Hugging Face's logo

AdaptiveChunking
/
pilot_Bpc_s0

PyTorch
Safetensors
dynamic-chunking
tokenization
hnet
byte-level
beetle-hnet
legacy-trainer
Model card Files
xet
 Community
suchirsalhan commited on
Commit
01a2aa1
·
verified ·
1 Parent(s): 7d40208

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_Bpc_s0",
4
+ "condition": "B",
5
+ "step": 6000,
6
+ "params": 22516992,
7
+ "per_lang": {
8
+ "en": {
9
+ "cpc": 0.211630178704821,
10
+ "cps": 28.7734375,
11
+ "bpb": 2.0236759541310367,
12
+ "bytes_per_chunk": 4.849803431848666
13
+ },
14
+ "de": {
15
+ "cpc": 0.21010081332489164,
16
+ "cps": 33.703125,
17
+ "bpb": 2.184274127428656,
18
+ "bytes_per_chunk": 4.781991645232492
19
+ },
20
+ "vi": {
21
+ "cpc": 0.1920260374288039,
22
+ "cps": 27.65625,
23
+ "bpb": 1.907216450995669,
24
+ "bytes_per_chunk": 6.8501247762520086
25
+ },
26
+ "tr": {
27
+ "cpc": 0.20410981605258796,
28
+ "cps": 28.8671875,
29
+ "bpb": 2.1666493561171514,
30
+ "bytes_per_chunk": 5.170697068917906
31
+ },
32
+ "fi": {
33
+ "cpc": 0.21577800223273616,
34
+ "cps": 31.7109375,
35
+ "bpb": 2.349011138461699,
36
+ "bytes_per_chunk": 4.563708849079926
37
+ },
38
+ "ru": {
39
+ "cpc": 0.21308707124010554,
40
+ "cps": 31.546875,
41
+ "bpb": 1.3587628852341709,
42
+ "bytes_per_chunk": 8.085748214864052
43
+ },
44
+ "el": {
45
+ "cpc": 0.22574672318861766,
46
+ "cps": 36.1953125,
47
+ "bpb": 1.4326475242048349,
48
+ "bytes_per_chunk": 8.065721054736777
49
+ },
50
+ "hi": {
51
+ "cpc": 0.28053491827637445,
52
+ "cps": 36.875,
53
+ "bpb": 1.1991079776574975,
54
+ "bytes_per_chunk": 9.450547073562017
55
+ },
56
+ "ta": {
57
+ "cpc": 0.22877774806097886,
58
+ "cps": 33.4140625,
59
+ "bpb": 0.9890050605123927,
60
+ "bytes_per_chunk": 12.070912188608004
61
+ },
62
+ "th": {
63
+ "cpc": 0.2103823437985701,
64
+ "cps": 26.4375,
65
+ "bpb": 1.1282025457776095,
66
+ "bytes_per_chunk": 13.016410536507857
67
+ },
68
+ "zh": {
69
+ "cpc": 0.16915001693193363,
70
+ "cps": 7.8046875,
71
+ "bpb": 3.135393780091879,
72
+ "bytes_per_chunk": 15.04542715298304
73
+ },
74
+ "my": {
75
+ "cpc": 0.13579269579902892,
76
+ "cps": 20.1015625,
77
+ "bpb": 1.1188346722829436,
78
+ "bytes_per_chunk": 20.079200337022712
79
+ }
80
+ },
81
+ "gini_cpc": 0.08120721436107003,
82
+ "gini_cps": 0.13498994269232245,
83
+ "bpic": 8.73917795741774,
84
+ "seed": 0,
85
+ "mean_boundary_entropy": 0.30649498105049133,
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": "parity_a",
101
+ "router_trainable": true,
102
+ "encoder_trainable": true,
103
+ "frozen_router_params": 0
104
+ },
105
+ {
106
+ "step": 4200,
107
+ "phase": 3,
108
+ "objective": "none",
109
+ "router_trainable": false,
110
+ "encoder_trainable": true,
111
+ "frozen_router_params": 131584
112
+ }
113
+ ]
114
+ }
log.jsonl ADDED
The diff for this file is too large to render. See raw diff