
10
+ "cps": 23.78125,
11
+ "bpb": 1.4344827618211076,
12
+ "bytes_per_chunk": 5.6980393861670215
13
+ },
14
+ "de": {
15
+ "cpc": 0.17742171139141868,
16
+ "cps": 28.4609375,
17
+ "bpb": 1.5354745000520915,
18
+ "bytes_per_chunk": 5.662744166203131
19
+ },
20
+ "vi": {
21
+ "cpc": 0.22723081095741796,
22
+ "cps": 32.7265625,
23
+ "bpb": 1.226269461056695,
24
+ "bytes_per_chunk": 5.933846789390587
25
+ },
26
+ "tr": {
27
+ "cpc": 0.1846102855880241,
28
+ "cps": 26.109375,
29
+ "bpb": 1.4144888596260154,
30
+ "bytes_per_chunk": 5.9682966116647975
31
+ },
32
+ "fi": {
33
+ "cpc": 0.19095210249322206,
34
+ "cps": 28.0625,
35
+ "bpb": 1.5563237203400448,
36
+ "bytes_per_chunk": 5.352335255984891
37
+ },
38
+ "ru": {
39
+ "cpc": 0.30337730870712404,
40
+ "cps": 44.9140625,
41
+ "bpb": 0.8736008536716351,
42
+ "bytes_per_chunk": 5.916870746763872
43
+ },
44
+ "el": {
45
+ "cpc": 0.29547337133947277,
46
+ "cps": 47.375,
47
+ "bpb": 0.8757178472401416,
48
+ "bytes_per_chunk": 6.034101568119695
49
+ },
50
+ "hi": {
51
+ "cpc": 0.45426448736998515,
52
+ "cps": 59.7109375,
53
+ "bpb": 0.7227330449265795,
54
+ "bytes_per_chunk": 5.6788141697400984
55
+ },
56
+ "ta": {
57
+ "cpc": 0.4721583311045734,
58
+ "cps": 68.9609375,
59
+ "bpb": 0.6253251161082508,
60
+ "bytes_per_chunk": 5.932336101744778
61
+ },
62
+ "th": {
63
+ "cpc": 0.49132732359341,
64
+ "cps": 61.7421875,
65
+ "bpb": 0.7201171780580405,
66
+ "bytes_per_chunk": 5.851297962099061
67
+ },
68
+ "zh": {
69
+ "cpc": 0.6232644768032509,
70
+ "cps": 28.7578125,
71
+ "bpb": 2.163128118829892,
72
+ "bytes_per_chunk": 4.561009134406263
73
+ },
74
+ "my": {
75
+ "cpc": 0.5171522060375765,
76
+ "cps": 76.5546875,
77
+ "bpb": 0.6932816340289751,
78
+ "bytes_per_chunk": 5.569118990461218
79
+ }
80
+ },
81
+ "gini_cpc": 0.24787354735650236,
82
+ "gini_cps": 0.22545794060110258,
83
+ "bpic": 5.687666725947003,
84
+ "seed": 1,
85
+ "mean_boundary_entropy": 0.09993767738342285,
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": 30520,
99
+ "phase": 2,
100
+ "objective": "global",
101
+ "router_trainable": true,
102
+ "encoder_trainable": true,
103
+ "frozen_router_params": 0
104
+ },
105
+ {
106
+ "step": 53410,
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