File size: 2,347 Bytes
cbacfbc
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
[16:37:39] longeval: T=32768 prompt_len=32744 max_new=24 n=100 greedy=1 gpus=0,1
[16:37:39] haystack: /root/Adhi/BluTrain/dist/Context_Parallelism/lendata/haystack/b2val_haystack.bin
[16:37:39] prompt signature changed -- regenerating prompts and clearing state
[16:37:39] generating prompts: niah_single
[16:37:40] fan-out: 2 GPU(s), NP=1 -> 2 concurrent job(s)
[16:37:40] 2 job(s) to run
[16:37:40]   [gpu 0] ext32k / niah_single
[16:37:40]   [gpu 1] l2 / niah_single
[16:48:52]   scoring ext32k / niah_single

=== niah_single  (prompt_len=32744, n=100) ===
  accuracy: 27/100 = 27.0%
    depth      acc   (needle position in context)
     0.00   0/20    0.0%
     0.25   0/20    0.0%
     0.50   2/20   10.0%
     0.75   8/20   40.0%
     1.00  17/20   85.0%

  --- 2 sample generations ---
  [MISS] expect=8056020
         got= “kitty-cat”.\n
  [MISS] expect=5343902
         got= 7. This number is unique to cactus, making it a unique identifier for each plant.\n

  wrote ext32k_len_32744/scores/ext32k_niah_single.csv
[16:48:52]   scoring l2 / niah_single

=== niah_single  (prompt_len=32744, n=100) ===
  accuracy: 11/100 = 11.0%
    depth      acc   (needle position in context)
     0.00   0/20    0.0%
     0.25   0/20    0.0%
     0.50   0/20    0.0%
     0.75   0/20    0.0%
     1.00  11/20   55.0%

  --- 2 sample generations ---
  [MISS] expect=8056020
         got= the number of the number of the number of the number of the number of the number of the number of the number of
  [MISS] expect=5343902
         got= the number for the cactus. This number is a unique identifier that identifies the plant, making it easier to identify the

  wrote ext32k_len_32744/scores/l2_niah_single.csv
[16:48:53] building summary

=== per-task accuracy %  (secondary/descriptive: ~10 per depth cell) ===
  task                 ext32k       l2
  niah_single            27.0     11.0

=== PRIMARY: accuracy % pooled by depth across tasks (+-95% Wilson) ===
  depth             ext32k              l2
  0.00           0.0+- 8.1       0.0+- 8.1
  0.25           0.0+- 8.1       0.0+- 8.1
  0.50          10.0+-13.7       0.0+- 8.1
  0.75          40.0+-19.7       0.0+- 8.1
  1.00          85.0+-15.4      55.0+-20.0

  wrote ext32k_len_32744/summary.csv and ext32k_len_32744/depth_summary.csv
[16:48:53] DONE -> ext32k_len_32744/summary.csv