Hugging Face's logo

AbstractPhil
/
captionbert-8192-rotary

Feature Extraction
Transformers
TensorBoard
Safetensors
English
captionbert_structured_rotary
sentence-similarity
consensus-distillation
rotary-embeddings
custom_code
Model card Files
xet
 Metrics Community
AbstractPhil commited on
Commit
4e75920
·
verified ·
1 Parent(s): 1928c8c

Ship A3-seed2 (all arms retained)

Browse files
Files changed (2) hide show
  1. experiments/817daa9e66a8119dc6b18a9b63fad32490d1c891e5f26750e91125d9e5e14e6c/A3-seed2/analyses/captionbert-rotary-mechanisms-v1/rotary_diagnostics.json +0 -0
  2. experiments/817daa9e66a8119dc6b18a9b63fad32490d1c891e5f26750e91125d9e5e14e6c/A3-seed2/analyses/captionbert-rotary-mechanisms-v1/source_result.json +114 -0
experiments/817daa9e66a8119dc6b18a9b63fad32490d1c891e5f26750e91125d9e5e14e6c/A3-seed2/analyses/captionbert-rotary-mechanisms-v1/rotary_diagnostics.json ADDED
The diff for this file is too large to render. See raw diff
 
experiments/817daa9e66a8119dc6b18a9b63fad32490d1c891e5f26750e91125d9e5e14e6c/A3-seed2/analyses/captionbert-rotary-mechanisms-v1/source_result.json ADDED
@@ -0,0 +1,114 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "arm": "A3",
3
+ "batch_size": 2048,
4
+ "benchmark": {
5
+ "dataset_fingerprints": {
6
+ "BIOSSES": "f1568c59b08bf901",
7
+ "SICK-R": "ec39ecc7fe3d35c5",
8
+ "STS-B": "a5cf888e07361adf",
9
+ "STS12": "babcef74f2c87568",
10
+ "STS13": "0b95e9bf5b389b20",
11
+ "STS14": "60e6b4f337fc83b1",
12
+ "STS15": "2ce54f5616944e48",
13
+ "STS16": "5f0ab25e659613f2"
14
+ },
15
+ "dataset_revisions": {
16
+ "mteb/biosses-sts": "f4f1265a7a4b0c99e6da09859b887bdb1f36d8ed",
17
+ "mteb/sickr-sts": "b93b4e08b707e9c9bc478e543495c3ced0b9d81e",
18
+ "mteb/sts12-sts": "3a51344839ec620dd3b753e0945bdf6c7e37af3e",
19
+ "mteb/sts13-sts": "f475812ca75868066186343ebf60e01f95c75164",
20
+ "mteb/sts14-sts": "8ab5e540b20b58d42fa6b944d3accee9df3fbaca",
21
+ "mteb/sts15-sts": "ae8e1ecbd3b3bcf5c1c8fd798290033715d98b93",
22
+ "mteb/sts16-sts": "d8e3be175a943eecfea09c96795087771996197c",
23
+ "mteb/stsbenchmark-sts": "96943a16ea6a35129e253c659081cb59daf81b30"
24
+ },
25
+ "max_tokens": 64,
26
+ "mean": 0.5119058055770798,
27
+ "split": "test",
28
+ "tasks": {
29
+ "BIOSSES": {
30
+ "erank": 22.499313389092332,
31
+ "self_cos": 0.5144970868437793,
32
+ "spearman": 0.4168046779002283
33
+ },
34
+ "SICK-R": {
35
+ "erank": 32.69667592129721,
36
+ "self_cos": 0.3534099986818124,
37
+ "spearman": 0.6351555990383906
38
+ },
39
+ "STS-B": {
40
+ "erank": 30.86173992005158,
41
+ "self_cos": 0.16574616691650723,
42
+ "spearman": 0.48451588582040767
43
+ },
44
+ "STS12": {
45
+ "erank": 23.39586909066291,
46
+ "self_cos": 0.20808495528481685,
47
+ "spearman": 0.4527847991684722
48
+ },
49
+ "STS13": {
50
+ "erank": 26.496608738739667,
51
+ "self_cos": 0.1839827052222016,
52
+ "spearman": 0.4537737087035938
53
+ },
54
+ "STS14": {
55
+ "erank": 33.728987734221725,
56
+ "self_cos": 0.1289306871030488,
57
+ "spearman": 0.4750671770567304
58
+ },
59
+ "STS15": {
60
+ "erank": 27.97555468947359,
61
+ "self_cos": 0.21997887047403303,
62
+ "spearman": 0.6761810995006643
63
+ },
64
+ "STS16": {
65
+ "erank": 24.455906112111638,
66
+ "self_cos": 0.22248843671704485,
67
+ "spearman": 0.5009634974281512
68
+ }
69
+ }
70
+ },
71
+ "config_id": "817daa9e66a8119dc6b18a9b63fad32490d1c891e5f26750e91125d9e5e14e6c",
72
+ "dataset_id": "21d5e137bbcdb91a743746b7f2e6fb3660266fd4320d0ce13061c678645e7c64",
73
+ "elapsed_seconds": 961.8552551740004,
74
+ "max_tokens": 256,
75
+ "memory": {
76
+ "max_observed_device_gb": 34.90185216,
77
+ "peak_allocated_gb": 31.081823232,
78
+ "peak_reserved_gb": 34.198257664,
79
+ "total_device_limit_gb": 80.0
80
+ },
81
+ "model_config": {
82
+ "d_ff": 2048,
83
+ "d_model": 512,
84
+ "dropout": 0.1,
85
+ "grad_checkpointing": true,
86
+ "max_len": 8192,
87
+ "n_heads": 8,
88
+ "n_layers": 12,
89
+ "output_dim": 768,
90
+ "pad_token_id": 0,
91
+ "vocab_size": 30522
92
+ },
93
+ "parameters": 54114560,
94
+ "seconds_per_step": 0.46468605498155685,
95
+ "seed": 2,
96
+ "status": "complete",
97
+ "steps": 1952,
98
+ "validation": {
99
+ "above_128_tokens": {
100
+ "cos_to_target": 0.6833597421646118,
101
+ "erank": 52.04497791239141,
102
+ "n": 5000,
103
+ "self_cos": 0.061107762241960975
104
+ },
105
+ "above_128_total_rows": 29240,
106
+ "heldout": {
107
+ "cos_to_target": 0.7621751427650452,
108
+ "erank": 73.07822921530978,
109
+ "n": 5000,
110
+ "self_cos": 0.016853114730957516
111
+ },
112
+ "heldout_total_rows": 500000
113
+ }
114
+ }