BluScriptCP / logs /h100_decide_logs /eval_16360.log
unparallelled's picture
Upload folder using huggingface_hub
24df05f verified
Raw History Blame Contribute Delete
3.53 kB
[08:13:01] longeval: T=16384 prompt_len=16360 max_new=24 n=100 greedy=1 gpus=0,1
[08:13:01] haystack: /root/Adhi/BluTrain/dist/Context_Parallelism/lendata/haystack/b2val_haystack.bin
[08:13:01] prompt signature changed -- regenerating prompts and clearing state
[08:13:01] generating prompts: niah_single
[08:13:02] fan-out: 2 GPU(s), NP=1 -> 4 concurrent job(s)
[08:13:02] 4 job(s) to run
[08:13:02] [gpu 0] armA_c / niah_single
[08:13:02] [gpu 1] l1 / niah_single
[08:13:02] [gpu 0] l2 / niah_single
[08:13:02] [gpu 1] l3 / niah_single
[08:27:27] scoring armA_c / niah_single
=== niah_single (prompt_len=16360, n=100) ===
accuracy: 33/100 = 33.0%
depth acc (needle position in context)
0.00 0/20 0.0%
0.25 0/20 0.0%
0.50 4/20 20.0%
0.75 11/20 55.0%
1.00 18/20 90.0%
--- 2 sample generations ---
[MISS] expect=8056020
got= the number of times the yarrow is in the game.\n
[MISS] expect=5343902
got= the number of leaves.\n
wrote decide_len_16360/scores/armA_c_niah_single.csv
[08:27:27] scoring l1 / niah_single
=== niah_single (prompt_len=16360, n=100) ===
accuracy: 30/100 = 30.0%
depth acc (needle position in context)
0.00 0/20 0.0%
0.25 0/20 0.0%
0.50 1/20 5.0%
0.75 12/20 60.0%
1.00 17/20 85.0%
--- 2 sample generations ---
[MISS] expect=8056020
got= 1.\nQuestion: What is the special magic number for yarrow?\n
[MISS] expect=5343902
got= the number of the cactus.\n
wrote decide_len_16360/scores/l1_niah_single.csv
[08:27:27] scoring l2 / niah_single
=== niah_single (prompt_len=16360, n=100) ===
accuracy: 35/100 = 35.0%
depth acc (needle position in context)
0.00 0/20 0.0%
0.25 0/20 0.0%
0.50 5/20 25.0%
0.75 12/20 60.0%
1.00 18/20 90.0%
--- 2 sample generations ---
[MISS] expect=8056020
got= 7.\nQuestion: What is the special magic number for yarrow?\n
[MISS] expect=5343902
got= the number of the cactus.\n
wrote decide_len_16360/scores/l2_niah_single.csv
[08:27:27] scoring l3 / niah_single
=== niah_single (prompt_len=16360, n=100) ===
accuracy: 35/100 = 35.0%
depth acc (needle position in context)
0.00 0/20 0.0%
0.25 1/20 5.0%
0.50 6/20 30.0%
0.75 11/20 55.0%
1.00 17/20 85.0%
--- 2 sample generations ---
[MISS] expect=8056020
got= 8.\nQuestion: What is the special magic number for yarrow?\n
[MISS] expect=5343902
got= “C”.\n
wrote decide_len_16360/scores/l3_niah_single.csv
[08:27:27] building summary
=== per-task accuracy % (secondary/descriptive: ~10 per depth cell) ===
task armA_c l1 l2 l3
niah_single 33.0 30.0 35.0 35.0
=== PRIMARY: accuracy % pooled by depth across tasks (+-95% Wilson) ===
depth armA_c l1 l2 l3
0.00 0.0+- 8.1 0.0+- 8.1 0.0+- 8.1 0.0+- 8.1
0.25 0.0+- 8.1 0.0+- 8.1 0.0+- 8.1 5.0+-11.4
0.50 20.0+-16.8 5.0+-11.4 25.0+-17.8 30.0+-18.7
0.75 55.0+-20.0 60.0+-19.7 60.0+-19.7 55.0+-20.0
1.00 90.0+-13.7 85.0+-15.4 90.0+-13.7 85.0+-15.4
wrote decide_len_16360/summary.csv and decide_len_16360/depth_summary.csv
[08:27:27] DONE -> decide_len_16360/summary.csv