Hugging Face's logo

dzungpham
/
graphcodebert-code-classification

Transformers
Safetensors
detection
AI-generated
bert
Model card Files
xet
 Community
dzungpham commited on
Commit
abf9e49
·
verified ·
1 Parent(s): 206f558

Upload folder using huggingface_hub

Browse files
Files changed (40) hide show
  1. semeval-2026-challenge.log +456 -0
  2. taskA-codebert-base-focal/checkpoint-600/config.json +28 -0
  3. taskA-codebert-base-focal/checkpoint-600/merges.txt +0 -0
  4. taskA-codebert-base-focal/checkpoint-600/model.safetensors +3 -0
  5. taskA-codebert-base-focal/checkpoint-600/optimizer.pt +3 -0
  6. taskA-codebert-base-focal/checkpoint-600/rng_state.pth +3 -0
  7. taskA-codebert-base-focal/checkpoint-600/scaler.pt +3 -0
  8. taskA-codebert-base-focal/checkpoint-600/scheduler.pt +3 -0
  9. taskA-codebert-base-focal/checkpoint-600/special_tokens_map.json +51 -0
  10. taskA-codebert-base-focal/checkpoint-600/tokenizer.json +0 -0
  11. taskA-codebert-base-focal/checkpoint-600/tokenizer_config.json +58 -0
  12. taskA-codebert-base-focal/checkpoint-600/trainer_state.json +502 -0
  13. taskA-codebert-base-focal/checkpoint-600/training_args.bin +3 -0
  14. taskA-codebert-base-focal/checkpoint-600/vocab.json +0 -0
  15. taskA-codebert-base-focal/checkpoint-800/config.json +28 -0
  16. taskA-codebert-base-focal/checkpoint-800/merges.txt +0 -0
  17. taskA-codebert-base-focal/checkpoint-800/model.safetensors +3 -0
  18. taskA-codebert-base-focal/checkpoint-800/optimizer.pt +3 -0
  19. taskA-codebert-base-focal/checkpoint-800/rng_state.pth +3 -0
  20. taskA-codebert-base-focal/checkpoint-800/scaler.pt +3 -0
  21. taskA-codebert-base-focal/checkpoint-800/scheduler.pt +3 -0
  22. taskA-codebert-base-focal/checkpoint-800/special_tokens_map.json +51 -0
  23. taskA-codebert-base-focal/checkpoint-800/tokenizer.json +0 -0
  24. taskA-codebert-base-focal/checkpoint-800/tokenizer_config.json +58 -0
  25. taskA-codebert-base-focal/checkpoint-800/trainer_state.json +655 -0
  26. taskA-codebert-base-focal/checkpoint-800/training_args.bin +3 -0
  27. taskA-codebert-base-focal/checkpoint-800/vocab.json +0 -0
  28. taskA-unixcoder-focal/checkpoint-2200/config.json +29 -0
  29. taskA-unixcoder-focal/checkpoint-2200/merges.txt +0 -0
  30. taskA-unixcoder-focal/checkpoint-2200/model.safetensors +3 -0
  31. taskA-unixcoder-focal/checkpoint-2200/optimizer.pt +3 -0
  32. taskA-unixcoder-focal/checkpoint-2200/rng_state.pth +3 -0
  33. taskA-unixcoder-focal/checkpoint-2200/scaler.pt +3 -0
  34. taskA-unixcoder-focal/checkpoint-2200/scheduler.pt +3 -0
  35. taskA-unixcoder-focal/checkpoint-2200/special_tokens_map.json +51 -0
  36. taskA-unixcoder-focal/checkpoint-2200/tokenizer.json +0 -0
  37. taskA-unixcoder-focal/checkpoint-2200/tokenizer_config.json +58 -0
  38. taskA-unixcoder-focal/checkpoint-2200/trainer_state.json +1869 -0
  39. taskA-unixcoder-focal/checkpoint-2200/training_args.bin +3 -0
  40. taskA-unixcoder-focal/checkpoint-2200/vocab.json +0 -0
semeval-2026-challenge.log ADDED
@@ -0,0 +1,456 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ [{"stream_name":"stderr","time":6.934681857,"data":"0.00s - Debugger warning: It seems that frozen modules are being used, which may\n"}
2
+ ,{"stream_name":"stderr","time":6.93477418,"data":"0.00s - make the debugger miss breakpoints. Please pass -Xfrozen_modules=off\n"}
3
+ ,{"stream_name":"stderr","time":6.93478288,"data":"0.00s - to python to disable frozen modules.\n"}
4
+ ,{"stream_name":"stderr","time":6.934787409,"data":"0.00s - Note: Debugging will proceed. Set PYDEVD_DISABLE_FILE_VALIDATION=1 to disable this validation.\n"}
5
+ ,{"stream_name":"stderr","time":7.564948339,"data":"0.00s - Debugger warning: It seems that frozen modules are being used, which may\n"}
6
+ ,{"stream_name":"stderr","time":7.565021886,"data":"0.00s - make the debugger miss breakpoints. Please pass -Xfrozen_modules=off\n"}
7
+ ,{"stream_name":"stderr","time":7.565029956,"data":"0.00s - to python to disable frozen modules.\n"}
8
+ ,{"stream_name":"stderr","time":7.565034449,"data":"0.00s - Note: Debugging will proceed. Set PYDEVD_DISABLE_FILE_VALIDATION=1 to disable this validation.\n"}
9
+ ,{"stream_name":"stdout","time":9.064594643,"data":"\u001b[2mUsing Python 3.12.12 environment at: /usr\u001b[0m\r\n"}
10
+ ,{"stream_name":"stdout","time":9.36997085,"data":"\u001b[37m⠋\u001b[0m \u001b[2mResolving dependencies... \u001b[0m\r\u001b[2K\u001b[37m⠙\u001b[0m \u001b[2mResolving dependencies... \u001b[0m\r\u001b[2K\u001b[37m⠋\u001b[0m \u001b[2mResolving dependencies... \u001b[0m\r\u001b[2K\u001b[37m⠙\u001b[0m \u001b[2mResolving dependencies... \u001b[0m\r\u001b[2K\u001b[37m⠙\u001b[0m \u001b[2mpip==26.0.1 \u001b[0m\r\u001b[2K\u001b[37m⠙\u001b[0m \u001b[2m \u001b[0m\r\u001b[2K\u001b[2mResolved \u001b[1m1 package\u001b[0m \u001b[2min 193ms\u001b[0m\u001b[0m\r\n"}
11
+ ,{"stream_name":"stdout","time":9.370014731,"data":"\u001b[37m⠋\u001b[0m \u001b[2mPreparing packages...\u001b[0m (0/0) \r\u001b[2K\u001b[37m⠋\u001b[0m \u001b[2mPreparing packages...\u001b[0m (0/1) \r\u001b[2K\u001b[37m⠙\u001b[0m \u001b[2mPreparing packages...\u001b[0m (0/1) \r\u001b[2K\u001b[37m⠙\u001b[0m \u001b[2mPreparing packages...\u001b[0m (0/1)\r\n"}
12
+ ,{"stream_name":"stdout","time":9.370024369,"data":"\u001b[2mpip \u001b[0m \u001b[32m\u001b[30m\u001b[2m------------------------------\u001b[0m\u001b[0m 0 B/1.70 MiB \u001b[1A\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[1A\u001b[37m⠙\u001b[0m \u001b[2mPreparing packages...\u001b[0m (0/1)\r\n"}
13
+ ,{"stream_name":"stdout","time":9.370029894,"data":"\u001b[2mpip \u001b[0m \u001b[32m\u001b[30m\u001b[2m------------------------------\u001b[0m\u001b[0m 16.00 KiB/1.70 MiB \u001b[1A\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[1A\u001b[37m⠙\u001b[0m \u001b[2mPreparing packages...\u001b[0m (0/1)\r\n"}
14
+ ,{"stream_name":"stdout","time":9.370035732,"data":"\u001b[2mpip \u001b[0m \u001b[32m\u001b[30m\u001b[2m------------------------------\u001b[0m\u001b[0m 32.00 KiB/1.70 MiB \u001b[1A\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[1A\u001b[37m⠙\u001b[0m \u001b[2mPreparing packages...\u001b[0m (0/1)\r\n"}
15
+ ,{"stream_name":"stdout","time":9.370041545,"data":"\u001b[2mpip \u001b[0m \u001b[32m\u001b[30m\u001b[2m------------------------------\u001b[0m\u001b[0m 48.00 KiB/1.70 MiB \u001b[1A\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[1A\u001b[37m⠙\u001b[0m \u001b[2mPreparing packages...\u001b[0m (0/1)\r\n"}
16
+ ,{"stream_name":"stdout","time":9.370070574,"data":"\u001b[2mpip \u001b[0m \u001b[32m-\u001b[30m\u001b[2m-----------------------------\u001b[0m\u001b[0m 60.49 KiB/1.70 MiB \u001b[1A\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[1A\u001b[37m⠙\u001b[0m \u001b[2mPreparing packages...\u001b[0m (0/1)\r\n"}
17
+ ,{"stream_name":"stdout","time":9.37007608,"data":"\u001b[2mpip \u001b[0m \u001b[32m-\u001b[30m\u001b[2m-----------------------------\u001b[0m\u001b[0m 76.49 KiB/1.70 MiB \u001b[1A\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[1A\u001b[37m⠙\u001b[0m \u001b[2mPreparing packages...\u001b[0m (0/1)\r\n"}
18
+ ,{"stream_name":"stdout","time":9.370081126,"data":"\u001b[2mpip \u001b[0m \u001b[32m-\u001b[30m\u001b[2m-----------------------------\u001b[0m\u001b[0m 92.49 KiB/1.70 MiB \u001b[1A\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[1A\u001b[37m⠙\u001b[0m \u001b[2mPreparing packages...\u001b[0m (0/1)\r\n"}
19
+ ,{"stream_name":"stdout","time":9.370086014,"data":"\u001b[2mpip \u001b[0m \u001b[32m-\u001b[30m\u001b[2m-----------------------------\u001b[0m\u001b[0m 108.49 KiB/1.70 MiB \u001b[1A\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[1A\u001b[37m⠙\u001b[0m \u001b[2mPreparing packages...\u001b[0m (0/1)\r\n"}
20
+ ,{"stream_name":"stdout","time":9.420437835,"data":"\u001b[2mpip \u001b[0m \u001b[32m--\u001b[30m\u001b[2m----------------------------\u001b[0m\u001b[0m 124.49 KiB/1.70 MiB \u001b[1A\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[1A\u001b[37m⠙\u001b[0m \u001b[2mPreparing packages...\u001b[0m (0/1)\r\n"}
21
+ ,{"stream_name":"stdout","time":9.420476405,"data":"\u001b[2mpip \u001b[0m \u001b[32m--\u001b[30m\u001b[2m----------------------------\u001b[0m\u001b[0m 140.49 KiB/1.70 MiB \u001b[1A\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[1A\u001b[37m⠙\u001b[0m \u001b[2mPreparing packages...\u001b[0m (0/1)\r\n"}
22
+ ,{"stream_name":"stdout","time":9.420483687,"data":"\u001b[2mpip \u001b[0m \u001b[32m--\u001b[30m\u001b[2m----------------------------\u001b[0m\u001b[0m 156.49 KiB/1.70 MiB \u001b[1A\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[1A\u001b[37m⠙\u001b[0m \u001b[2mPreparing packages...\u001b[0m (0/1)\r\n"}
23
+ ,{"stream_name":"stdout","time":9.420488965,"data":"\u001b[2mpip \u001b[0m \u001b[32m--\u001b[30m\u001b[2m----------------------------\u001b[0m\u001b[0m 172.49 KiB/1.70 MiB \u001b[1A\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[1A\u001b[37m⠙\u001b[0m \u001b[2mPreparing packages...\u001b[0m (0/1)\r\n"}
24
+ ,{"stream_name":"stdout","time":9.420494795,"data":"\u001b[2mpip \u001b[0m \u001b[32m---\u001b[30m\u001b[2m---------------------------\u001b[0m\u001b[0m 188.49 KiB/1.70 MiB \u001b[1A\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[1A\u001b[37m⠙\u001b[0m \u001b[2mPreparing packages...\u001b[0m (0/1)\r\n"}
25
+ ,{"stream_name":"stdout","time":9.420499343,"data":"\u001b[2mpip \u001b[0m \u001b[32m---\u001b[30m\u001b[2m---------------------------\u001b[0m\u001b[0m 204.49 KiB/1.70 MiB \u001b[1A\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[1A\u001b[37m⠙\u001b[0m \u001b[2mPreparing packages...\u001b[0m (0/1)\r\n"}
26
+ ,{"stream_name":"stdout","time":9.420503452,"data":"\u001b[2mpip \u001b[0m \u001b[32m---\u001b[30m\u001b[2m---------------------------\u001b[0m\u001b[0m 220.49 KiB/1.70 MiB \u001b[1A\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[1A\u001b[37m⠙\u001b[0m \u001b[2mPreparing packages...\u001b[0m (0/1)\r\n"}
27
+ ,{"stream_name":"stdout","time":9.420507875,"data":"\u001b[2mpip \u001b[0m \u001b[32m----\u001b[30m\u001b[2m--------------------------\u001b[0m\u001b[0m 236.49 KiB/1.70 MiB \u001b[1A\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[1A\u001b[37m⠙\u001b[0m \u001b[2mPreparing packages...\u001b[0m (0/1)\r\n"}
28
+ ,{"stream_name":"stdout","time":9.420512194,"data":"\u001b[2mpip \u001b[0m \u001b[32m----\u001b[30m\u001b[2m--------------------------\u001b[0m\u001b[0m 252.49 KiB/1.70 MiB \u001b[1A\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[1A\u001b[37m⠙\u001b[0m \u001b[2mPreparing packages...\u001b[0m (0/1)\r\n"}
29
+ ,{"stream_name":"stdout","time":9.471352073,"data":"\u001b[2mpip \u001b[0m \u001b[32m----\u001b[30m\u001b[2m--------------------------\u001b[0m\u001b[0m 268.49 KiB/1.70 MiB \u001b[1A\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[1A\u001b[37m⠙\u001b[0m \u001b[2mPreparing packages...\u001b[0m (0/1)\r\n"}
30
+ ,{"stream_name":"stdout","time":9.522502958,"data":"\u001b[2mpip \u001b[0m \u001b[32m---------------------\u001b[30m\u001b[2m---------\u001b[0m\u001b[0m 1.24 MiB/1.70 MiB \u001b[1A\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[1A\u001b[37m⠙\u001b[0m \u001b[2mPreparing packages...\u001b[0m (0/1) \r\u001b[2K\u001b[2mPrepared \u001b[1m1 package\u001b[0m \u001b[2min 146ms\u001b[0m\u001b[0m\r\n"}
31
+ ,{"stream_name":"stdout","time":9.621699951,"data":"\u001b[2mUninstalled \u001b[1m1 package\u001b[0m \u001b[2min 118ms\u001b[0m\u001b[0m\r\n"}
32
+ ,{"stream_name":"stdout","time":9.621738457,"data":"░░░░░░░░░░░░░░░░░░░░ [0/0] \u001b[2mInstalling wheels... \u001b[0m\r\u001b[2K░░░░░░░░░░░░░░░░░░░░ [0/1] \u001b[2mInstalling wheels... \u001b[0m\r\u001b[2K░░░░░░░░░░░░░░░░░░░░ [0/1] \u001b[2mpip==26.0.1 \u001b[0m\r\u001b[2K████████████████████ [1/1] \u001b[2mpip==26.0.1 \u001b[0m\r\u001b[2K\u001b[2mInstalled \u001b[1m1 package\u001b[0m \u001b[2min 19ms\u001b[0m\u001b[0m\r\n"}
33
+ ,{"stream_name":"stdout","time":9.62175176,"data":" \u001b[31m-\u001b[39m \u001b[1mpip\u001b[0m\u001b[2m==24.1.2\u001b[0m\r\n"}
34
+ ,{"stream_name":"stdout","time":9.621755164,"data":" \u001b[32m+\u001b[39m \u001b[1mpip\u001b[0m\u001b[2m==26.0.1\u001b[0m\r\n"}
35
+ ,{"stream_name":"stdout","time":9.883503858,"data":"\u001b[2mUsing Python 3.12.12 environment at: /usr\u001b[0m\r\n"}
36
+ ,{"stream_name":"stdout","time":9.95107307,"data":"\u001b[2mAudited \u001b[1m3 packages\u001b[0m \u001b[2min 206ms\u001b[0m\u001b[0m\r\n"}
37
+ ,{"stream_name":"stdout","time":10.212319219,"data":"\u001b[2mUsing Python 3.12.12 environment at: /usr\u001b[0m\r\n"}
38
+ ,{"stream_name":"stdout","time":10.567629065,"data":"\u001b[2mUninstalled \u001b[1m2 packages\u001b[0m \u001b[2min 495ms\u001b[0m\u001b[0m\r\n"}
39
+ ,{"stream_name":"stdout","time":10.567688807,"data":" \u001b[31m-\u001b[39m \u001b[1mpandas\u001b[0m\u001b[2m==2.3.3\u001b[0m\r\n"}
40
+ ,{"stream_name":"stdout","time":10.5676969,"data":" \u001b[31m-\u001b[39m \u001b[1mtorchvision\u001b[0m\u001b[2m==0.25.0+cu128\u001b[0m\r\n"}
41
+ ,{"stream_name":"stdout","time":14.321278121,"data":"\u001b[2mUsing Python 3.12.12 environment at: /usr\u001b[0m\r\n"}
42
+ ,{"stream_name":"stdout","time":14.331177011,"data":"\u001b[2mAudited \u001b[1m2 packages\u001b[0m \u001b[2min 151ms\u001b[0m\u001b[0m\r\n"}
43
+ ,{"stream_name":"stdout","time":14.59154696,"data":"\u001b[2mUsing Python 3.12.12 environment at: /usr\u001b[0m\r\n"}
44
+ ,{"stream_name":"stdout","time":14.744366572,"data":"\u001b[37m⠋\u001b[0m \u001b[2mResolving dependencies... \u001b[0m\r\u001b[2K\u001b[37m⠙\u001b[0m \u001b[2mResolving dependencies... \u001b[0m\r\u001b[2K\u001b[37m⠋\u001b[0m \u001b[2mResolving dependencies... \u001b[0m\r\u001b[2K\u001b[37m⠙\u001b[0m \u001b[2mResolving dependencies... \u001b[0m\r\u001b[2K\u001b[37m⠙\u001b[0m \u001b[2mbitsandbytes==0.49.2 \u001b[0m\r\u001b[2K\u001b[37m⠙\u001b[0m \u001b[2maccelerate==1.13.0 \u001b[0m\r\u001b[2K\u001b[37m⠙\u001b[0m \u001b[2mhf-transfer==0.1.9 \u001b[0m\r\u001b[2K\u001b[37m⠙\u001b[0m \u001b[2mtorch==2.10.0+cu128 \u001b[0m\r\u001b[2K\u001b[37m⠙\u001b[0m \u001b[2mcuda-bindings==12.9.4 \u001b[0m\r\u001b[2K\u001b[37m⠙\u001b[0m \u001b[2mcuda-bindings==12.9.4 \u001b[0m\r\u001b[2K\u001b[37m⠙\u001b[0m \u001b[2mnvidia-cuda-nvrtc-cu12==12.8.93 \u001b[0m\r\u001b[2K\u001b[37m⠙\u001b[0m \u001b[2mnvidia-cuda-nvrtc-cu12==12.8.93 \u001b[0m\r\u001b[2K\u001b[37m⠙\u001b[0m \u001b[2mnvidia-cuda-runtime-cu12==12.8.90 \u001b[0m\r\u001b[2K\u001b[37m⠙\u001b[0m \u001b[2mnvidia-cuda-runtime-cu12==12.8.90 \u001b[0m\r\u001b[2K\u001b[37m⠙\u001b[0m \u001b[2mnvidia-cuda-cupti-cu12==12.8.90 \u001b[0m\r\u001b[2K\u001b[37m⠙\u001b[0m \u001b[2mnvidia-cuda-cupti-cu12==12.8.90 \u001b[0m\r\u001b[2K\u001b[37m⠙\u001b[0m \u001b[2mnvidia-cudnn-cu12==9.10.2.21 \u001b[0m\r\u001b[2K\u001b[37m⠙\u001b[0m \u001b[2mnvidia-cudnn-cu12==9.10.2.21 \u001b[0m\r\u001b[2K\u001b[37m⠙\u001b[0m \u001b[2mnvidia-cublas-cu12==12.8.4.1 \u001b[0m\r\u001b[2K\u001b[37m⠙\u001b[0m \u001b[2mnvidia-cublas-cu12==12.8.4.1 \u001b[0m\r\u001b[2K\u001b[37m⠙\u001b[0m \u001b[2mnvidia-cufft-cu12==11.3.3.83 \u001b[0m\r\u001b[2K\u001b[37m⠙\u001b[0m \u001b[2mnvidia-cufft-cu12==11.3.3.83 \u001b[0m\r\u001b[2K\u001b[37m⠙\u001b[0m \u001b[2mnvidia-curand-cu12==10.3.9.90 \u001b[0m\r\u001b[2K\u001b[37m⠙\u001b[0m \u001b[2mnvidia-curand-cu12==10.3.9.90 \u001b[0m\r\u001b[2K\u001b[2mResolved \u001b[1m44 packages\u001b[0m \u001b[2min 145ms\u001b[0m\u001b[0m\r\n"}
45
+ ,{"stream_name":"stdout","time":14.744449589,"data":"\u001b[37m⠋\u001b[0m \u001b[2mPreparing packages...\u001b[0m (0/0) \r\u001b[2K\u001b[37m⠙\u001b[0m \u001b[2mPreparing packages...\u001b[0m (0/0) \r\u001b[2K\u001b[37m⠙\u001b[0m \u001b[2mPreparing packages...\u001b[0m (0/3) \r\u001b[2K\u001b[37m⠙\u001b[0m \u001b[2mPreparing packages...\u001b[0m (0/3)\r\n"}
46
+ ,{"stream_name":"stdout","time":14.744458505,"data":"\u001b[2maccelerate \u001b[0m \u001b[32m\u001b[30m\u001b[2m------------------------------\u001b[0m\u001b[0m 0 B/374.75 KiB \u001b[1A\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[1A\u001b[37m⠙\u001b[0m \u001b[2mPreparing packages...\u001b[0m (0/3)\r\n"}
47
+ ,{"stream_name":"stdout","time":14.796469306,"data":"\u001b[2maccelerate \u001b[0m \u001b[32m-\u001b[30m\u001b[2m-----------------------------\u001b[0m\u001b[0m 16.00 KiB/374.75 KiB \u001b[1A\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[1A\u001b[37m⠙\u001b[0m \u001b[2mPreparing packages...\u001b[0m (0/3)\r\n"}
48
+ ,{"stream_name":"stdout","time":14.79650495,"data":"\u001b[2maccelerate \u001b[0m \u001b[32m--\u001b[30m\u001b[2m----------------------------\u001b[0m\u001b[0m 32.00 KiB/374.75 KiB \u001b[1A\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[1A\u001b[37m⠙\u001b[0m \u001b[2mPreparing packages...\u001b[0m (0/3)\r\n"}
49
+ ,{"stream_name":"stdout","time":14.796520342,"data":"\u001b[2maccelerate \u001b[0m \u001b[32m---\u001b[30m\u001b[2m---------------------------\u001b[0m\u001b[0m 48.00 KiB/374.75 KiB \u001b[1A\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[1A\u001b[37m⠙\u001b[0m \u001b[2mPreparing packages...\u001b[0m (0/3)\r\n"}
50
+ ,{"stream_name":"stdout","time":14.796525757,"data":"\u001b[2maccelerate \u001b[0m \u001b[32m-----\u001b[30m\u001b[2m-------------------------\u001b[0m\u001b[0m 64.00 KiB/374.75 KiB \u001b[1A\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[1A\u001b[37m⠙\u001b[0m \u001b[2mPreparing packages...\u001b[0m (0/3)\r\n"}
51
+ ,{"stream_name":"stdout","time":14.796530738,"data":"\u001b[2maccelerate \u001b[0m \u001b[32m------\u001b[30m\u001b[2m------------------------\u001b[0m\u001b[0m 80.00 KiB/374.75 KiB \u001b[1A\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[1A\u001b[37m⠙\u001b[0m \u001b[2mPreparing packages...\u001b[0m (0/3)\r\n"}
52
+ ,{"stream_name":"stdout","time":14.796535724,"data":"\u001b[2maccelerate \u001b[0m \u001b[32m-------\u001b[30m\u001b[2m-----------------------\u001b[0m\u001b[0m 96.00 KiB/374.75 KiB \u001b[1A\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[1A\u001b[37m⠙\u001b[0m \u001b[2mPreparing packages...\u001b[0m (0/3)\r\n"}
53
+ ,{"stream_name":"stdout","time":14.796540629,"data":"\u001b[2maccelerate \u001b[0m \u001b[32m--------\u001b[30m\u001b[2m----------------------\u001b[0m\u001b[0m 112.00 KiB/374.75 KiB \u001b[1A\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[1A\u001b[37m⠙\u001b[0m \u001b[2mPreparing packages...\u001b[0m (0/3)\r\n"}
54
+ ,{"stream_name":"stdout","time":14.796545599,"data":"\u001b[2maccelerate \u001b[0m \u001b[32m----------\u001b[30m\u001b[2m--------------------\u001b[0m\u001b[0m 128.00 KiB/374.75 KiB \u001b[1A\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[1A\u001b[37m⠙\u001b[0m \u001b[2mPreparing packages...\u001b[0m (0/3)\r\n"}
55
+ ,{"stream_name":"stdout","time":14.796550469,"data":"\u001b[2maccelerate \u001b[0m \u001b[32m----------\u001b[30m\u001b[2m--------------------\u001b[0m\u001b[0m 128.00 KiB/374.75 KiB \u001b[1A\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[1A\u001b[37m⠙\u001b[0m \u001b[2mPreparing packages...\u001b[0m (0/3)\r\n"}
56
+ ,{"stream_name":"stdout","time":14.796555465,"data":"\u001b[2maccelerate \u001b[0m \u001b[32m----------\u001b[30m\u001b[2m--------------------\u001b[0m\u001b[0m 128.00 KiB/374.75 KiB\r\n"}
57
+ ,{"stream_name":"stdout","time":14.796559844,"data":"\u001b[2mbitsandbytes \u001b[0m \u001b[32m\u001b[30m\u001b[2m------------------------------\u001b[0m\u001b[0m 0 B/57.84 MiB \u001b[2A\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[2A\u001b[37m⠙\u001b[0m \u001b[2mPreparing packages...\u001b[0m (0/3)\r\n"}
58
+ ,{"stream_name":"stdout","time":14.796565133,"data":"\u001b[2maccelerate \u001b[0m \u001b[32m----------\u001b[30m\u001b[2m--------------------\u001b[0m\u001b[0m 128.00 KiB/374.75 KiB\r\n"}
59
+ ,{"stream_name":"stdout","time":14.79657062,"data":"\u001b[2mbitsandbytes \u001b[0m \u001b[32m\u001b[30m\u001b[2m------------------------------\u001b[0m\u001b[0m 14.91 KiB/57.84 MiB \u001b[2A\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[2A\u001b[37m⠙\u001b[0m \u001b[2mPreparing packages...\u001b[0m (0/3)\r\n"}
60
+ ,{"stream_name":"stdout","time":14.796575829,"data":"\u001b[2maccelerate \u001b[0m \u001b[32m----------\u001b[30m\u001b[2m--------------------\u001b[0m\u001b[0m 128.00 KiB/374.75 KiB\r\n"}
61
+ ,{"stream_name":"stdout","time":14.796580013,"data":"\u001b[2mbitsandbytes \u001b[0m \u001b[32m\u001b[30m\u001b[2m------------------------------\u001b[0m\u001b[0m 14.91 KiB/57.84 MiB \u001b[2A\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[2A\u001b[37m⠙\u001b[0m \u001b[2mPreparing packages...\u001b[0m (0/3)\r\n"}
62
+ ,{"stream_name":"stdout","time":14.796588992,"data":"\u001b[2maccelerate \u001b[0m \u001b[32m----------\u001b[30m\u001b[2m--------------------\u001b[0m\u001b[0m 128.00 KiB/374.75 KiB\r\n"}
63
+ ,{"stream_name":"stdout","time":14.796593858,"data":"\u001b[2mbitsandbytes \u001b[0m \u001b[32m\u001b[30m\u001b[2m------------------------------\u001b[0m\u001b[0m 14.91 KiB/57.84 MiB \u001b[2A\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[2A\u001b[37m⠙\u001b[0m \u001b[2mPreparing packages...\u001b[0m (0/3)\r\n"}
64
+ ,{"stream_name":"stdout","time":14.796599062,"data":"\u001b[2maccelerate \u001b[0m \u001b[32m----------\u001b[30m\u001b[2m--------------------\u001b[0m\u001b[0m 128.00 KiB/374.75 KiB\r\n"}
65
+ ,{"stream_name":"stdout","time":14.796603362,"data":"\u001b[2mhf-transfer \u001b[0m \u001b[32m\u001b[30m\u001b[2m------------------------------\u001b[0m\u001b[0m 0 B/3.39 MiB\r\n"}
66
+ ,{"stream_name":"stdout","time":14.796607582,"data":"\u001b[2mbitsandbytes \u001b[0m \u001b[32m\u001b[30m\u001b[2m------------------------------\u001b[0m\u001b[0m 14.91 KiB/57.84 MiB \u001b[3A\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[3A\u001b[37m⠙\u001b[0m \u001b[2mPreparing packages...\u001b[0m (0/3)\r\n"}
67
+ ,{"stream_name":"stdout","time":14.796612824,"data":"\u001b[2maccelerate \u001b[0m \u001b[32m-----------\u001b[30m\u001b[2m-------------------\u001b[0m\u001b[0m 144.00 KiB/374.75 KiB\r\n"}
68
+ ,{"stream_name":"stdout","time":14.796617041,"data":"\u001b[2mhf-transfer \u001b[0m \u001b[32m\u001b[30m\u001b[2m------------------------------\u001b[0m\u001b[0m 0 B/3.39 MiB\r\n"}
69
+ ,{"stream_name":"stdout","time":14.796621255,"data":"\u001b[2mbitsandbytes \u001b[0m \u001b[32m\u001b[30m\u001b[2m------------------------------\u001b[0m\u001b[0m 14.91 KiB/57.84 MiB \u001b[3A\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[3A\u001b[37m⠙\u001b[0m \u001b[2mPreparing packages...\u001b[0m (0/3)\r\n"}
70
+ ,{"stream_name":"stdout","time":14.796626309,"data":"\u001b[2maccelerate \u001b[0m \u001b[32m------------\u001b[30m\u001b[2m------------------\u001b[0m\u001b[0m 160.00 KiB/374.75 KiB\r\n"}
71
+ ,{"stream_name":"stdout","time":14.796630563,"data":"\u001b[2mhf-transfer \u001b[0m \u001b[32m\u001b[30m\u001b[2m------------------------------\u001b[0m\u001b[0m 0 B/3.39 MiB\r\n"}
72
+ ,{"stream_name":"stdout","time":14.796634805,"data":"\u001b[2mbitsandbytes \u001b[0m \u001b[32m\u001b[30m\u001b[2m------------------------------\u001b[0m\u001b[0m 14.91 KiB/57.84 MiB \u001b[3A\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[3A\u001b[37m⠙\u001b[0m \u001b[2mPreparing packages...\u001b[0m (0/3)\r\n"}
73
+ ,{"stream_name":"stdout","time":14.796643965,"data":"\u001b[2maccelerate \u001b[0m \u001b[32m----------------\u001b[30m\u001b[2m--------------\u001b[0m\u001b[0m 208.00 KiB/374.75 KiB\r\n"}
74
+ ,{"stream_name":"stdout","time":14.796648815,"data":"\u001b[2mhf-transfer \u001b[0m \u001b[32m\u001b[30m\u001b[2m------------------------------\u001b[0m\u001b[0m 64.00 KiB/3.39 MiB\r\n"}
75
+ ,{"stream_name":"stdout","time":14.796653102,"data":"\u001b[2mbitsandbytes \u001b[0m \u001b[32m\u001b[30m\u001b[2m------------------------------\u001b[0m\u001b[0m 78.91 KiB/57.84 MiB \u001b[3A\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[3A\u001b[37m⠙\u001b[0m \u001b[2mPreparing packages...\u001b[0m (0/3)\r\n"}
76
+ ,{"stream_name":"stdout","time":14.796658188,"data":"\u001b[2mhf-transfer \u001b[0m \u001b[32m--\u001b[30m\u001b[2m----------------------------\u001b[0m\u001b[0m 336.00 KiB/3.39 MiB\r\n"}
77
+ ,{"stream_name":"stdout","time":14.846617166,"data":"\u001b[2mbitsandbytes \u001b[0m \u001b[32m\u001b[30m\u001b[2m------------------------------\u001b[0m\u001b[0m 392.56 KiB/57.84 MiB \u001b[2A\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[2A\u001b[37m⠙\u001b[0m \u001b[2mPreparing packages...\u001b[0m (0/3)\r\n"}
78
+ ,{"stream_name":"stdout","time":14.846656595,"data":"\u001b[2mhf-transfer \u001b[0m \u001b[32m-------------\u001b[30m\u001b[2m-----------------\u001b[0m\u001b[0m 1.49 MiB/3.39 MiB\r\n"}
79
+ ,{"stream_name":"stdout","time":14.897725836,"data":"\u001b[2mbitsandbytes \u001b[0m \u001b[32m\u001b[30m\u001b[2m------------------------------\u001b[0m\u001b[0m 1.92 MiB/57.84 MiB \u001b[2A\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[2A\u001b[37m⠙\u001b[0m \u001b[2mPreparing packages...\u001b[0m (0/3)\r\n"}
80
+ ,{"stream_name":"stdout","time":14.897762478,"data":"\u001b[2mhf-transfer \u001b[0m \u001b[32m------------------------------\u001b[30m\u001b[2m\u001b[0m\u001b[0m 3.39 MiB/3.39 MiB\r\n"}
81
+ ,{"stream_name":"stdout","time":14.897769426,"data":"\u001b[2mbitsandbytes \u001b[0m \u001b[32m--\u001b[30m\u001b[2m----------------------------\u001b[0m\u001b[0m 4.62 MiB/57.84 MiB \u001b[2A\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[2A\u001b[37m⠙\u001b[0m \u001b[2mPreparing packages...\u001b[0m (0/3)\r\n"}
82
+ ,{"stream_name":"stdout","time":14.948672247,"data":"\u001b[2mbitsandbytes \u001b[0m \u001b[32m--\u001b[30m\u001b[2m----------------------------\u001b[0m\u001b[0m 5.01 MiB/57.84 MiB \u001b[1A\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[1A\u001b[37m⠙\u001b[0m \u001b[2mPreparing packages...\u001b[0m (0/3)\r\n"}
83
+ ,{"stream_name":"stdout","time":15.000232332,"data":"\u001b[2mbitsandbytes \u001b[0m \u001b[32m----\u001b[30m\u001b[2m--------------------------\u001b[0m\u001b[0m 8.88 MiB/57.84 MiB \u001b[1A\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[1A\u001b[37m⠹\u001b[0m \u001b[2mPreparing packages...\u001b[0m (2/3)\r\n"}
84
+ ,{"stream_name":"stdout","time":15.05090943,"data":"\u001b[2mbitsandbytes \u001b[0m \u001b[32m-------\u001b[30m\u001b[2m-----------------------\u001b[0m\u001b[0m 13.76 MiB/57.84 MiB \u001b[1A\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[1A\u001b[37m⠹\u001b[0m \u001b[2mPreparing packages...\u001b[0m (2/3)\r\n"}
85
+ ,{"stream_name":"stdout","time":15.102215159,"data":"\u001b[2mbitsandbytes \u001b[0m \u001b[32m---------\u001b[30m\u001b[2m---------------------\u001b[0m\u001b[0m 18.24 MiB/57.84 MiB \u001b[1A\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[1A\u001b[37m⠹\u001b[0m \u001b[2mPreparing packages...\u001b[0m (2/3)\r\n"}
86
+ ,{"stream_name":"stdout","time":15.152723195,"data":"\u001b[2mbitsandbytes \u001b[0m \u001b[32m-----------\u001b[30m\u001b[2m-------------------\u001b[0m\u001b[0m 22.14 MiB/57.84 MiB \u001b[1A\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[1A\u001b[37m⠸\u001b[0m \u001b[2mPreparing packages...\u001b[0m (2/3)\r\n"}
87
+ ,{"stream_name":"stdout","time":15.203708716,"data":"\u001b[2mbitsandbytes \u001b[0m \u001b[32m-------------\u001b[30m\u001b[2m-----------------\u001b[0m\u001b[0m 25.90 MiB/57.84 MiB \u001b[1A\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[1A\u001b[37m⠸\u001b[0m \u001b[2mPreparing packages...\u001b[0m (2/3)\r\n"}
88
+ ,{"stream_name":"stdout","time":15.255148809,"data":"\u001b[2mbitsandbytes \u001b[0m \u001b[32m---------------\u001b[30m\u001b[2m---------------\u001b[0m\u001b[0m 30.63 MiB/57.84 MiB \u001b[1A\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[1A\u001b[37m⠸\u001b[0m \u001b[2mPreparing packages...\u001b[0m (2/3)\r\n"}
89
+ ,{"stream_name":"stdout","time":15.30559422,"data":"\u001b[2mbitsandbytes \u001b[0m \u001b[32m------------------\u001b[30m\u001b[2m------------\u001b[0m\u001b[0m 35.09 MiB/57.84 MiB \u001b[1A\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[1A\u001b[37m⠸\u001b[0m \u001b[2mPreparing packages...\u001b[0m (2/3)\r\n"}
90
+ ,{"stream_name":"stdout","time":15.357182334,"data":"\u001b[2mbitsandbytes \u001b[0m \u001b[32m--------------------\u001b[30m\u001b[2m----------\u001b[0m\u001b[0m 39.58 MiB/57.84 MiB \u001b[1A\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[1A\u001b[37m⠼\u001b[0m \u001b[2mPreparing packages...\u001b[0m (2/3)\r\n"}
91
+ ,{"stream_name":"stdout","time":15.408409431,"data":"\u001b[2mbitsandbytes \u001b[0m \u001b[32m----------------------\u001b[30m\u001b[2m--------\u001b[0m\u001b[0m 43.73 MiB/57.84 MiB \u001b[1A\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[1A\u001b[37m⠼\u001b[0m \u001b[2mPreparing packages...\u001b[0m (2/3)\r\n"}
92
+ ,{"stream_name":"stdout","time":15.459847163,"data":"\u001b[2mbitsandbytes \u001b[0m \u001b[32m------------------------\u001b[30m\u001b[2m------\u001b[0m\u001b[0m 47.89 MiB/57.84 MiB \u001b[1A\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[1A\u001b[37m⠼\u001b[0m \u001b[2mPreparing packages...\u001b[0m (2/3)\r\n"}
93
+ ,{"stream_name":"stdout","time":15.502070377,"data":"\u001b[2mbitsandbytes \u001b[0m \u001b[32m---------------------------\u001b[30m\u001b[2m---\u001b[0m\u001b[0m 52.76 MiB/57.84 MiB \u001b[1A\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[1A\u001b[37m⠼\u001b[0m \u001b[2mPreparing packages...\u001b[0m (2/3)\r\n"}
94
+ ,{"stream_name":"stdout","time":15.502123259,"data":"\u001b[2mbitsandbytes \u001b[0m \u001b[32m-----------------------------\u001b[30m\u001b[2m-\u001b[0m\u001b[0m 57.78 MiB/57.84 MiB \u001b[1A\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[1A\u001b[37m⠼\u001b[0m \u001b[2mPreparing packages...\u001b[0m (2/3) \r\u001b[2K\u001b[2mPrepared \u001b[1m3 packages\u001b[0m \u001b[2min 760ms\u001b[0m\u001b[0m\r\n"}
95
+ ,{"stream_name":"stdout","time":15.502191272,"data":"░░░░░░░░░░░░░░░░░░░░ [0/0] \u001b[2mInstalling wheels... \u001b[0m\r\u001b[2K░░░░░░░░░░░░░░░░░░░░ [0/3] \u001b[2mInstalling wheels... \u001b[0m\r\u001b[2K░░░░░░░░░░░░░░░░░░░░ [0/3] \u001b[2mhf-transfer==0.1.9 \u001b[0m\r\u001b[2K██████░░░░░░░░░░░░░░ [1/3] \u001b[2mhf-transfer==0.1.9 \u001b[0m\r\u001b[2K██████░░░░░░░░░░░░░░ [1/3] \u001b[2mbitsandbytes==0.49.2 \u001b[0m\r\u001b[2K█████████████░░░░░░░ [2/3] \u001b[2mbitsandbytes==0.49.2 \u001b[0m\r\u001b[2K█████████████░░░░░░░ [2/3] \u001b[2maccelerate==1.13.0 \u001b[0m\r\u001b[2K████████████████████ [3/3] \u001b[2maccelerate==1.13.0 \u001b[0m\r\u001b[2K\u001b[2mInstalled \u001b[1m3 packages\u001b[0m \u001b[2min 8ms\u001b[0m\u001b[0m\r\n"}
96
+ ,{"stream_name":"stdout","time":15.502207801,"data":" \u001b[32m+\u001b[39m \u001b[1maccelerate\u001b[0m\u001b[2m==1.13.0\u001b[0m\r\n"}
97
+ ,{"stream_name":"stdout","time":15.502213719,"data":" \u001b[32m+\u001b[39m \u001b[1mbitsandbytes\u001b[0m\u001b[2m==0.49.2\u001b[0m\r\n"}
98
+ ,{"stream_name":"stdout","time":15.502219661,"data":" \u001b[32m+\u001b[39m \u001b[1mhf-transfer\u001b[0m\u001b[2m==0.1.9\u001b[0m\r\n"}
99
+ ,{"stream_name":"stdout","time":15.762840652,"data":"\u001b[2mUsing Python 3.12.12 environment at: /usr\u001b[0m\r\n"}
100
+ ,{"stream_name":"stdout","time":15.915771622,"data":"\u001b[37m⠋\u001b[0m \u001b[2mResolving dependencies... \u001b[0m\r\u001b[2K\u001b[37m⠙\u001b[0m \u001b[2mResolving dependencies... \u001b[0m\r\u001b[2K\u001b[37m⠋\u001b[0m \u001b[2mResolving dependencies... \u001b[0m\r\u001b[2K\u001b[37m⠙\u001b[0m \u001b[2mResolving dependencies... \u001b[0m\r\u001b[2K\u001b[37m⠙\u001b[0m \u001b[2mnumpy==2.0.0 \u001b[0m\r\u001b[2K\u001b[37m⠙\u001b[0m \u001b[2m \u001b[0m\r\u001b[2K\u001b[2mResolved \u001b[1m1 package\u001b[0m \u001b[2min 145ms\u001b[0m\u001b[0m\r\n"}
101
+ ,{"stream_name":"stdout","time":15.915823579,"data":"\u001b[37m⠋\u001b[0m \u001b[2mPreparing packages...\u001b[0m (0/0) \r\u001b[2K\u001b[37m⠙\u001b[0m \u001b[2mPreparing packages...\u001b[0m (0/0) \r\u001b[2K\u001b[37m⠙\u001b[0m \u001b[2mPreparing packages...\u001b[0m (0/1) \r\u001b[2K\u001b[37m⠙\u001b[0m \u001b[2mPreparing packages...\u001b[0m (0/1)\r\n"}
102
+ ,{"stream_name":"stdout","time":15.966224581,"data":"\u001b[2mnumpy \u001b[0m \u001b[32m\u001b[30m\u001b[2m------------------------------\u001b[0m\u001b[0m 0 B/18.11 MiB \u001b[1A\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[1A\u001b[37m⠙\u001b[0m \u001b[2mPreparing packages...\u001b[0m (0/1)\r\n"}
103
+ ,{"stream_name":"stdout","time":15.966361951,"data":"\u001b[2mnumpy \u001b[0m \u001b[32m\u001b[30m\u001b[2m------------------------------\u001b[0m\u001b[0m 16.00 KiB/18.11 MiB \u001b[1A\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[1A\u001b[37m⠙\u001b[0m \u001b[2mPreparing packages...\u001b[0m (0/1)\r\n"}
104
+ ,{"stream_name":"stdout","time":15.966372515,"data":"\u001b[2mnumpy \u001b[0m \u001b[32m\u001b[30m\u001b[2m------------------------------\u001b[0m\u001b[0m 20.08 KiB/18.11 MiB \u001b[1A\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[1A\u001b[37m⠙\u001b[0m \u001b[2mPreparing packages...\u001b[0m (0/1)\r\n"}
105
+ ,{"stream_name":"stdout","time":15.966378565,"data":"\u001b[2mnumpy \u001b[0m \u001b[32m\u001b[30m\u001b[2m------------------------------\u001b[0m\u001b[0m 36.08 KiB/18.11 MiB \u001b[1A\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[1A\u001b[37m⠙\u001b[0m \u001b[2mPreparing packages...\u001b[0m (0/1)\r\n"}
106
+ ,{"stream_name":"stdout","time":15.966384009,"data":"\u001b[2mnumpy \u001b[0m \u001b[32m\u001b[30m\u001b[2m------------------------------\u001b[0m\u001b[0m 37.49 KiB/18.11 MiB \u001b[1A\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[1A\u001b[37m⠙\u001b[0m \u001b[2mPreparing packages...\u001b[0m (0/1)\r\n"}
107
+ ,{"stream_name":"stdout","time":15.966389511,"data":"\u001b[2mnumpy \u001b[0m \u001b[32m\u001b[30m\u001b[2m------------------------------\u001b[0m\u001b[0m 48.20 KiB/18.11 MiB \u001b[1A\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[1A\u001b[37m⠙\u001b[0m \u001b[2mPreparing packages...\u001b[0m (0/1)\r\n"}
108
+ ,{"stream_name":"stdout","time":15.966394784,"data":"\u001b[2mnumpy \u001b[0m \u001b[32m\u001b[30m\u001b[2m------------------------------\u001b[0m\u001b[0m 50.88 KiB/18.11 MiB \u001b[1A\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[1A\u001b[37m⠙\u001b[0m \u001b[2mPreparing packages...\u001b[0m (0/1)\r\n"}
109
+ ,{"stream_name":"stdout","time":15.966400029999999,"data":"\u001b[2mnumpy \u001b[0m \u001b[32m\u001b[30m\u001b[2m------------------------------\u001b[0m\u001b[0m 52.22 KiB/18.11 MiB \u001b[1A\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[1A\u001b[37m⠙\u001b[0m \u001b[2mPreparing packages...\u001b[0m (0/1)\r\n"}
110
+ ,{"stream_name":"stdout","time":15.966405298,"data":"\u001b[2mnumpy \u001b[0m \u001b[32m\u001b[30m\u001b[2m------------------------------\u001b[0m\u001b[0m 54.89 KiB/18.11 MiB \u001b[1A\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[1A\u001b[37m⠙\u001b[0m \u001b[2mPreparing packages...\u001b[0m (0/1)\r\n"}
111
+ ,{"stream_name":"stdout","time":15.966419539,"data":"\u001b[2mnumpy \u001b[0m \u001b[32m\u001b[30m\u001b[2m------------------------------\u001b[0m\u001b[0m 57.57 KiB/18.11 MiB \u001b[1A\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[1A\u001b[37m⠙\u001b[0m \u001b[2mPreparing packages...\u001b[0m (0/1)\r\n"}
112
+ ,{"stream_name":"stdout","time":15.966424798,"data":"\u001b[2mnumpy \u001b[0m \u001b[32m\u001b[30m\u001b[2m------------------------------\u001b[0m\u001b[0m 58.91 KiB/18.11 MiB \u001b[1A\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[1A\u001b[37m⠙\u001b[0m \u001b[2mPreparing packages...\u001b[0m (0/1)\r\n"}
113
+ ,{"stream_name":"stdout","time":15.966432059,"data":"\u001b[2mnumpy \u001b[0m \u001b[32m\u001b[30m\u001b[2m------------------------------\u001b[0m\u001b[0m 60.25 KiB/18.11 MiB \u001b[1A\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[1A\u001b[37m⠙\u001b[0m \u001b[2mPreparing packages...\u001b[0m (0/1)\r\n"}
114
+ ,{"stream_name":"stdout","time":15.966437489,"data":"\u001b[2mnumpy \u001b[0m \u001b[32m\u001b[30m\u001b[2m------------------------------\u001b[0m\u001b[0m 60.81 KiB/18.11 MiB \u001b[1A\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[1A\u001b[37m⠙\u001b[0m \u001b[2mPreparing packages...\u001b[0m (0/1)\r\n"}
115
+ ,{"stream_name":"stdout","time":15.966442657,"data":"\u001b[2mnumpy \u001b[0m \u001b[32m\u001b[30m\u001b[2m------------------------------\u001b[0m\u001b[0m 76.81 KiB/18.11 MiB \u001b[1A\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[1A\u001b[37m⠙\u001b[0m \u001b[2mPreparing packages...\u001b[0m (0/1)\r\n"}
116
+ ,{"stream_name":"stdout","time":15.966447919,"data":"\u001b[2mnumpy \u001b[0m \u001b[32m\u001b[30m\u001b[2m------------------------------\u001b[0m\u001b[0m 92.81 KiB/18.11 MiB \u001b[1A\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[1A\u001b[37m⠙\u001b[0m \u001b[2mPreparing packages...\u001b[0m (0/1)\r\n"}
117
+ ,{"stream_name":"stdout","time":15.966453434,"data":"\u001b[2mnumpy \u001b[0m \u001b[32m\u001b[30m\u001b[2m------------------------------\u001b[0m\u001b[0m 108.81 KiB/18.11 MiB \u001b[1A\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[1A\u001b[37m⠙\u001b[0m \u001b[2mPreparing packages...\u001b[0m (0/1)\r\n"}
118
+ ,{"stream_name":"stdout","time":15.96645867,"data":"\u001b[2mnumpy \u001b[0m \u001b[32m\u001b[30m\u001b[2m------------------------------\u001b[0m\u001b[0m 124.81 KiB/18.11 MiB \u001b[1A\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[1A\u001b[37m⠙\u001b[0m \u001b[2mPreparing packages...\u001b[0m (0/1)\r\n"}
119
+ ,{"stream_name":"stdout","time":16.017797587,"data":"\u001b[2mnumpy \u001b[0m \u001b[32m\u001b[30m\u001b[2m------------------------------\u001b[0m\u001b[0m 444.81 KiB/18.11 MiB \u001b[1A\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[1A\u001b[37m⠙\u001b[0m \u001b[2mPreparing packages...\u001b[0m (0/1)\r\n"}
120
+ ,{"stream_name":"stdout","time":16.068515576,"data":"\u001b[2mnumpy \u001b[0m \u001b[32m---\u001b[30m\u001b[2m---------------------------\u001b[0m\u001b[0m 1.87 MiB/18.11 MiB \u001b[1A\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[1A\u001b[37m⠙\u001b[0m \u001b[2mPreparing packages...\u001b[0m (0/1)\r\n"}
121
+ ,{"stream_name":"stdout","time":16.119875534,"data":"\u001b[2mnumpy \u001b[0m \u001b[32m------\u001b[30m\u001b[2m------------------------\u001b[0m\u001b[0m 4.01 MiB/18.11 MiB \u001b[1A\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[1A\u001b[37m⠙\u001b[0m \u001b[2mPreparing packages...\u001b[0m (0/1)\r\n"}
122
+ ,{"stream_name":"stdout","time":16.171031808,"data":"\u001b[2mnumpy \u001b[0m \u001b[32m---------\u001b[30m\u001b[2m---------------------\u001b[0m\u001b[0m 5.87 MiB/18.11 MiB \u001b[1A\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[1A\u001b[37m⠹\u001b[0m \u001b[2mPreparing packages...\u001b[0m (0/1)\r\n"}
123
+ ,{"stream_name":"stdout","time":16.221317616,"data":"\u001b[2mnumpy \u001b[0m \u001b[32m-----------\u001b[30m\u001b[2m-------------------\u001b[0m\u001b[0m 7.08 MiB/18.11 MiB \u001b[1A\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[1A\u001b[37m⠹\u001b[0m \u001b[2mPreparing packages...\u001b[0m (0/1)\r\n"}
124
+ ,{"stream_name":"stdout","time":16.27319406,"data":"\u001b[2mnumpy \u001b[0m \u001b[32m---------------\u001b[30m\u001b[2m---------------\u001b[0m\u001b[0m 9.18 MiB/18.11 MiB \u001b[1A\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[1A\u001b[37m⠹\u001b[0m \u001b[2mPreparing packages...\u001b[0m (0/1)\r\n"}
125
+ ,{"stream_name":"stdout","time":16.323819272,"data":"\u001b[2mnumpy \u001b[0m \u001b[32m---------------------\u001b[30m\u001b[2m---------\u001b[0m\u001b[0m 12.81 MiB/18.11 MiB \u001b[1A\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[1A\u001b[37m⠸\u001b[0m \u001b[2mPreparing packages...\u001b[0m (0/1)\r\n"}
126
+ ,{"stream_name":"stdout","time":16.374572931,"data":"\u001b[2mnumpy \u001b[0m \u001b[32m----------------------------\u001b[30m\u001b[2m--\u001b[0m\u001b[0m 16.92 MiB/18.11 MiB \u001b[1A\r\u001b[2K\u001b[1B\r\u001b[2K\u001b[1A\u001b[37m⠸\u001b[0m \u001b[2mPreparing packages...\u001b[0m (0/1) \r\u001b[2K\u001b[2mPrepared \u001b[1m1 package\u001b[0m \u001b[2min 424ms\u001b[0m\u001b[0m\r\n"}
127
+ ,{"stream_name":"stdout","time":16.425564129,"data":"\u001b[2mUninstalled \u001b[1m1 package\u001b[0m \u001b[2min 95ms\u001b[0m\u001b[0m\r\n"}
128
+ ,{"stream_name":"stdout","time":16.458088441,"data":"░░░░░░░░░░░░░░░░░░░░ [0/0] \u001b[2mInstalling wheels... \u001b[0m\r\u001b[2K░░░░░░░░░░░░░░░░░░░░ [0/1] \u001b[2mInstalling wheels... \u001b[0m\r\u001b[2K░░░░░░░░░░░░░░░░░░░░ [0/1] \u001b[2mnumpy==2.0.0 \u001b[0m\r\u001b[2K████████████████████ [1/1] \u001b[2mnumpy==2.0.0 \u001b[0m\r\u001b[2K\u001b[2mInstalled \u001b[1m1 package\u001b[0m \u001b[2min 23ms\u001b[0m\u001b[0m\r\n"}
129
+ ,{"stream_name":"stdout","time":16.458262602,"data":" \u001b[31m-\u001b[39m \u001b[1mnumpy\u001b[0m\u001b[2m==2.0.2\u001b[0m\r\n"}
130
+ ,{"stream_name":"stdout","time":16.458275089,"data":" \u001b[32m+\u001b[39m \u001b[1mnumpy\u001b[0m\u001b[2m==2.0.0\u001b[0m\r\n"}
131
+ ,{"stream_name":"stderr","time":16.592229747,"data":"2026-04-13 17:57:28,260 - INFO - ✅ Environment detected: Kaggle\n"}
132
+ ,{"stream_name":"stderr","time":16.592276218,"data":"\n"}
133
+ ,{"stream_name":"stderr","time":16.592282049,"data":"2026-04-13 17:57:28,260 - INFO - ✅ Environment detected: Kaggle\n"}
134
+ ,{"stream_name":"stderr","time":16.592286348000002,"data":"2026-04-13 17:57:28,262 - DEBUG - Created/verified output directory: /kaggle/working/\n"}
135
+ ,{"stream_name":"stderr","time":16.592290444,"data":"\n"}
136
+ ,{"stream_name":"stderr","time":16.592294454,"data":"2026-04-13 17:57:28,262 - DEBUG - Created/verified output directory: /kaggle/working/\n"}
137
+ ,{"stream_name":"stderr","time":16.593417424,"data":"2026-04-13 17:57:28,262 - INFO - 📂 Data Path: /kaggle/input/\n"}
138
+ ,{"stream_name":"stderr","time":16.59344082,"data":"\n"}
139
+ ,{"stream_name":"stderr","time":16.593445553,"data":"2026-04-13 17:57:28,262 - INFO - 📂 Data Path: /kaggle/input/\n"}
140
+ ,{"stream_name":"stderr","time":16.59437892,"data":"2026-04-13 17:57:28,263 - INFO - 📦 Output Path: /kaggle/working/\n"}
141
+ ,{"stream_name":"stderr","time":16.59440825,"data":"\n"}
142
+ ,{"stream_name":"stderr","time":16.594413326,"data":"2026-04-13 17:57:28,263 - INFO - 📦 Output Path: /kaggle/working/\n"}
143
+ ,{"stream_name":"stderr","time":16.595322766,"data":"2026-04-13 17:57:28,264 - INFO - \n"}
144
+ ,{"stream_name":"stderr","time":16.595351986,"data":"🔧 System Information\n"}
145
+ ,{"stream_name":"stderr","time":16.595356947,"data":"\n"}
146
+ ,{"stream_name":"stderr","time":16.595360737,"data":"2026-04-13 17:57:28,264 - INFO - \n"}
147
+ ,{"stream_name":"stderr","time":16.595364793999998,"data":"🔧 System Information\n"}
148
+ ,{"stream_name":"stderr","time":16.596292477,"data":"2026-04-13 17:57:28,265 - INFO - Python version: 3.12.12\n"}
149
+ ,{"stream_name":"stderr","time":16.596335602,"data":"\n"}
150
+ ,{"stream_name":"stderr","time":16.596341166,"data":"2026-04-13 17:57:28,265 - INFO - Python version: 3.12.12\n"}
151
+ ,{"stream_name":"stderr","time":21.477821506,"data":"2026-04-13 17:57:32,900 - INFO - PyTorch version: 2.10.0+cu128\n"}
152
+ ,{"stream_name":"stderr","time":21.477865996,"data":"\n"}
153
+ ,{"stream_name":"stderr","time":21.477871789,"data":"2026-04-13 17:57:32,900 - INFO - PyTorch version: 2.10.0+cu128\n"}
154
+ ,{"stream_name":"stderr","time":21.478899895,"data":"2026-04-13 17:57:33,147 - INFO - CUDA version: 12.8\n"}
155
+ ,{"stream_name":"stderr","time":21.47891908,"data":"\n"}
156
+ ,{"stream_name":"stderr","time":21.478923793,"data":"2026-04-13 17:57:33,147 - INFO - CUDA version: 12.8\n"}
157
+ ,{"stream_name":"stderr","time":21.514320166,"data":"2026-04-13 17:57:33,176 - INFO - GPU count: 2\n"}
158
+ ,{"stream_name":"stderr","time":21.514386476,"data":"\n"}
159
+ ,{"stream_name":"stderr","time":21.514392634,"data":"2026-04-13 17:57:33,176 - INFO - GPU count: 2\n"}
160
+ ,{"stream_name":"stderr","time":21.521550771,"data":"2026-04-13 17:57:33,191 - INFO - GPU 0: Tesla T4\n"}
161
+ ,{"stream_name":"stderr","time":21.521576685,"data":"\n"}
162
+ ,{"stream_name":"stderr","time":21.52158218,"data":"2026-04-13 17:57:33,191 - INFO - GPU 0: Tesla T4\n"}
163
+ ,{"stream_name":"stderr","time":21.522522321,"data":"2026-04-13 17:57:33,192 - INFO - GPU 1: Tesla T4\n"}
164
+ ,{"stream_name":"stderr","time":21.522540595,"data":"\n"}
165
+ ,{"stream_name":"stderr","time":21.522545401,"data":"2026-04-13 17:57:33,192 - INFO - GPU 1: Tesla T4\n"}
166
+ ,{"stream_name":"stderr","time":21.868764658,"data":"2026-04-13 17:57:33,537 - INFO - nvidia-smi output:\n"}
167
+ ,{"stream_name":"stderr","time":21.868808744,"data":"Mon Apr 13 17:57:33 2026 \n"}
168
+ ,{"stream_name":"stderr","time":21.868814653,"data":"+-----------------------------------------------------------------------------------------+\n"}
169
+ ,{"stream_name":"stderr","time":21.868819217,"data":"| NVIDIA-SMI 580.105.08 Driver Version: 580.105.08 CUDA Version: 13.0 |\n"}
170
+ ,{"stream_name":"stderr","time":21.868822789,"data":"+-----------------------------------------+------------------------+----------------------+\n"}
171
+ ,{"stream_name":"stderr","time":21.868825984,"data":"| GPU Name Persistence-M | Bus-Id Disp.A | Volatile Uncorr. ECC |\n"}
172
+ ,{"stream_name":"stderr","time":21.86882915,"data":"| Fan Temp Perf Pwr:Usage/Cap | Memory-Usage | GPU-Util Compute M. |\n"}
173
+ ,{"stream_name":"stderr","time":21.868832332,"data":"| | | MIG M. |\n"}
174
+ ,{"stream_name":"stderr","time":21.868835444,"data":"|=========================================+========================+======================|\n"}
175
+ ,{"stream_name":"stderr","time":21.868839266,"data":"| 0 Tesla T4 Off | 00000000:00:04.0 Off | 0 |\n"}
176
+ ,{"stream_name":"stderr","time":21.868842704,"data":"| N/A 45C P8 12W / 70W | 3MiB / 15360MiB | 0% Default |\n"}
177
+ ,{"stream_name":"stderr","time":21.868846502,"data":"| | | N/A |\n"}
178
+ ,{"stream_name":"stderr","time":21.868862289,"data":"+-----------------------------------------+------------------------+----------------------+\n"}
179
+ ,{"stream_name":"stderr","time":21.86886581,"data":"| 1 Tesla T4 Off | 00000000:00:05.0 Off | 0 |\n"}
180
+ ,{"stream_name":"stderr","time":21.868869224,"data":"| N/A 35C P8 10W / 70W | 3MiB / 15360MiB | 0% Default |\n"}
181
+ ,{"stream_name":"stderr","time":21.868872711,"data":"| | | N/A |\n"}
182
+ ,{"stream_name":"stderr","time":21.868876127,"data":"+-----------------------------------------+------------------------+----------------------+\n"}
183
+ ,{"stream_name":"stderr","time":21.86887945,"data":"\n"}
184
+ ,{"stream_name":"stderr","time":21.868882861,"data":"+-----------------------------------------------------------------------------------------+\n"}
185
+ ,{"stream_name":"stderr","time":21.86888645,"data":"| Processes: |\n"}
186
+ ,{"stream_name":"stderr","time":21.868890322,"data":"| GPU GI CI PID Type Process name GPU Memory |\n"}
187
+ ,{"stream_name":"stderr","time":21.868893985,"data":"| ID ID Usage |\n"}
188
+ ,{"stream_name":"stderr","time":21.868897549,"data":"|=========================================================================================|\n"}
189
+ ,{"stream_name":"stderr","time":21.868901089,"data":"| No running processes found |\n"}
190
+ ,{"stream_name":"stderr","time":21.868906463,"data":"+-----------------------------------------------------------------------------------------+\n"}
191
+ ,{"stream_name":"stderr","time":21.868909928,"data":"\n"}
192
+ ,{"stream_name":"stderr","time":21.868913698,"data":"2026-04-13 17:57:33,537 - INFO - nvidia-smi output:\n"}
193
+ ,{"stream_name":"stderr","time":21.868917285,"data":"Mon Apr 13 17:57:33 2026 \n"}
194
+ ,{"stream_name":"stderr","time":21.868920889,"data":"+-----------------------------------------------------------------------------------------+\n"}
195
+ ,{"stream_name":"stderr","time":21.86892455,"data":"| NVIDIA-SMI 580.105.08 Driver Version: 580.105.08 CUDA Version: 13.0 |\n"}
196
+ ,{"stream_name":"stderr","time":21.868928254,"data":"+-----------------------------------------+------------------------+----------------------+\n"}
197
+ ,{"stream_name":"stderr","time":21.868931939,"data":"| GPU Name Persistence-M | Bus-Id Disp.A | Volatile Uncorr. ECC |\n"}
198
+ ,{"stream_name":"stderr","time":21.868935667,"data":"| Fan Temp Perf Pwr:Usage/Cap | Memory-Usage | GPU-Util Compute M. |\n"}
199
+ ,{"stream_name":"stderr","time":21.868939337,"data":"| | | MIG M. |\n"}
200
+ ,{"stream_name":"stderr","time":21.86894312,"data":"|=========================================+========================+======================|\n"}
201
+ ,{"stream_name":"stderr","time":21.868946882,"data":"| 0 Tesla T4 Off | 00000000:00:04.0 Off | 0 |\n"}
202
+ ,{"stream_name":"stderr","time":21.86895045,"data":"| N/A 45C P8 12W / 70W | 3MiB / 15360MiB | 0% Default |\n"}
203
+ ,{"stream_name":"stderr","time":21.868954037,"data":"| | | N/A |\n"}
204
+ ,{"stream_name":"stderr","time":21.868957617,"data":"+-----------------------------------------+------------------------+----------------------+\n"}
205
+ ,{"stream_name":"stderr","time":21.86896529,"data":"| 1 Tesla T4 Off | 00000000:00:05.0 Off | 0 |\n"}
206
+ ,{"stream_name":"stderr","time":21.868969136,"data":"| N/A 35C P8 10W / 70W | 3MiB / 15360MiB | 0% Default |\n"}
207
+ ,{"stream_name":"stderr","time":21.868972891,"data":"| | | N/A |\n"}
208
+ ,{"stream_name":"stderr","time":21.868976534,"data":"+-----------------------------------------+------------------------+----------------------+\n"}
209
+ ,{"stream_name":"stderr","time":21.868980252,"data":"\n"}
210
+ ,{"stream_name":"stderr","time":21.868983595,"data":"+-----------------------------------------------------------------------------------------+\n"}
211
+ ,{"stream_name":"stderr","time":21.868987303,"data":"| Processes: |\n"}
212
+ ,{"stream_name":"stderr","time":21.868991184,"data":"| GPU GI CI PID Type Process name GPU Memory |\n"}
213
+ ,{"stream_name":"stderr","time":21.86899514,"data":"| ID ID Usage |\n"}
214
+ ,{"stream_name":"stderr","time":21.86899882,"data":"|=========================================================================================|\n"}
215
+ ,{"stream_name":"stderr","time":21.8690029,"data":"| No running processes found |\n"}
216
+ ,{"stream_name":"stderr","time":21.869006678,"data":"+-----------------------------------------------------------------------------------------+\n"}
217
+ ,{"stream_name":"stderr","time":21.869966751,"data":"2026-04-13 17:57:33,539 - INFO - Attempting to load secret 'WANDB_API_KEY' from the 'kaggle' environment...\n"}
218
+ ,{"stream_name":"stderr","time":21.870003761,"data":"\n"}
219
+ ,{"stream_name":"stderr","time":21.870009043,"data":"2026-04-13 17:57:33,539 - INFO - Attempting to load secret 'WANDB_API_KEY' from the 'kaggle' environment...\n"}
220
+ ,{"stream_name":"stderr","time":21.96523335,"data":"2026-04-13 17:57:33,634 - INFO - ✅ Successfully loaded secret 'WANDB_API_KEY'.\n"}
221
+ ,{"stream_name":"stderr","time":21.965265613,"data":"\n"}
222
+ ,{"stream_name":"stderr","time":21.965270989,"data":"2026-04-13 17:57:33,634 - INFO - ✅ Successfully loaded secret 'WANDB_API_KEY'.\n"}
223
+ ,{"stream_name":"stderr","time":21.966399236,"data":"2026-04-13 17:57:33,635 - INFO - Attempting to load secret 'HF_TOKEN' from the 'kaggle' environment...\n"}
224
+ ,{"stream_name":"stderr","time":21.966431686,"data":"\n"}
225
+ ,{"stream_name":"stderr","time":21.966437831,"data":"2026-04-13 17:57:33,635 - INFO - Attempting to load secret 'HF_TOKEN' from the 'kaggle' environment...\n"}
226
+ ,{"stream_name":"stderr","time":22.038620533,"data":"2026-04-13 17:57:33,708 - INFO - ✅ Successfully loaded secret 'HF_TOKEN'.\n"}
227
+ ,{"stream_name":"stderr","time":22.03864997,"data":"\n"}
228
+ ,{"stream_name":"stderr","time":22.038655755,"data":"2026-04-13 17:57:33,708 - INFO - ✅ Successfully loaded secret 'HF_TOKEN'.\n"}
229
+ ,{"stream_name":"stderr","time":22.039928454,"data":"2026-04-13 17:57:33,709 - INFO - Attempting to load secret 'GITHUB_TOKEN' from the 'kaggle' environment...\n"}
230
+ ,{"stream_name":"stderr","time":22.039953904,"data":"\n"}
231
+ ,{"stream_name":"stderr","time":22.039958612,"data":"2026-04-13 17:57:33,709 - INFO - Attempting to load secret 'GITHUB_TOKEN' from the 'kaggle' environment...\n"}
232
+ ,{"stream_name":"stderr","time":22.107468348,"data":"2026-04-13 17:57:33,776 - INFO - ✅ Successfully loaded secret 'GITHUB_TOKEN'.\n"}
233
+ ,{"stream_name":"stderr","time":22.10754115,"data":"\n"}
234
+ ,{"stream_name":"stderr","time":22.107548407,"data":"2026-04-13 17:57:33,776 - INFO - ✅ Successfully loaded secret 'GITHUB_TOKEN'.\n"}
235
+ ,{"stream_name":"stderr","time":23.914041327,"data":"\u001b[34m\u001b[1mwandb\u001b[0m: [wandb.login()] Loaded credentials for https://api.wandb.ai from WANDB_API_KEY.\n"}
236
+ ,{"stream_name":"stderr","time":23.914091046,"data":"\n"}
237
+ ,{"stream_name":"stderr","time":23.91409753,"data":"\u001b[34m\u001b[1mwandb\u001b[0m: [wandb.login()] Loaded credentials for https://api.wandb.ai from WANDB_API_KEY.\n"}
238
+ ,{"stream_name":"stderr","time":24.040656149,"data":"\u001b[34m\u001b[1mwandb\u001b[0m: Currently logged in as: \u001b[33mdungngocpham171\u001b[0m (\u001b[33mdungngocpham171-university-of-science\u001b[0m) to \u001b[32mhttps://api.wandb.ai\u001b[0m. Use \u001b[1m`wandb login --relogin`\u001b[0m to force relogin\n"}
239
+ ,{"stream_name":"stderr","time":24.040686868,"data":"\n"}
240
+ ,{"stream_name":"stderr","time":24.040691629,"data":"\u001b[34m\u001b[1mwandb\u001b[0m: Currently logged in as: \u001b[33mdungngocpham171\u001b[0m (\u001b[33mdungngocpham171-university-of-science\u001b[0m) to \u001b[32mhttps://api.wandb.ai\u001b[0m. Use \u001b[1m`wandb login --relogin`\u001b[0m to force relogin\n"}
241
+ ,{"stream_name":"stderr","time":24.660173215,"data":"Note: Environment variable`HF_TOKEN` is set and is the current active token independently from the token you've just configured.\n"}
242
+ ,{"stream_name":"stderr","time":24.660211173,"data":"\n"}
243
+ ,{"stream_name":"stderr","time":24.660217118,"data":"Note: Environment variable`HF_TOKEN` is set and is the current active token independently from the token you've just configured.\n"}
244
+ ,{"stream_name":"stdout","time":24.901499082,"data":"📥 Downloading checkpoint from Hugging Face Hub to 'taskA-unixcoder-focal'...\n"}
245
+ ,{"stream_name":"stdout","time":24.901528656,"data":"\n"}
246
+ ,{"stream_name":"stdout","time":34.524673503,"data":"\n"}
247
+ ,{"stream_name":"stdout","time":34.524719663,"data":"✅ Download complete!\n"}
248
+ ,{"stream_name":"stdout","time":34.524728012,"data":"\n"}
249
+ ,{"stream_name":"stdout","time":34.524732334,"data":"📂 Files in taskA-unixcoder-focal/:\n"}
250
+ ,{"stream_name":"stderr","time":51.238458078,"data":"2026-04-13 17:58:02.073726: E external/local_xla/xla/stream_executor/cuda/cuda_fft.cc:467] Unable to register cuFFT factory: Attempting to register factory for plugin cuFFT when one has already been registered\n"}
251
+ ,{"stream_name":"stderr","time":51.238507799,"data":"WARNING: All log messages before absl::InitializeLog() is called are written to STDERR\n"}
252
+ ,{"stream_name":"stderr","time":51.238515871,"data":"E0000 00:00:1776103082.289616 23 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered\n"}
253
+ ,{"stream_name":"stderr","time":51.23852112,"data":"E0000 00:00:1776103082.343697 23 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered\n"}
254
+ ,{"stream_name":"stderr","time":51.238526872,"data":"W0000 00:00:1776103082.848433 23 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once.\n"}
255
+ ,{"stream_name":"stderr","time":51.238531394,"data":"W0000 00:00:1776103082.848476 23 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once.\n"}
256
+ ,{"stream_name":"stderr","time":51.238536423,"data":"W0000 00:00:1776103082.848479 23 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once.\n"}
257
+ ,{"stream_name":"stderr","time":51.239044537,"data":"W0000 00:00:1776103082.848482 23 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once.\n"}
258
+ ,{"stream_name":"stderr","time":51.441415848,"data":"2026-04-13 17:58:02.073726: E external/local_xla/xla/stream_executor/cuda/cuda_fft.cc:467] Unable to register cuFFT factory: Attempting to register factory for plugin cuFFT when one has already been registered\n"}
259
+ ,{"stream_name":"stderr","time":51.441450452,"data":"WARNING: All log messages before absl::InitializeLog() is called are written to STDERR\n"}
260
+ ,{"stream_name":"stderr","time":51.441456374,"data":"E0000 00:00:1776103082.289616 23 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered\n"}
261
+ ,{"stream_name":"stderr","time":51.441461099,"data":"E0000 00:00:1776103082.343697 23 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered\n"}
262
+ ,{"stream_name":"stderr","time":51.441465757,"data":"W0000 00:00:1776103082.848433 23 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once.\n"}
263
+ ,{"stream_name":"stderr","time":51.441470186,"data":"W0000 00:00:1776103082.848476 23 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once.\n"}
264
+ ,{"stream_name":"stderr","time":51.441474575,"data":"W0000 00:00:1776103082.848479 23 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once.\n"}
265
+ ,{"stream_name":"stderr","time":51.441478871,"data":"W0000 00:00:1776103082.848482 23 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once.\n"}
266
+ ,{"stream_name":"stderr","time":51.441483601,"data":"\n"}
267
+ ,{"stream_name":"stderr","time":51.441487313,"data":"2026-04-13 17:58:02.073726: E external/local_xla/xla/stream_executor/cuda/cuda_fft.cc:467] Unable to register cuFFT factory: Attempting to register factory for plugin cuFFT when one has already been registered\n"}
268
+ ,{"stream_name":"stderr","time":51.441491795,"data":"WARNING: All log messages before absl::InitializeLog() is called are written to STDERR\n"}
269
+ ,{"stream_name":"stderr","time":51.441495677,"data":"E0000 00:00:1776103082.289616 23 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered\n"}
270
+ ,{"stream_name":"stderr","time":51.441500294,"data":"E0000 00:00:1776103082.343697 23 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered\n"}
271
+ ,{"stream_name":"stderr","time":51.441505609000004,"data":"W0000 00:00:1776103082.848433 23 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once.\n"}
272
+ ,{"stream_name":"stderr","time":51.441509989,"data":"W0000 00:00:1776103082.848476 23 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once.\n"}
273
+ ,{"stream_name":"stderr","time":51.441514263,"data":"W0000 00:00:1776103082.848479 23 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once.\n"}
274
+ ,{"stream_name":"stderr","time":51.44151867,"data":"W0000 00:00:1776103082.848482 23 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once.\n"}
275
+ ,{"stream_name":"stdout","time":74.444641201,"data":"Torch version: 2.10.0+cu128\n"}
276
+ ,{"stream_name":"stdout","time":74.444680023,"data":"CUDA available: True\n"}
277
+ ,{"stream_name":"stderr","time":74.628803559,"data":"2026-04-13 17:58:26,297 - INFO - ✅ Dataset balancing monkey-patch applied\n"}
278
+ ,{"stream_name":"stderr","time":74.628867127,"data":"\n"}
279
+ ,{"stream_name":"stderr","time":74.628873377,"data":"2026-04-13 17:58:26,297 - INFO - ✅ Dataset balancing monkey-patch applied\n"}
280
+ ,{"stream_name":"stderr","time":74.629919953,"data":"INFO:__main__:✅ Dataset balancing monkey-patch applied\n"}
281
+ ,{"stream_name":"stderr","time":74.629963528,"data":"\n"}
282
+ ,{"stream_name":"stderr","time":74.629969508,"data":"INFO:__main__:✅ Dataset balancing monkey-patch applied\n"}
283
+ ,{"stream_name":"stderr","time":74.653908693,"data":"2026-04-13 17:58:26,322 - INFO - Loading data...\n"}
284
+ ,{"stream_name":"stderr","time":74.653944829,"data":"\n"}
285
+ ,{"stream_name":"stderr","time":74.653950697,"data":"2026-04-13 17:58:26,322 - INFO - Loading data...\n"}
286
+ ,{"stream_name":"stderr","time":74.655119498,"data":"INFO:__main__:Loading data...\n"}
287
+ ,{"stream_name":"stderr","time":74.655157819,"data":"\n"}
288
+ ,{"stream_name":"stderr","time":74.655163341,"data":"INFO:__main__:Loading data...\n"}
289
+ ,{"stream_name":"stderr","time":74.656107226,"data":"2026-04-13 17:58:26,324 - INFO - Loading datasets from Hugging Face Hub...\n"}
290
+ ,{"stream_name":"stderr","time":74.656121843,"data":"\n"}
291
+ ,{"stream_name":"stderr","time":74.65614068,"data":"2026-04-13 17:58:26,324 - INFO - Loading datasets from Hugging Face Hub...\n"}
292
+ ,{"stream_name":"stderr","time":74.65758463,"data":"INFO:__main__:Loading datasets from Hugging Face Hub...\n"}
293
+ ,{"stream_name":"stderr","time":74.657630758,"data":"\n"}
294
+ ,{"stream_name":"stderr","time":74.657637297,"data":"INFO:__main__:Loading datasets from Hugging Face Hub...\n"}
295
+ ,{"stream_name":"stderr","time":89.025503373,"data":"2026-04-13 17:58:40,694 - INFO - Train: 500000 | Val: 100000 | Labels: 2\n"}
296
+ ,{"stream_name":"stderr","time":89.025543944,"data":"\n"}
297
+ ,{"stream_name":"stderr","time":89.025549803,"data":"2026-04-13 17:58:40,694 - INFO - Train: 500000 | Val: 100000 | Labels: 2\n"}
298
+ ,{"stream_name":"stderr","time":89.026560121,"data":"INFO:__main__:Train: 500000 | Val: 100000 | Labels: 2\n"}
299
+ ,{"stream_name":"stderr","time":89.026588122,"data":"\n"}
300
+ ,{"stream_name":"stderr","time":89.02664,"data":"INFO:__main__:Train: 500000 | Val: 100000 | Labels: 2\n"}
301
+ ,{"stream_name":"stderr","time":89.027618215,"data":"2026-04-13 17:58:40,696 - INFO - Initialising model...\n"}
302
+ ,{"stream_name":"stderr","time":89.027645621,"data":"\n"}
303
+ ,{"stream_name":"stderr","time":89.027652672,"data":"2026-04-13 17:58:40,696 - INFO - Initialising model...\n"}
304
+ ,{"stream_name":"stderr","time":89.028506098,"data":"INFO:__main__:Initialising model...\n"}
305
+ ,{"stream_name":"stderr","time":89.028531453,"data":"\n"}
306
+ ,{"stream_name":"stderr","time":89.028537471,"data":"INFO:__main__:Initialising model...\n"}
307
+ ,{"stream_name":"stderr","time":89.029429236,"data":"2026-04-13 17:58:40,697 - INFO - Initialising backbone: microsoft/codebert-base\n"}
308
+ ,{"stream_name":"stderr","time":89.029453957,"data":"\n"}
309
+ ,{"stream_name":"stderr","time":89.029459701,"data":"2026-04-13 17:58:40,697 - INFO - Initialising backbone: microsoft/codebert-base\n"}
310
+ ,{"stream_name":"stderr","time":89.030438103,"data":"INFO:__main__:Initialising backbone: microsoft/codebert-base\n"}
311
+ ,{"stream_name":"stderr","time":89.030458786,"data":"\n"}
312
+ ,{"stream_name":"stderr","time":89.030465459,"data":"INFO:__main__:Initialising backbone: microsoft/codebert-base\n"}
313
+ ,{"stream_name":"stderr","time":95.047087462,"data":"Some weights of RobertaForSequenceClassification were not initialized from the model checkpoint at microsoft/codebert-base and are newly initialized: ['classifier.dense.bias', 'classifier.dense.weight', 'classifier.out_proj.bias', 'classifier.out_proj.weight']\n"}
314
+ ,{"stream_name":"stderr","time":95.047141494,"data":"You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.\n"}
315
+ ,{"stream_name":"stderr","time":95.047148154,"data":"\n"}
316
+ ,{"stream_name":"stderr","time":95.047153112,"data":"Some weights of RobertaForSequenceClassification were not initialized from the model checkpoint at microsoft/codebert-base and are newly initialized: ['classifier.dense.bias', 'classifier.dense.weight', 'classifier.out_proj.bias', 'classifier.out_proj.weight']\n"}
317
+ ,{"stream_name":"stderr","time":95.047157654,"data":"You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.\n"}
318
+ ,{"stream_name":"stderr","time":95.048277549,"data":"2026-04-13 17:58:46,717 - INFO - Replacing classifier with CodeBertStylometryClassifier\n"}
319
+ ,{"stream_name":"stderr","time":95.04830557,"data":"\n"}
320
+ ,{"stream_name":"stderr","time":95.048311241,"data":"2026-04-13 17:58:46,717 - INFO - Replacing classifier with CodeBertStylometryClassifier\n"}
321
+ ,{"stream_name":"stderr","time":95.049301703,"data":"INFO:__main__:Replacing classifier with CodeBertStylometryClassifier\n"}
322
+ ,{"stream_name":"stderr","time":95.049344632,"data":"\n"}
323
+ ,{"stream_name":"stderr","time":95.049350837,"data":"INFO:__main__:Replacing classifier with CodeBertStylometryClassifier\n"}
324
+ ,{"stream_name":"stderr","time":95.055875813,"data":"2026-04-13 17:58:46,725 - INFO - Trainable parameters: 124,351,106\n"}
325
+ ,{"stream_name":"stderr","time":95.055897855,"data":"\n"}
326
+ ,{"stream_name":"stderr","time":95.055903405,"data":"2026-04-13 17:58:46,725 - INFO - Trainable parameters: 124,351,106\n"}
327
+ ,{"stream_name":"stderr","time":95.057090211,"data":"INFO:__main__:Trainable parameters: 124,351,106\n"}
328
+ ,{"stream_name":"stderr","time":95.057115801,"data":"\n"}
329
+ ,{"stream_name":"stderr","time":95.057121181,"data":"INFO:__main__:Trainable parameters: 124,351,106\n"}
330
+ ,{"stream_name":"stderr","time":95.058150363,"data":"2026-04-13 17:58:46,727 - INFO - Preparing datasets...\n"}
331
+ ,{"stream_name":"stderr","time":95.058169077,"data":"\n"}
332
+ ,{"stream_name":"stderr","time":95.058173367,"data":"2026-04-13 17:58:46,727 - INFO - Preparing datasets...\n"}
333
+ ,{"stream_name":"stderr","time":95.059183709,"data":"INFO:__main__:Preparing datasets...\n"}
334
+ ,{"stream_name":"stderr","time":95.059204659,"data":"\n"}
335
+ ,{"stream_name":"stderr","time":95.059211017,"data":"INFO:__main__:Preparing datasets...\n"}
336
+ ,{"stream_name":"stderr","time":95.060118367,"data":"2026-04-13 17:58:46,728 - INFO - Preparing datasets...\n"}
337
+ ,{"stream_name":"stderr","time":95.060128149,"data":"\n"}
338
+ ,{"stream_name":"stderr","time":95.060132494,"data":"2026-04-13 17:58:46,728 - INFO - Preparing datasets...\n"}
339
+ ,{"stream_name":"stderr","time":95.069963716,"data":"INFO:__main__:Preparing datasets...\n"}
340
+ ,{"stream_name":"stderr","time":95.070005454,"data":"\n"}
341
+ ,{"stream_name":"stderr","time":95.070012079,"data":"INFO:__main__:Preparing datasets...\n"}
342
+ ,{"stream_name":"stderr","time":587.008951729,"data":"2026-04-13 18:06:58,678 - INFO - Before balancing - Class 0: 238475, Class 1: 261525\n"}
343
+ ,{"stream_name":"stderr","time":587.00903261,"data":"\n"}
344
+ ,{"stream_name":"stderr","time":587.009040542,"data":"2026-04-13 18:06:58,678 - INFO - Before balancing - Class 0: 238475, Class 1: 261525\n"}
345
+ ,{"stream_name":"stderr","time":587.021390381,"data":"INFO:__main__:Before balancing - Class 0: 238475, Class 1: 261525\n"}
346
+ ,{"stream_name":"stderr","time":587.021434138,"data":"\n"}
347
+ ,{"stream_name":"stderr","time":587.021439556,"data":"INFO:__main__:Before balancing - Class 0: 238475, Class 1: 261525\n"}
348
+ ,{"stream_name":"stderr","time":587.295967334,"data":"2026-04-13 18:06:58,965 - INFO - After balancing - Added 23050 samples to Class 0. Total: 523050\n"}
349
+ ,{"stream_name":"stderr","time":587.296001037,"data":"\n"}
350
+ ,{"stream_name":"stderr","time":587.296006489,"data":"2026-04-13 18:06:58,965 - INFO - After balancing - Added 23050 samples to Class 0. Total: 523050\n"}
351
+ ,{"stream_name":"stderr","time":587.296879801,"data":"INFO:__main__:After balancing - Added 23050 samples to Class 0. Total: 523050\n"}
352
+ ,{"stream_name":"stderr","time":587.296902251,"data":"\n"}
353
+ ,{"stream_name":"stderr","time":587.296907617,"data":"INFO:__main__:After balancing - Added 23050 samples to Class 0. Total: 523050\n"}
354
+ ,{"stream_name":"stderr","time":587.595908403,"data":"2026-04-13 18:06:59,265 - INFO - Loss: FocalLossWithSmoothing\n"}
355
+ ,{"stream_name":"stderr","time":587.595954177,"data":"\n"}
356
+ ,{"stream_name":"stderr","time":587.595959893,"data":"2026-04-13 18:06:59,265 - INFO - Loss: FocalLossWithSmoothing\n"}
357
+ ,{"stream_name":"stderr","time":587.596929639,"data":"INFO:__main__:Loss: FocalLossWithSmoothing\n"}
358
+ ,{"stream_name":"stderr","time":587.596960025,"data":"\n"}
359
+ ,{"stream_name":"stderr","time":587.596966398,"data":"INFO:__main__:Loss: FocalLossWithSmoothing\n"}
360
+ ,{"stream_name":"stderr","time":587.7727685,"data":"2026-04-13 18:06:59,442 - INFO - ***** Training Configuration *****\n"}
361
+ ,{"stream_name":"stderr","time":587.772798186,"data":"\n"}
362
+ ,{"stream_name":"stderr","time":587.772802579,"data":"2026-04-13 18:06:59,442 - INFO - ***** Training Configuration *****\n"}
363
+ ,{"stream_name":"stderr","time":587.77387027,"data":"INFO:__main__:***** Training Configuration *****\n"}
364
+ ,{"stream_name":"stderr","time":587.773893148,"data":"\n"}
365
+ ,{"stream_name":"stderr","time":587.773897876,"data":"INFO:__main__:***** Training Configuration *****\n"}
366
+ ,{"stream_name":"stderr","time":587.774937444,"data":"2026-04-13 18:06:59,443 - INFO - Backbone : microsoft/codebert-base\n"}
367
+ ,{"stream_name":"stderr","time":587.774955894,"data":"\n"}
368
+ ,{"stream_name":"stderr","time":587.774961115,"data":"2026-04-13 18:06:59,443 - INFO - Backbone : microsoft/codebert-base\n"}
369
+ ,{"stream_name":"stderr","time":587.776010097,"data":"INFO:__main__: Backbone : microsoft/codebert-base\n"}
370
+ ,{"stream_name":"stderr","time":587.77602899,"data":"\n"}
371
+ ,{"stream_name":"stderr","time":587.776034534,"data":"INFO:__main__: Backbone : microsoft/codebert-base\n"}
372
+ ,{"stream_name":"stderr","time":587.777150108,"data":"2026-04-13 18:06:59,444 - INFO - Max epochs : 1\n"}
373
+ ,{"stream_name":"stderr","time":587.777174675,"data":"\n"}
374
+ ,{"stream_name":"stderr","time":587.777180141,"data":"2026-04-13 18:06:59,444 - INFO - Max epochs : 1\n"}
375
+ ,{"stream_name":"stderr","time":587.77833302,"data":"INFO:__main__: Max epochs : 1\n"}
376
+ ,{"stream_name":"stderr","time":587.778356753,"data":"\n"}
377
+ ,{"stream_name":"stderr","time":587.778362978,"data":"INFO:__main__: Max epochs : 1\n"}
378
+ ,{"stream_name":"stderr","time":587.779316062,"data":"2026-04-13 18:06:59,446 - INFO - Batch size : 64\n"}
379
+ ,{"stream_name":"stderr","time":587.779331726,"data":"\n"}
380
+ ,{"stream_name":"stderr","time":587.779336165,"data":"2026-04-13 18:06:59,446 - INFO - Batch size : 64\n"}
381
+ ,{"stream_name":"stderr","time":587.780659705,"data":"INFO:__main__: Batch size : 64\n"}
382
+ ,{"stream_name":"stderr","time":587.780675418,"data":"\n"}
383
+ ,{"stream_name":"stderr","time":587.780681428,"data":"INFO:__main__: Batch size : 64\n"}
384
+ ,{"stream_name":"stderr","time":587.781800687,"data":"2026-04-13 18:06:59,447 - INFO - Learning rate : 2e-05\n"}
385
+ ,{"stream_name":"stderr","time":587.781817501,"data":"\n"}
386
+ ,{"stream_name":"stderr","time":587.781822588,"data":"2026-04-13 18:06:59,447 - INFO - Learning rate : 2e-05\n"}
387
+ ,{"stream_name":"stderr","time":587.782922023,"data":"INFO:__main__: Learning rate : 2e-05\n"}
388
+ ,{"stream_name":"stderr","time":587.782937803,"data":"\n"}
389
+ ,{"stream_name":"stderr","time":587.782942921,"data":"INFO:__main__: Learning rate : 2e-05\n"}
390
+ ,{"stream_name":"stderr","time":587.783951691,"data":"2026-04-13 18:06:59,448 - INFO - LR scheduler : cosine_with_restarts (num_cycles=2)\n"}
391
+ ,{"stream_name":"stderr","time":587.783968337,"data":"\n"}
392
+ ,{"stream_name":"stderr","time":587.783973215,"data":"2026-04-13 18:06:59,448 - INFO - LR scheduler : cosine_with_restarts (num_cycles=2)\n"}
393
+ ,{"stream_name":"stderr","time":587.785031185,"data":"INFO:__main__: LR scheduler : cosine_with_restarts (num_cycles=2)\n"}
394
+ ,{"stream_name":"stderr","time":587.785047826,"data":"\n"}
395
+ ,{"stream_name":"stderr","time":587.785053045,"data":"INFO:__main__: LR scheduler : cosine_with_restarts (num_cycles=2)\n"}
396
+ ,{"stream_name":"stderr","time":587.786147401,"data":"2026-04-13 18:06:59,451 - INFO - Warmup steps : 490 / 8172\n"}
397
+ ,{"stream_name":"stderr","time":587.786163435,"data":"\n"}
398
+ ,{"stream_name":"stderr","time":587.786167972,"data":"2026-04-13 18:06:59,451 - INFO - Warmup steps : 490 / 8172\n"}
399
+ ,{"stream_name":"stderr","time":587.787228368,"data":"INFO:__main__: Warmup steps : 490 / 8172\n"}
400
+ ,{"stream_name":"stderr","time":587.78725123,"data":"\n"}
401
+ ,{"stream_name":"stderr","time":587.787256436,"data":"INFO:__main__: Warmup steps : 490 / 8172\n"}
402
+ ,{"stream_name":"stderr","time":587.788311197,"data":"2026-04-13 18:06:59,453 - INFO - LLRD : True (decay=0.9)\n"}
403
+ ,{"stream_name":"stderr","time":587.788326261,"data":"\n"}
404
+ ,{"stream_name":"stderr","time":587.788330257,"data":"2026-04-13 18:06:59,453 - INFO - LLRD : True (decay=0.9)\n"}
405
+ ,{"stream_name":"stderr","time":587.789366629,"data":"INFO:__main__: LLRD : True (decay=0.9)\n"}
406
+ ,{"stream_name":"stderr","time":587.789384716,"data":"\n"}
407
+ ,{"stream_name":"stderr","time":587.789387858,"data":"INFO:__main__: LLRD : True (decay=0.9)\n"}
408
+ ,{"stream_name":"stderr","time":587.790427516,"data":"2026-04-13 18:06:59,454 - INFO - Mixed precision : fp16=True, bf16=False\n"}
409
+ ,{"stream_name":"stderr","time":587.790446725,"data":"\n"}
410
+ ,{"stream_name":"stderr","time":587.790451728,"data":"2026-04-13 18:06:59,454 - INFO - Mixed precision : fp16=True, bf16=False\n"}
411
+ ,{"stream_name":"stderr","time":587.791930033,"data":"INFO:__main__: Mixed precision : fp16=True, bf16=False\n"}
412
+ ,{"stream_name":"stderr","time":587.791948484,"data":"\n"}
413
+ ,{"stream_name":"stderr","time":587.791953492,"data":"INFO:__main__: Mixed precision : fp16=True, bf16=False\n"}
414
+ ,{"stream_name":"stderr","time":587.792490874,"data":"2026-04-13 18:06:59,456 - INFO - Augmentation : True (p=0.25)\n"}
415
+ ,{"stream_name":"stderr","time":587.792509142,"data":"\n"}
416
+ ,{"stream_name":"stderr","time":587.792513896,"data":"2026-04-13 18:06:59,456 - INFO - Augmentation : True (p=0.25)\n"}
417
+ ,{"stream_name":"stderr","time":587.793638115,"data":"INFO:__main__: Augmentation : True (p=0.25)\n"}
418
+ ,{"stream_name":"stderr","time":587.793679381,"data":"\n"}
419
+ ,{"stream_name":"stderr","time":587.793685256,"data":"INFO:__main__: Augmentation : True (p=0.25)\n"}
420
+ ,{"stream_name":"stderr","time":587.79451031,"data":"2026-04-13 18:06:59,457 - INFO - Multi-sample dropout: False\n"}
421
+ ,{"stream_name":"stderr","time":587.794527128,"data":"\n"}
422
+ ,{"stream_name":"stderr","time":587.7945318,"data":"2026-04-13 18:06:59,457 - INFO - Multi-sample dropout: False\n"}
423
+ ,{"stream_name":"stderr","time":587.795423125,"data":"INFO:__main__: Multi-sample dropout: False\n"}
424
+ ,{"stream_name":"stderr","time":587.79543997,"data":"\n"}
425
+ ,{"stream_name":"stderr","time":587.795444907,"data":"INFO:__main__: Multi-sample dropout: False\n"}
426
+ ,{"stream_name":"stderr","time":587.796330766,"data":"2026-04-13 18:06:59,458 - INFO - R-Drop : True (weight=0.5)\n"}
427
+ ,{"stream_name":"stderr","time":587.796349675,"data":"\n"}
428
+ ,{"stream_name":"stderr","time":587.796354457,"data":"2026-04-13 18:06:59,458 - INFO - R-Drop : True (weight=0.5)\n"}
429
+ ,{"stream_name":"stderr","time":587.797150205,"data":"INFO:__main__: R-Drop : True (weight=0.5)\n"}
430
+ ,{"stream_name":"stderr","time":587.797167375,"data":"\n"}
431
+ ,{"stream_name":"stderr","time":587.797171993,"data":"INFO:__main__: R-Drop : True (weight=0.5)\n"}
432
+ ,{"stream_name":"stderr","time":587.797916311,"data":"2026-04-13 18:06:59,460 - INFO - Triplet loss : True (margin=1.0, weight=0.1)\n"}
433
+ ,{"stream_name":"stderr","time":587.797928785,"data":"\n"}
434
+ ,{"stream_name":"stderr","time":587.797932973,"data":"2026-04-13 18:06:59,460 - INFO - Triplet loss : True (margin=1.0, weight=0.1)\n"}
435
+ ,{"stream_name":"stderr","time":587.798719449,"data":"INFO:__main__: Triplet loss : True (margin=1.0, weight=0.1)\n"}
436
+ ,{"stream_name":"stderr","time":587.798736471,"data":"\n"}
437
+ ,{"stream_name":"stderr","time":587.798740873,"data":"INFO:__main__: Triplet loss : True (margin=1.0, weight=0.1)\n"}
438
+ ,{"stream_name":"stderr","time":587.799555214,"data":"2026-04-13 18:06:59,461 - INFO - MixCode : True (alpha=0.2)\n"}
439
+ ,{"stream_name":"stderr","time":587.799572463,"data":"\n"}
440
+ ,{"stream_name":"stderr","time":587.799576708,"data":"2026-04-13 18:06:59,461 - INFO - MixCode : True (alpha=0.2)\n"}
441
+ ,{"stream_name":"stderr","time":587.800416505,"data":"INFO:__main__: MixCode : True (alpha=0.2)\n"}
442
+ ,{"stream_name":"stderr","time":587.800433784,"data":"\n"}
443
+ ,{"stream_name":"stderr","time":587.800438689,"data":"INFO:__main__: MixCode : True (alpha=0.2)\n"}
444
+ ,{"stream_name":"stderr","time":587.801273718,"data":"2026-04-13 18:06:59,462 - INFO - PEFT (LoRA) : False\n"}
445
+ ,{"stream_name":"stderr","time":587.801291278,"data":"\n"}
446
+ ,{"stream_name":"stderr","time":587.801295981,"data":"2026-04-13 18:06:59,462 - INFO - PEFT (LoRA) : False\n"}
447
+ ,{"stream_name":"stderr","time":587.802085359,"data":"INFO:__main__: PEFT (LoRA) : False\n"}
448
+ ,{"stream_name":"stderr","time":587.8020979,"data":"\n"}
449
+ ,{"stream_name":"stderr","time":587.802114957,"data":"INFO:__main__: PEFT (LoRA) : False\n"}
450
+ ,{"stream_name":"stderr","time":587.802926621,"data":"2026-04-13 18:06:59,464 - INFO - Gradient checkpoint : True\n"}
451
+ ,{"stream_name":"stderr","time":587.802943748,"data":"\n"}
452
+ ,{"stream_name":"stderr","time":587.802948168,"data":"2026-04-13 18:06:59,464 - INFO - Gradient checkpoint : True\n"}
453
+ ,{"stream_name":"stderr","time":588.286185663,"data":"INFO:__main__: Gradient checkpoint : True\n"}
454
+ ,{"stream_name":"stderr","time":588.286246382,"data":"\n"}
455
+ ,{"stream_name":"stderr","time":588.286254056,"data":"INFO:__main__: Gradient checkpoint : True\n"}
456
+ ]
taskA-codebert-base-focal/checkpoint-600/config.json ADDED
@@ -0,0 +1,28 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "RobertaForSequenceClassification"
4
+ ],
5
+ "attention_probs_dropout_prob": 0.1,
6
+ "bos_token_id": 0,
7
+ "classifier_dropout": null,
8
+ "dtype": "float32",
9
+ "eos_token_id": 2,
10
+ "hidden_act": "gelu",
11
+ "hidden_dropout_prob": 0.1,
12
+ "hidden_size": 768,
13
+ "initializer_range": 0.02,
14
+ "intermediate_size": 3072,
15
+ "layer_norm_eps": 1e-05,
16
+ "max_position_embeddings": 514,
17
+ "model_type": "roberta",
18
+ "num_attention_heads": 12,
19
+ "num_hidden_layers": 12,
20
+ "output_past": true,
21
+ "pad_token_id": 1,
22
+ "position_embedding_type": "absolute",
23
+ "problem_type": "single_label_classification",
24
+ "transformers_version": "4.56.0",
25
+ "type_vocab_size": 1,
26
+ "use_cache": true,
27
+ "vocab_size": 50265
28
+ }
taskA-codebert-base-focal/checkpoint-600/merges.txt ADDED
The diff for this file is too large to render. See raw diff
 
taskA-codebert-base-focal/checkpoint-600/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:47bd788b85a09dffa89bae3ba9569ff70403400bb38dae2d3c9a242a64010342
3
+ size 500185080
taskA-codebert-base-focal/checkpoint-600/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:879a9b07c456c6dca73868fa2e3278d0bb5aaaf47d2613f8ac75871f309b0174
3
+ size 994980619
taskA-codebert-base-focal/checkpoint-600/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d13d28cbd8dea6f99b08a5286e6843028dcf2fd9a6e4cd2045450a8a6ee7fc04
3
+ size 14645
taskA-codebert-base-focal/checkpoint-600/scaler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b21c5349d5e7d02de630ebc1cb53ade1d9c6079eeb8594d223bb786011a0428b
3
+ size 1383
taskA-codebert-base-focal/checkpoint-600/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2f55aca7b64a847467387b78ff8558964def3d2cb5259c8271ce255811645cc2
3
+ size 1977
taskA-codebert-base-focal/checkpoint-600/special_tokens_map.json ADDED
@@ -0,0 +1,51 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": {
3
+ "content": "<s>",
4
+ "lstrip": false,
5
+ "normalized": true,
6
+ "rstrip": false,
7
+ "single_word": false
8
+ },
9
+ "cls_token": {
10
+ "content": "<s>",
11
+ "lstrip": false,
12
+ "normalized": true,
13
+ "rstrip": false,
14
+ "single_word": false
15
+ },
16
+ "eos_token": {
17
+ "content": "</s>",
18
+ "lstrip": false,
19
+ "normalized": true,
20
+ "rstrip": false,
21
+ "single_word": false
22
+ },
23
+ "mask_token": {
24
+ "content": "<mask>",
25
+ "lstrip": true,
26
+ "normalized": false,
27
+ "rstrip": false,
28
+ "single_word": false
29
+ },
30
+ "pad_token": {
31
+ "content": "<pad>",
32
+ "lstrip": false,
33
+ "normalized": true,
34
+ "rstrip": false,
35
+ "single_word": false
36
+ },
37
+ "sep_token": {
38
+ "content": "</s>",
39
+ "lstrip": false,
40
+ "normalized": true,
41
+ "rstrip": false,
42
+ "single_word": false
43
+ },
44
+ "unk_token": {
45
+ "content": "<unk>",
46
+ "lstrip": false,
47
+ "normalized": true,
48
+ "rstrip": false,
49
+ "single_word": false
50
+ }
51
+ }
taskA-codebert-base-focal/checkpoint-600/tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
taskA-codebert-base-focal/checkpoint-600/tokenizer_config.json ADDED
@@ -0,0 +1,58 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_prefix_space": false,
3
+ "added_tokens_decoder": {
4
+ "0": {
5
+ "content": "<s>",
6
+ "lstrip": false,
7
+ "normalized": true,
8
+ "rstrip": false,
9
+ "single_word": false,
10
+ "special": true
11
+ },
12
+ "1": {
13
+ "content": "<pad>",
14
+ "lstrip": false,
15
+ "normalized": true,
16
+ "rstrip": false,
17
+ "single_word": false,
18
+ "special": true
19
+ },
20
+ "2": {
21
+ "content": "</s>",
22
+ "lstrip": false,
23
+ "normalized": true,
24
+ "rstrip": false,
25
+ "single_word": false,
26
+ "special": true
27
+ },
28
+ "3": {
29
+ "content": "<unk>",
30
+ "lstrip": false,
31
+ "normalized": true,
32
+ "rstrip": false,
33
+ "single_word": false,
34
+ "special": true
35
+ },
36
+ "50264": {
37
+ "content": "<mask>",
38
+ "lstrip": true,
39
+ "normalized": false,
40
+ "rstrip": false,
41
+ "single_word": false,
42
+ "special": true
43
+ }
44
+ },
45
+ "bos_token": "<s>",
46
+ "clean_up_tokenization_spaces": false,
47
+ "cls_token": "<s>",
48
+ "eos_token": "</s>",
49
+ "errors": "replace",
50
+ "extra_special_tokens": {},
51
+ "mask_token": "<mask>",
52
+ "model_max_length": 512,
53
+ "pad_token": "<pad>",
54
+ "sep_token": "</s>",
55
+ "tokenizer_class": "RobertaTokenizer",
56
+ "trim_offsets": true,
57
+ "unk_token": "<unk>"
58
+ }
taskA-codebert-base-focal/checkpoint-600/trainer_state.json ADDED
@@ -0,0 +1,502 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": 600,
3
+ "best_metric": 0.9695240197651653,
4
+ "best_model_checkpoint": "taskA-codebert-base-focal/checkpoint-600",
5
+ "epoch": 0.14680694886224616,
6
+ "eval_steps": 200,
7
+ "global_step": 600,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 0.002446782481037436,
14
+ "grad_norm": 1.0,
15
+ "learning_rate": 3.6734693877551025e-07,
16
+ "loss": 1.1949,
17
+ "step": 10
18
+ },
19
+ {
20
+ "epoch": 0.004893564962074872,
21
+ "grad_norm": 0.9999998807907104,
22
+ "learning_rate": 7.755102040816327e-07,
23
+ "loss": 1.1644,
24
+ "step": 20
25
+ },
26
+ {
27
+ "epoch": 0.007340347443112307,
28
+ "grad_norm": 0.9999999403953552,
29
+ "learning_rate": 1.1836734693877552e-06,
30
+ "loss": 0.96,
31
+ "step": 30
32
+ },
33
+ {
34
+ "epoch": 0.009787129924149743,
35
+ "grad_norm": 1.0,
36
+ "learning_rate": 1.5918367346938775e-06,
37
+ "loss": 0.7557,
38
+ "step": 40
39
+ },
40
+ {
41
+ "epoch": 0.012233912405187179,
42
+ "grad_norm": 1.0,
43
+ "learning_rate": 2.0000000000000003e-06,
44
+ "loss": 0.5786,
45
+ "step": 50
46
+ },
47
+ {
48
+ "epoch": 0.014680694886224614,
49
+ "grad_norm": 0.9999999403953552,
50
+ "learning_rate": 2.4081632653061225e-06,
51
+ "loss": 0.4813,
52
+ "step": 60
53
+ },
54
+ {
55
+ "epoch": 0.01712747736726205,
56
+ "grad_norm": 0.9999999403953552,
57
+ "learning_rate": 2.816326530612245e-06,
58
+ "loss": 0.4477,
59
+ "step": 70
60
+ },
61
+ {
62
+ "epoch": 0.019574259848299486,
63
+ "grad_norm": 1.0,
64
+ "learning_rate": 3.2244897959183672e-06,
65
+ "loss": 0.4254,
66
+ "step": 80
67
+ },
68
+ {
69
+ "epoch": 0.022021042329336923,
70
+ "grad_norm": 1.0,
71
+ "learning_rate": 3.6326530612244903e-06,
72
+ "loss": 0.4017,
73
+ "step": 90
74
+ },
75
+ {
76
+ "epoch": 0.024467824810374357,
77
+ "grad_norm": 0.9999999403953552,
78
+ "learning_rate": 4.040816326530612e-06,
79
+ "loss": 0.379,
80
+ "step": 100
81
+ },
82
+ {
83
+ "epoch": 0.026914607291411794,
84
+ "grad_norm": 1.0000001192092896,
85
+ "learning_rate": 4.448979591836735e-06,
86
+ "loss": 0.3749,
87
+ "step": 110
88
+ },
89
+ {
90
+ "epoch": 0.029361389772449228,
91
+ "grad_norm": 1.0,
92
+ "learning_rate": 4.857142857142858e-06,
93
+ "loss": 0.3536,
94
+ "step": 120
95
+ },
96
+ {
97
+ "epoch": 0.031808172253486665,
98
+ "grad_norm": 0.9999999403953552,
99
+ "learning_rate": 5.26530612244898e-06,
100
+ "loss": 0.336,
101
+ "step": 130
102
+ },
103
+ {
104
+ "epoch": 0.0342549547345241,
105
+ "grad_norm": 1.0000001192092896,
106
+ "learning_rate": 5.673469387755103e-06,
107
+ "loss": 0.3261,
108
+ "step": 140
109
+ },
110
+ {
111
+ "epoch": 0.03670173721556154,
112
+ "grad_norm": 0.9999999403953552,
113
+ "learning_rate": 6.0816326530612245e-06,
114
+ "loss": 0.3134,
115
+ "step": 150
116
+ },
117
+ {
118
+ "epoch": 0.03914851969659897,
119
+ "grad_norm": 1.0,
120
+ "learning_rate": 6.489795918367348e-06,
121
+ "loss": 0.308,
122
+ "step": 160
123
+ },
124
+ {
125
+ "epoch": 0.041595302177636406,
126
+ "grad_norm": 1.0,
127
+ "learning_rate": 6.8979591836734705e-06,
128
+ "loss": 0.2961,
129
+ "step": 170
130
+ },
131
+ {
132
+ "epoch": 0.04404208465867385,
133
+ "grad_norm": 1.0,
134
+ "learning_rate": 7.306122448979592e-06,
135
+ "loss": 0.2998,
136
+ "step": 180
137
+ },
138
+ {
139
+ "epoch": 0.04648886713971128,
140
+ "grad_norm": 1.0,
141
+ "learning_rate": 7.714285714285716e-06,
142
+ "loss": 0.2895,
143
+ "step": 190
144
+ },
145
+ {
146
+ "epoch": 0.048935649620748714,
147
+ "grad_norm": 1.0,
148
+ "learning_rate": 8.122448979591837e-06,
149
+ "loss": 0.2773,
150
+ "step": 200
151
+ },
152
+ {
153
+ "epoch": 0.048935649620748714,
154
+ "eval_accuracy": 0.81444,
155
+ "eval_f1_weighted": 0.8120194753530078,
156
+ "eval_loss": 0.10672978311777115,
157
+ "eval_macro_f1": 0.8128217919062601,
158
+ "eval_precision": 0.8426688032270206,
159
+ "eval_recall": 0.81444,
160
+ "eval_runtime": 1609.4299,
161
+ "eval_samples_per_second": 62.134,
162
+ "eval_steps_per_second": 0.243,
163
+ "step": 200
164
+ },
165
+ {
166
+ "epoch": 0.051382432101786155,
167
+ "grad_norm": 1.0,
168
+ "learning_rate": 8.530612244897961e-06,
169
+ "loss": 0.2658,
170
+ "step": 210
171
+ },
172
+ {
173
+ "epoch": 0.05382921458282359,
174
+ "grad_norm": 1.0000001192092896,
175
+ "learning_rate": 8.938775510204082e-06,
176
+ "loss": 0.2574,
177
+ "step": 220
178
+ },
179
+ {
180
+ "epoch": 0.05627599706386102,
181
+ "grad_norm": 1.0,
182
+ "learning_rate": 9.346938775510204e-06,
183
+ "loss": 0.2633,
184
+ "step": 230
185
+ },
186
+ {
187
+ "epoch": 0.058722779544898455,
188
+ "grad_norm": 1.0,
189
+ "learning_rate": 9.755102040816327e-06,
190
+ "loss": 0.2503,
191
+ "step": 240
192
+ },
193
+ {
194
+ "epoch": 0.061169562025935896,
195
+ "grad_norm": 1.0,
196
+ "learning_rate": 1.016326530612245e-05,
197
+ "loss": 0.2359,
198
+ "step": 250
199
+ },
200
+ {
201
+ "epoch": 0.06361634450697333,
202
+ "grad_norm": 0.9999999403953552,
203
+ "learning_rate": 1.0571428571428572e-05,
204
+ "loss": 0.2327,
205
+ "step": 260
206
+ },
207
+ {
208
+ "epoch": 0.06606312698801077,
209
+ "grad_norm": 1.0,
210
+ "learning_rate": 1.0979591836734695e-05,
211
+ "loss": 0.2305,
212
+ "step": 270
213
+ },
214
+ {
215
+ "epoch": 0.0685099094690482,
216
+ "grad_norm": 1.0,
217
+ "learning_rate": 1.1387755102040819e-05,
218
+ "loss": 0.2284,
219
+ "step": 280
220
+ },
221
+ {
222
+ "epoch": 0.07095669195008564,
223
+ "grad_norm": 0.9999999403953552,
224
+ "learning_rate": 1.179591836734694e-05,
225
+ "loss": 0.226,
226
+ "step": 290
227
+ },
228
+ {
229
+ "epoch": 0.07340347443112308,
230
+ "grad_norm": 0.9999999403953552,
231
+ "learning_rate": 1.2204081632653062e-05,
232
+ "loss": 0.2225,
233
+ "step": 300
234
+ },
235
+ {
236
+ "epoch": 0.0758502569121605,
237
+ "grad_norm": 1.0,
238
+ "learning_rate": 1.2612244897959185e-05,
239
+ "loss": 0.2156,
240
+ "step": 310
241
+ },
242
+ {
243
+ "epoch": 0.07829703939319795,
244
+ "grad_norm": 1.0,
245
+ "learning_rate": 1.3020408163265308e-05,
246
+ "loss": 0.2033,
247
+ "step": 320
248
+ },
249
+ {
250
+ "epoch": 0.08074382187423539,
251
+ "grad_norm": 0.9999999403953552,
252
+ "learning_rate": 1.3428571428571429e-05,
253
+ "loss": 0.2021,
254
+ "step": 330
255
+ },
256
+ {
257
+ "epoch": 0.08319060435527281,
258
+ "grad_norm": 1.0,
259
+ "learning_rate": 1.3836734693877551e-05,
260
+ "loss": 0.1902,
261
+ "step": 340
262
+ },
263
+ {
264
+ "epoch": 0.08563738683631025,
265
+ "grad_norm": 1.0,
266
+ "learning_rate": 1.4244897959183674e-05,
267
+ "loss": 0.1975,
268
+ "step": 350
269
+ },
270
+ {
271
+ "epoch": 0.0880841693173477,
272
+ "grad_norm": 1.0,
273
+ "learning_rate": 1.4653061224489798e-05,
274
+ "loss": 0.1921,
275
+ "step": 360
276
+ },
277
+ {
278
+ "epoch": 0.09053095179838512,
279
+ "grad_norm": 0.9999999403953552,
280
+ "learning_rate": 1.506122448979592e-05,
281
+ "loss": 0.2039,
282
+ "step": 370
283
+ },
284
+ {
285
+ "epoch": 0.09297773427942256,
286
+ "grad_norm": 1.0000001192092896,
287
+ "learning_rate": 1.546938775510204e-05,
288
+ "loss": 0.202,
289
+ "step": 380
290
+ },
291
+ {
292
+ "epoch": 0.09542451676046,
293
+ "grad_norm": 0.9999999403953552,
294
+ "learning_rate": 1.5877551020408162e-05,
295
+ "loss": 0.1885,
296
+ "step": 390
297
+ },
298
+ {
299
+ "epoch": 0.09787129924149743,
300
+ "grad_norm": 0.9999998807907104,
301
+ "learning_rate": 1.6285714285714287e-05,
302
+ "loss": 0.1818,
303
+ "step": 400
304
+ },
305
+ {
306
+ "epoch": 0.09787129924149743,
307
+ "eval_accuracy": 0.9495,
308
+ "eval_f1_weighted": 0.9495267406510758,
309
+ "eval_loss": 0.039080820977687836,
310
+ "eval_macro_f1": 0.9494696691689186,
311
+ "eval_precision": 0.9504394545408126,
312
+ "eval_recall": 0.9495,
313
+ "eval_runtime": 1609.6857,
314
+ "eval_samples_per_second": 62.124,
315
+ "eval_steps_per_second": 0.243,
316
+ "step": 400
317
+ },
318
+ {
319
+ "epoch": 0.10031808172253487,
320
+ "grad_norm": 1.0,
321
+ "learning_rate": 1.669387755102041e-05,
322
+ "loss": 0.1813,
323
+ "step": 410
324
+ },
325
+ {
326
+ "epoch": 0.10276486420357231,
327
+ "grad_norm": 0.9999999403953552,
328
+ "learning_rate": 1.7102040816326532e-05,
329
+ "loss": 0.1758,
330
+ "step": 420
331
+ },
332
+ {
333
+ "epoch": 0.10521164668460974,
334
+ "grad_norm": 1.0000001192092896,
335
+ "learning_rate": 1.7510204081632653e-05,
336
+ "loss": 0.1778,
337
+ "step": 430
338
+ },
339
+ {
340
+ "epoch": 0.10765842916564718,
341
+ "grad_norm": 0.9999998807907104,
342
+ "learning_rate": 1.7918367346938777e-05,
343
+ "loss": 0.1681,
344
+ "step": 440
345
+ },
346
+ {
347
+ "epoch": 0.1101052116466846,
348
+ "grad_norm": 0.9999998807907104,
349
+ "learning_rate": 1.8326530612244898e-05,
350
+ "loss": 0.1715,
351
+ "step": 450
352
+ },
353
+ {
354
+ "epoch": 0.11255199412772204,
355
+ "grad_norm": 1.0,
356
+ "learning_rate": 1.8734693877551022e-05,
357
+ "loss": 0.1773,
358
+ "step": 460
359
+ },
360
+ {
361
+ "epoch": 0.11499877660875948,
362
+ "grad_norm": 1.0,
363
+ "learning_rate": 1.9142857142857146e-05,
364
+ "loss": 0.1859,
365
+ "step": 470
366
+ },
367
+ {
368
+ "epoch": 0.11744555908979691,
369
+ "grad_norm": 1.0,
370
+ "learning_rate": 1.9551020408163267e-05,
371
+ "loss": 0.173,
372
+ "step": 480
373
+ },
374
+ {
375
+ "epoch": 0.11989234157083435,
376
+ "grad_norm": 1.0,
377
+ "learning_rate": 1.9959183673469388e-05,
378
+ "loss": 0.1753,
379
+ "step": 490
380
+ },
381
+ {
382
+ "epoch": 0.12233912405187179,
383
+ "grad_norm": 0.9999999403953552,
384
+ "learning_rate": 1.9998764266160687e-05,
385
+ "loss": 0.1712,
386
+ "step": 500
387
+ },
388
+ {
389
+ "epoch": 0.12478590653290922,
390
+ "grad_norm": 1.0,
391
+ "learning_rate": 1.9994492985725524e-05,
392
+ "loss": 0.1644,
393
+ "step": 510
394
+ },
395
+ {
396
+ "epoch": 0.12723268901394666,
397
+ "grad_norm": 0.9999998807907104,
398
+ "learning_rate": 1.9987172205655365e-05,
399
+ "loss": 0.1585,
400
+ "step": 520
401
+ },
402
+ {
403
+ "epoch": 0.1296794714949841,
404
+ "grad_norm": 1.0000001192092896,
405
+ "learning_rate": 1.9976804159651927e-05,
406
+ "loss": 0.1625,
407
+ "step": 530
408
+ },
409
+ {
410
+ "epoch": 0.13212625397602154,
411
+ "grad_norm": 1.0,
412
+ "learning_rate": 1.9963392011192586e-05,
413
+ "loss": 0.1656,
414
+ "step": 540
415
+ },
416
+ {
417
+ "epoch": 0.13457303645705898,
418
+ "grad_norm": 0.9999999403953552,
419
+ "learning_rate": 1.994693985256516e-05,
420
+ "loss": 0.1528,
421
+ "step": 550
422
+ },
423
+ {
424
+ "epoch": 0.1370198189380964,
425
+ "grad_norm": 0.9999999403953552,
426
+ "learning_rate": 1.9927452703619262e-05,
427
+ "loss": 0.1547,
428
+ "step": 560
429
+ },
430
+ {
431
+ "epoch": 0.13946660141913383,
432
+ "grad_norm": 1.0,
433
+ "learning_rate": 1.9904936510234648e-05,
434
+ "loss": 0.1685,
435
+ "step": 570
436
+ },
437
+ {
438
+ "epoch": 0.14191338390017127,
439
+ "grad_norm": 0.9999999403953552,
440
+ "learning_rate": 1.987939814250705e-05,
441
+ "loss": 0.1568,
442
+ "step": 580
443
+ },
444
+ {
445
+ "epoch": 0.14436016638120872,
446
+ "grad_norm": 1.0,
447
+ "learning_rate": 1.985084539265195e-05,
448
+ "loss": 0.1589,
449
+ "step": 590
450
+ },
451
+ {
452
+ "epoch": 0.14680694886224616,
453
+ "grad_norm": 1.0,
454
+ "learning_rate": 1.9819286972627066e-05,
455
+ "loss": 0.1617,
456
+ "step": 600
457
+ },
458
+ {
459
+ "epoch": 0.14680694886224616,
460
+ "eval_accuracy": 0.96956,
461
+ "eval_f1_weighted": 0.9695722935959534,
462
+ "eval_loss": 0.02569369599223137,
463
+ "eval_macro_f1": 0.9695240197651653,
464
+ "eval_precision": 0.9698519718113257,
465
+ "eval_recall": 0.96956,
466
+ "eval_runtime": 1608.5431,
467
+ "eval_samples_per_second": 62.168,
468
+ "eval_steps_per_second": 0.243,
469
+ "step": 600
470
+ }
471
+ ],
472
+ "logging_steps": 10,
473
+ "max_steps": 4087,
474
+ "num_input_tokens_seen": 0,
475
+ "num_train_epochs": 1,
476
+ "save_steps": 200,
477
+ "stateful_callbacks": {
478
+ "EarlyStoppingCallback": {
479
+ "args": {
480
+ "early_stopping_patience": 3,
481
+ "early_stopping_threshold": 0.0
482
+ },
483
+ "attributes": {
484
+ "early_stopping_patience_counter": 0
485
+ }
486
+ },
487
+ "TrainerControl": {
488
+ "args": {
489
+ "should_epoch_stop": false,
490
+ "should_evaluate": false,
491
+ "should_log": false,
492
+ "should_save": true,
493
+ "should_training_stop": false
494
+ },
495
+ "attributes": {}
496
+ }
497
+ },
498
+ "total_flos": 2.0299639947264e+16,
499
+ "train_batch_size": 128,
500
+ "trial_name": null,
501
+ "trial_params": null
502
+ }
taskA-codebert-base-focal/checkpoint-600/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:25442a0fc63ab39a0b4278f6954102923d048206f65e3a867203c3b922f90a51
3
+ size 5777
taskA-codebert-base-focal/checkpoint-600/vocab.json ADDED
The diff for this file is too large to render. See raw diff
 
taskA-codebert-base-focal/checkpoint-800/config.json ADDED
@@ -0,0 +1,28 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "RobertaForSequenceClassification"
4
+ ],
5
+ "attention_probs_dropout_prob": 0.1,
6
+ "bos_token_id": 0,
7
+ "classifier_dropout": null,
8
+ "dtype": "float32",
9
+ "eos_token_id": 2,
10
+ "hidden_act": "gelu",
11
+ "hidden_dropout_prob": 0.1,
12
+ "hidden_size": 768,
13
+ "initializer_range": 0.02,
14
+ "intermediate_size": 3072,
15
+ "layer_norm_eps": 1e-05,
16
+ "max_position_embeddings": 514,
17
+ "model_type": "roberta",
18
+ "num_attention_heads": 12,
19
+ "num_hidden_layers": 12,
20
+ "output_past": true,
21
+ "pad_token_id": 1,
22
+ "position_embedding_type": "absolute",
23
+ "problem_type": "single_label_classification",
24
+ "transformers_version": "4.56.0",
25
+ "type_vocab_size": 1,
26
+ "use_cache": true,
27
+ "vocab_size": 50265
28
+ }
taskA-codebert-base-focal/checkpoint-800/merges.txt ADDED
The diff for this file is too large to render. See raw diff
 
taskA-codebert-base-focal/checkpoint-800/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7f885697b847c735434586c6d419aa01b48ecf1438d8541d53c24022a71382c5
3
+ size 500185080
taskA-codebert-base-focal/checkpoint-800/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:84e12c96f22c50f662c931eaa3a5541ca7e327e5bdfc79de81b6efd0648568c0
3
+ size 994980619
taskA-codebert-base-focal/checkpoint-800/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:16bdc36deeaf14775fd85d01ecedc37916349bbe5f07f55e0288ff89ba9db38c
3
+ size 14645
taskA-codebert-base-focal/checkpoint-800/scaler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b21c5349d5e7d02de630ebc1cb53ade1d9c6079eeb8594d223bb786011a0428b
3
+ size 1383
taskA-codebert-base-focal/checkpoint-800/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:615a9903cb94fb5599429d59be543997f77d8bf0001ac20b40aebe266da6dbec
3
+ size 1977
taskA-codebert-base-focal/checkpoint-800/special_tokens_map.json ADDED
@@ -0,0 +1,51 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": {
3
+ "content": "<s>",
4
+ "lstrip": false,
5
+ "normalized": true,
6
+ "rstrip": false,
7
+ "single_word": false
8
+ },
9
+ "cls_token": {
10
+ "content": "<s>",
11
+ "lstrip": false,
12
+ "normalized": true,
13
+ "rstrip": false,
14
+ "single_word": false
15
+ },
16
+ "eos_token": {
17
+ "content": "</s>",
18
+ "lstrip": false,
19
+ "normalized": true,
20
+ "rstrip": false,
21
+ "single_word": false
22
+ },
23
+ "mask_token": {
24
+ "content": "<mask>",
25
+ "lstrip": true,
26
+ "normalized": false,
27
+ "rstrip": false,
28
+ "single_word": false
29
+ },
30
+ "pad_token": {
31
+ "content": "<pad>",
32
+ "lstrip": false,
33
+ "normalized": true,
34
+ "rstrip": false,
35
+ "single_word": false
36
+ },
37
+ "sep_token": {
38
+ "content": "</s>",
39
+ "lstrip": false,
40
+ "normalized": true,
41
+ "rstrip": false,
42
+ "single_word": false
43
+ },
44
+ "unk_token": {
45
+ "content": "<unk>",
46
+ "lstrip": false,
47
+ "normalized": true,
48
+ "rstrip": false,
49
+ "single_word": false
50
+ }
51
+ }
taskA-codebert-base-focal/checkpoint-800/tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
taskA-codebert-base-focal/checkpoint-800/tokenizer_config.json ADDED
@@ -0,0 +1,58 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_prefix_space": false,
3
+ "added_tokens_decoder": {
4
+ "0": {
5
+ "content": "<s>",
6
+ "lstrip": false,
7
+ "normalized": true,
8
+ "rstrip": false,
9
+ "single_word": false,
10
+ "special": true
11
+ },
12
+ "1": {
13
+ "content": "<pad>",
14
+ "lstrip": false,
15
+ "normalized": true,
16
+ "rstrip": false,
17
+ "single_word": false,
18
+ "special": true
19
+ },
20
+ "2": {
21
+ "content": "</s>",
22
+ "lstrip": false,
23
+ "normalized": true,
24
+ "rstrip": false,
25
+ "single_word": false,
26
+ "special": true
27
+ },
28
+ "3": {
29
+ "content": "<unk>",
30
+ "lstrip": false,
31
+ "normalized": true,
32
+ "rstrip": false,
33
+ "single_word": false,
34
+ "special": true
35
+ },
36
+ "50264": {
37
+ "content": "<mask>",
38
+ "lstrip": true,
39
+ "normalized": false,
40
+ "rstrip": false,
41
+ "single_word": false,
42
+ "special": true
43
+ }
44
+ },
45
+ "bos_token": "<s>",
46
+ "clean_up_tokenization_spaces": false,
47
+ "cls_token": "<s>",
48
+ "eos_token": "</s>",
49
+ "errors": "replace",
50
+ "extra_special_tokens": {},
51
+ "mask_token": "<mask>",
52
+ "model_max_length": 512,
53
+ "pad_token": "<pad>",
54
+ "sep_token": "</s>",
55
+ "tokenizer_class": "RobertaTokenizer",
56
+ "trim_offsets": true,
57
+ "unk_token": "<unk>"
58
+ }
taskA-codebert-base-focal/checkpoint-800/trainer_state.json ADDED
@@ -0,0 +1,655 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": 800,
3
+ "best_metric": 0.9775107622870663,
4
+ "best_model_checkpoint": "taskA-codebert-base-focal/checkpoint-800",
5
+ "epoch": 0.19574259848299486,
6
+ "eval_steps": 200,
7
+ "global_step": 800,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 0.002446782481037436,
14
+ "grad_norm": 1.0,
15
+ "learning_rate": 3.6734693877551025e-07,
16
+ "loss": 1.1949,
17
+ "step": 10
18
+ },
19
+ {
20
+ "epoch": 0.004893564962074872,
21
+ "grad_norm": 0.9999998807907104,
22
+ "learning_rate": 7.755102040816327e-07,
23
+ "loss": 1.1644,
24
+ "step": 20
25
+ },
26
+ {
27
+ "epoch": 0.007340347443112307,
28
+ "grad_norm": 0.9999999403953552,
29
+ "learning_rate": 1.1836734693877552e-06,
30
+ "loss": 0.96,
31
+ "step": 30
32
+ },
33
+ {
34
+ "epoch": 0.009787129924149743,
35
+ "grad_norm": 1.0,
36
+ "learning_rate": 1.5918367346938775e-06,
37
+ "loss": 0.7557,
38
+ "step": 40
39
+ },
40
+ {
41
+ "epoch": 0.012233912405187179,
42
+ "grad_norm": 1.0,
43
+ "learning_rate": 2.0000000000000003e-06,
44
+ "loss": 0.5786,
45
+ "step": 50
46
+ },
47
+ {
48
+ "epoch": 0.014680694886224614,
49
+ "grad_norm": 0.9999999403953552,
50
+ "learning_rate": 2.4081632653061225e-06,
51
+ "loss": 0.4813,
52
+ "step": 60
53
+ },
54
+ {
55
+ "epoch": 0.01712747736726205,
56
+ "grad_norm": 0.9999999403953552,
57
+ "learning_rate": 2.816326530612245e-06,
58
+ "loss": 0.4477,
59
+ "step": 70
60
+ },
61
+ {
62
+ "epoch": 0.019574259848299486,
63
+ "grad_norm": 1.0,
64
+ "learning_rate": 3.2244897959183672e-06,
65
+ "loss": 0.4254,
66
+ "step": 80
67
+ },
68
+ {
69
+ "epoch": 0.022021042329336923,
70
+ "grad_norm": 1.0,
71
+ "learning_rate": 3.6326530612244903e-06,
72
+ "loss": 0.4017,
73
+ "step": 90
74
+ },
75
+ {
76
+ "epoch": 0.024467824810374357,
77
+ "grad_norm": 0.9999999403953552,
78
+ "learning_rate": 4.040816326530612e-06,
79
+ "loss": 0.379,
80
+ "step": 100
81
+ },
82
+ {
83
+ "epoch": 0.026914607291411794,
84
+ "grad_norm": 1.0000001192092896,
85
+ "learning_rate": 4.448979591836735e-06,
86
+ "loss": 0.3749,
87
+ "step": 110
88
+ },
89
+ {
90
+ "epoch": 0.029361389772449228,
91
+ "grad_norm": 1.0,
92
+ "learning_rate": 4.857142857142858e-06,
93
+ "loss": 0.3536,
94
+ "step": 120
95
+ },
96
+ {
97
+ "epoch": 0.031808172253486665,
98
+ "grad_norm": 0.9999999403953552,
99
+ "learning_rate": 5.26530612244898e-06,
100
+ "loss": 0.336,
101
+ "step": 130
102
+ },
103
+ {
104
+ "epoch": 0.0342549547345241,
105
+ "grad_norm": 1.0000001192092896,
106
+ "learning_rate": 5.673469387755103e-06,
107
+ "loss": 0.3261,
108
+ "step": 140
109
+ },
110
+ {
111
+ "epoch": 0.03670173721556154,
112
+ "grad_norm": 0.9999999403953552,
113
+ "learning_rate": 6.0816326530612245e-06,
114
+ "loss": 0.3134,
115
+ "step": 150
116
+ },
117
+ {
118
+ "epoch": 0.03914851969659897,
119
+ "grad_norm": 1.0,
120
+ "learning_rate": 6.489795918367348e-06,
121
+ "loss": 0.308,
122
+ "step": 160
123
+ },
124
+ {
125
+ "epoch": 0.041595302177636406,
126
+ "grad_norm": 1.0,
127
+ "learning_rate": 6.8979591836734705e-06,
128
+ "loss": 0.2961,
129
+ "step": 170
130
+ },
131
+ {
132
+ "epoch": 0.04404208465867385,
133
+ "grad_norm": 1.0,
134
+ "learning_rate": 7.306122448979592e-06,
135
+ "loss": 0.2998,
136
+ "step": 180
137
+ },
138
+ {
139
+ "epoch": 0.04648886713971128,
140
+ "grad_norm": 1.0,
141
+ "learning_rate": 7.714285714285716e-06,
142
+ "loss": 0.2895,
143
+ "step": 190
144
+ },
145
+ {
146
+ "epoch": 0.048935649620748714,
147
+ "grad_norm": 1.0,
148
+ "learning_rate": 8.122448979591837e-06,
149
+ "loss": 0.2773,
150
+ "step": 200
151
+ },
152
+ {
153
+ "epoch": 0.048935649620748714,
154
+ "eval_accuracy": 0.81444,
155
+ "eval_f1_weighted": 0.8120194753530078,
156
+ "eval_loss": 0.10672978311777115,
157
+ "eval_macro_f1": 0.8128217919062601,
158
+ "eval_precision": 0.8426688032270206,
159
+ "eval_recall": 0.81444,
160
+ "eval_runtime": 1609.4299,
161
+ "eval_samples_per_second": 62.134,
162
+ "eval_steps_per_second": 0.243,
163
+ "step": 200
164
+ },
165
+ {
166
+ "epoch": 0.051382432101786155,
167
+ "grad_norm": 1.0,
168
+ "learning_rate": 8.530612244897961e-06,
169
+ "loss": 0.2658,
170
+ "step": 210
171
+ },
172
+ {
173
+ "epoch": 0.05382921458282359,
174
+ "grad_norm": 1.0000001192092896,
175
+ "learning_rate": 8.938775510204082e-06,
176
+ "loss": 0.2574,
177
+ "step": 220
178
+ },
179
+ {
180
+ "epoch": 0.05627599706386102,
181
+ "grad_norm": 1.0,
182
+ "learning_rate": 9.346938775510204e-06,
183
+ "loss": 0.2633,
184
+ "step": 230
185
+ },
186
+ {
187
+ "epoch": 0.058722779544898455,
188
+ "grad_norm": 1.0,
189
+ "learning_rate": 9.755102040816327e-06,
190
+ "loss": 0.2503,
191
+ "step": 240
192
+ },
193
+ {
194
+ "epoch": 0.061169562025935896,
195
+ "grad_norm": 1.0,
196
+ "learning_rate": 1.016326530612245e-05,
197
+ "loss": 0.2359,
198
+ "step": 250
199
+ },
200
+ {
201
+ "epoch": 0.06361634450697333,
202
+ "grad_norm": 0.9999999403953552,
203
+ "learning_rate": 1.0571428571428572e-05,
204
+ "loss": 0.2327,
205
+ "step": 260
206
+ },
207
+ {
208
+ "epoch": 0.06606312698801077,
209
+ "grad_norm": 1.0,
210
+ "learning_rate": 1.0979591836734695e-05,
211
+ "loss": 0.2305,
212
+ "step": 270
213
+ },
214
+ {
215
+ "epoch": 0.0685099094690482,
216
+ "grad_norm": 1.0,
217
+ "learning_rate": 1.1387755102040819e-05,
218
+ "loss": 0.2284,
219
+ "step": 280
220
+ },
221
+ {
222
+ "epoch": 0.07095669195008564,
223
+ "grad_norm": 0.9999999403953552,
224
+ "learning_rate": 1.179591836734694e-05,
225
+ "loss": 0.226,
226
+ "step": 290
227
+ },
228
+ {
229
+ "epoch": 0.07340347443112308,
230
+ "grad_norm": 0.9999999403953552,
231
+ "learning_rate": 1.2204081632653062e-05,
232
+ "loss": 0.2225,
233
+ "step": 300
234
+ },
235
+ {
236
+ "epoch": 0.0758502569121605,
237
+ "grad_norm": 1.0,
238
+ "learning_rate": 1.2612244897959185e-05,
239
+ "loss": 0.2156,
240
+ "step": 310
241
+ },
242
+ {
243
+ "epoch": 0.07829703939319795,
244
+ "grad_norm": 1.0,
245
+ "learning_rate": 1.3020408163265308e-05,
246
+ "loss": 0.2033,
247
+ "step": 320
248
+ },
249
+ {
250
+ "epoch": 0.08074382187423539,
251
+ "grad_norm": 0.9999999403953552,
252
+ "learning_rate": 1.3428571428571429e-05,
253
+ "loss": 0.2021,
254
+ "step": 330
255
+ },
256
+ {
257
+ "epoch": 0.08319060435527281,
258
+ "grad_norm": 1.0,
259
+ "learning_rate": 1.3836734693877551e-05,
260
+ "loss": 0.1902,
261
+ "step": 340
262
+ },
263
+ {
264
+ "epoch": 0.08563738683631025,
265
+ "grad_norm": 1.0,
266
+ "learning_rate": 1.4244897959183674e-05,
267
+ "loss": 0.1975,
268
+ "step": 350
269
+ },
270
+ {
271
+ "epoch": 0.0880841693173477,
272
+ "grad_norm": 1.0,
273
+ "learning_rate": 1.4653061224489798e-05,
274
+ "loss": 0.1921,
275
+ "step": 360
276
+ },
277
+ {
278
+ "epoch": 0.09053095179838512,
279
+ "grad_norm": 0.9999999403953552,
280
+ "learning_rate": 1.506122448979592e-05,
281
+ "loss": 0.2039,
282
+ "step": 370
283
+ },
284
+ {
285
+ "epoch": 0.09297773427942256,
286
+ "grad_norm": 1.0000001192092896,
287
+ "learning_rate": 1.546938775510204e-05,
288
+ "loss": 0.202,
289
+ "step": 380
290
+ },
291
+ {
292
+ "epoch": 0.09542451676046,
293
+ "grad_norm": 0.9999999403953552,
294
+ "learning_rate": 1.5877551020408162e-05,
295
+ "loss": 0.1885,
296
+ "step": 390
297
+ },
298
+ {
299
+ "epoch": 0.09787129924149743,
300
+ "grad_norm": 0.9999998807907104,
301
+ "learning_rate": 1.6285714285714287e-05,
302
+ "loss": 0.1818,
303
+ "step": 400
304
+ },
305
+ {
306
+ "epoch": 0.09787129924149743,
307
+ "eval_accuracy": 0.9495,
308
+ "eval_f1_weighted": 0.9495267406510758,
309
+ "eval_loss": 0.039080820977687836,
310
+ "eval_macro_f1": 0.9494696691689186,
311
+ "eval_precision": 0.9504394545408126,
312
+ "eval_recall": 0.9495,
313
+ "eval_runtime": 1609.6857,
314
+ "eval_samples_per_second": 62.124,
315
+ "eval_steps_per_second": 0.243,
316
+ "step": 400
317
+ },
318
+ {
319
+ "epoch": 0.10031808172253487,
320
+ "grad_norm": 1.0,
321
+ "learning_rate": 1.669387755102041e-05,
322
+ "loss": 0.1813,
323
+ "step": 410
324
+ },
325
+ {
326
+ "epoch": 0.10276486420357231,
327
+ "grad_norm": 0.9999999403953552,
328
+ "learning_rate": 1.7102040816326532e-05,
329
+ "loss": 0.1758,
330
+ "step": 420
331
+ },
332
+ {
333
+ "epoch": 0.10521164668460974,
334
+ "grad_norm": 1.0000001192092896,
335
+ "learning_rate": 1.7510204081632653e-05,
336
+ "loss": 0.1778,
337
+ "step": 430
338
+ },
339
+ {
340
+ "epoch": 0.10765842916564718,
341
+ "grad_norm": 0.9999998807907104,
342
+ "learning_rate": 1.7918367346938777e-05,
343
+ "loss": 0.1681,
344
+ "step": 440
345
+ },
346
+ {
347
+ "epoch": 0.1101052116466846,
348
+ "grad_norm": 0.9999998807907104,
349
+ "learning_rate": 1.8326530612244898e-05,
350
+ "loss": 0.1715,
351
+ "step": 450
352
+ },
353
+ {
354
+ "epoch": 0.11255199412772204,
355
+ "grad_norm": 1.0,
356
+ "learning_rate": 1.8734693877551022e-05,
357
+ "loss": 0.1773,
358
+ "step": 460
359
+ },
360
+ {
361
+ "epoch": 0.11499877660875948,
362
+ "grad_norm": 1.0,
363
+ "learning_rate": 1.9142857142857146e-05,
364
+ "loss": 0.1859,
365
+ "step": 470
366
+ },
367
+ {
368
+ "epoch": 0.11744555908979691,
369
+ "grad_norm": 1.0,
370
+ "learning_rate": 1.9551020408163267e-05,
371
+ "loss": 0.173,
372
+ "step": 480
373
+ },
374
+ {
375
+ "epoch": 0.11989234157083435,
376
+ "grad_norm": 1.0,
377
+ "learning_rate": 1.9959183673469388e-05,
378
+ "loss": 0.1753,
379
+ "step": 490
380
+ },
381
+ {
382
+ "epoch": 0.12233912405187179,
383
+ "grad_norm": 0.9999999403953552,
384
+ "learning_rate": 1.9998764266160687e-05,
385
+ "loss": 0.1712,
386
+ "step": 500
387
+ },
388
+ {
389
+ "epoch": 0.12478590653290922,
390
+ "grad_norm": 1.0,
391
+ "learning_rate": 1.9994492985725524e-05,
392
+ "loss": 0.1644,
393
+ "step": 510
394
+ },
395
+ {
396
+ "epoch": 0.12723268901394666,
397
+ "grad_norm": 0.9999998807907104,
398
+ "learning_rate": 1.9987172205655365e-05,
399
+ "loss": 0.1585,
400
+ "step": 520
401
+ },
402
+ {
403
+ "epoch": 0.1296794714949841,
404
+ "grad_norm": 1.0000001192092896,
405
+ "learning_rate": 1.9976804159651927e-05,
406
+ "loss": 0.1625,
407
+ "step": 530
408
+ },
409
+ {
410
+ "epoch": 0.13212625397602154,
411
+ "grad_norm": 1.0,
412
+ "learning_rate": 1.9963392011192586e-05,
413
+ "loss": 0.1656,
414
+ "step": 540
415
+ },
416
+ {
417
+ "epoch": 0.13457303645705898,
418
+ "grad_norm": 0.9999999403953552,
419
+ "learning_rate": 1.994693985256516e-05,
420
+ "loss": 0.1528,
421
+ "step": 550
422
+ },
423
+ {
424
+ "epoch": 0.1370198189380964,
425
+ "grad_norm": 0.9999999403953552,
426
+ "learning_rate": 1.9927452703619262e-05,
427
+ "loss": 0.1547,
428
+ "step": 560
429
+ },
430
+ {
431
+ "epoch": 0.13946660141913383,
432
+ "grad_norm": 1.0,
433
+ "learning_rate": 1.9904936510234648e-05,
434
+ "loss": 0.1685,
435
+ "step": 570
436
+ },
437
+ {
438
+ "epoch": 0.14191338390017127,
439
+ "grad_norm": 0.9999999403953552,
440
+ "learning_rate": 1.987939814250705e-05,
441
+ "loss": 0.1568,
442
+ "step": 580
443
+ },
444
+ {
445
+ "epoch": 0.14436016638120872,
446
+ "grad_norm": 1.0,
447
+ "learning_rate": 1.985084539265195e-05,
448
+ "loss": 0.1589,
449
+ "step": 590
450
+ },
451
+ {
452
+ "epoch": 0.14680694886224616,
453
+ "grad_norm": 1.0,
454
+ "learning_rate": 1.9819286972627066e-05,
455
+ "loss": 0.1617,
456
+ "step": 600
457
+ },
458
+ {
459
+ "epoch": 0.14680694886224616,
460
+ "eval_accuracy": 0.96956,
461
+ "eval_f1_weighted": 0.9695722935959534,
462
+ "eval_loss": 0.02569369599223137,
463
+ "eval_macro_f1": 0.9695240197651653,
464
+ "eval_precision": 0.9698519718113257,
465
+ "eval_recall": 0.96956,
466
+ "eval_runtime": 1608.5431,
467
+ "eval_samples_per_second": 62.168,
468
+ "eval_steps_per_second": 0.243,
469
+ "step": 600
470
+ },
471
+ {
472
+ "epoch": 0.14925373134328357,
473
+ "grad_norm": 0.9999998807907104,
474
+ "learning_rate": 1.9784732511474162e-05,
475
+ "loss": 0.1533,
476
+ "step": 610
477
+ },
478
+ {
479
+ "epoch": 0.151700513824321,
480
+ "grad_norm": 1.0,
481
+ "learning_rate": 1.9747192552381052e-05,
482
+ "loss": 0.1569,
483
+ "step": 620
484
+ },
485
+ {
486
+ "epoch": 0.15414729630535845,
487
+ "grad_norm": 1.0,
488
+ "learning_rate": 1.9706678549464685e-05,
489
+ "loss": 0.1562,
490
+ "step": 630
491
+ },
492
+ {
493
+ "epoch": 0.1565940787863959,
494
+ "grad_norm": 1.0,
495
+ "learning_rate": 1.9663202864276302e-05,
496
+ "loss": 0.1552,
497
+ "step": 640
498
+ },
499
+ {
500
+ "epoch": 0.15904086126743333,
501
+ "grad_norm": 1.0,
502
+ "learning_rate": 1.9616778762029686e-05,
503
+ "loss": 0.1513,
504
+ "step": 650
505
+ },
506
+ {
507
+ "epoch": 0.16148764374847077,
508
+ "grad_norm": 1.0,
509
+ "learning_rate": 1.9567420407553707e-05,
510
+ "loss": 0.1564,
511
+ "step": 660
512
+ },
513
+ {
514
+ "epoch": 0.16393442622950818,
515
+ "grad_norm": 1.0,
516
+ "learning_rate": 1.95151428609704e-05,
517
+ "loss": 0.1497,
518
+ "step": 670
519
+ },
520
+ {
521
+ "epoch": 0.16638120871054562,
522
+ "grad_norm": 0.9999999403953552,
523
+ "learning_rate": 1.9459962073099824e-05,
524
+ "loss": 0.1541,
525
+ "step": 680
526
+ },
527
+ {
528
+ "epoch": 0.16882799119158307,
529
+ "grad_norm": 1.0,
530
+ "learning_rate": 1.940189488059319e-05,
531
+ "loss": 0.1489,
532
+ "step": 690
533
+ },
534
+ {
535
+ "epoch": 0.1712747736726205,
536
+ "grad_norm": 0.9999999403953552,
537
+ "learning_rate": 1.9340959000795707e-05,
538
+ "loss": 0.1509,
539
+ "step": 700
540
+ },
541
+ {
542
+ "epoch": 0.17372155615365795,
543
+ "grad_norm": 1.0,
544
+ "learning_rate": 1.927717302634068e-05,
545
+ "loss": 0.1455,
546
+ "step": 710
547
+ },
548
+ {
549
+ "epoch": 0.1761683386346954,
550
+ "grad_norm": 0.9999999403953552,
551
+ "learning_rate": 1.9210556419476582e-05,
552
+ "loss": 0.1448,
553
+ "step": 720
554
+ },
555
+ {
556
+ "epoch": 0.1786151211157328,
557
+ "grad_norm": 0.9999999403953552,
558
+ "learning_rate": 1.914112950612876e-05,
559
+ "loss": 0.1521,
560
+ "step": 730
561
+ },
562
+ {
563
+ "epoch": 0.18106190359677024,
564
+ "grad_norm": 0.9999999403953552,
565
+ "learning_rate": 1.9068913469697633e-05,
566
+ "loss": 0.153,
567
+ "step": 740
568
+ },
569
+ {
570
+ "epoch": 0.18350868607780768,
571
+ "grad_norm": 1.0,
572
+ "learning_rate": 1.899393034459526e-05,
573
+ "loss": 0.1461,
574
+ "step": 750
575
+ },
576
+ {
577
+ "epoch": 0.18595546855884512,
578
+ "grad_norm": 1.0,
579
+ "learning_rate": 1.8916203009522228e-05,
580
+ "loss": 0.1512,
581
+ "step": 760
582
+ },
583
+ {
584
+ "epoch": 0.18840225103988256,
585
+ "grad_norm": 1.0,
586
+ "learning_rate": 1.8835755180486984e-05,
587
+ "loss": 0.1459,
588
+ "step": 770
589
+ },
590
+ {
591
+ "epoch": 0.19084903352092,
592
+ "grad_norm": 1.0,
593
+ "learning_rate": 1.875261140356961e-05,
594
+ "loss": 0.1458,
595
+ "step": 780
596
+ },
597
+ {
598
+ "epoch": 0.19329581600195742,
599
+ "grad_norm": 0.9999999403953552,
600
+ "learning_rate": 1.8666797047432396e-05,
601
+ "loss": 0.1441,
602
+ "step": 790
603
+ },
604
+ {
605
+ "epoch": 0.19574259848299486,
606
+ "grad_norm": 1.0,
607
+ "learning_rate": 1.8578338295579413e-05,
608
+ "loss": 0.143,
609
+ "step": 800
610
+ },
611
+ {
612
+ "epoch": 0.19574259848299486,
613
+ "eval_accuracy": 0.97755,
614
+ "eval_f1_weighted": 0.9775540674957386,
615
+ "eval_loss": 0.019970815628767014,
616
+ "eval_macro_f1": 0.9775107622870663,
617
+ "eval_precision": 0.9775948396265879,
618
+ "eval_recall": 0.97755,
619
+ "eval_runtime": 1609.303,
620
+ "eval_samples_per_second": 62.139,
621
+ "eval_steps_per_second": 0.243,
622
+ "step": 800
623
+ }
624
+ ],
625
+ "logging_steps": 10,
626
+ "max_steps": 4087,
627
+ "num_input_tokens_seen": 0,
628
+ "num_train_epochs": 1,
629
+ "save_steps": 200,
630
+ "stateful_callbacks": {
631
+ "EarlyStoppingCallback": {
632
+ "args": {
633
+ "early_stopping_patience": 3,
634
+ "early_stopping_threshold": 0.0
635
+ },
636
+ "attributes": {
637
+ "early_stopping_patience_counter": 0
638
+ }
639
+ },
640
+ "TrainerControl": {
641
+ "args": {
642
+ "should_epoch_stop": false,
643
+ "should_evaluate": false,
644
+ "should_log": false,
645
+ "should_save": true,
646
+ "should_training_stop": false
647
+ },
648
+ "attributes": {}
649
+ }
650
+ },
651
+ "total_flos": 2.7066186596352e+16,
652
+ "train_batch_size": 128,
653
+ "trial_name": null,
654
+ "trial_params": null
655
+ }
taskA-codebert-base-focal/checkpoint-800/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:25442a0fc63ab39a0b4278f6954102923d048206f65e3a867203c3b922f90a51
3
+ size 5777
taskA-codebert-base-focal/checkpoint-800/vocab.json ADDED
The diff for this file is too large to render. See raw diff
 
taskA-unixcoder-focal/checkpoint-2200/config.json ADDED
@@ -0,0 +1,29 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "RobertaForSequenceClassification"
4
+ ],
5
+ "attention_probs_dropout_prob": 0.1,
6
+ "bos_token_id": 0,
7
+ "classifier_dropout": null,
8
+ "dtype": "float32",
9
+ "eos_token_id": 2,
10
+ "gradient_checkpointing": false,
11
+ "hidden_act": "gelu",
12
+ "hidden_dropout_prob": 0.1,
13
+ "hidden_size": 768,
14
+ "initializer_range": 0.02,
15
+ "intermediate_size": 3072,
16
+ "layer_norm_eps": 1e-05,
17
+ "max_position_embeddings": 1026,
18
+ "model_type": "roberta",
19
+ "num_attention_heads": 12,
20
+ "num_hidden_layers": 12,
21
+ "output_past": true,
22
+ "pad_token_id": 1,
23
+ "position_embedding_type": "absolute",
24
+ "problem_type": "single_label_classification",
25
+ "transformers_version": "4.56.0",
26
+ "type_vocab_size": 10,
27
+ "use_cache": true,
28
+ "vocab_size": 51416
29
+ }
taskA-unixcoder-focal/checkpoint-2200/merges.txt ADDED
The diff for this file is too large to render. See raw diff
 
taskA-unixcoder-focal/checkpoint-2200/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:335df2fb9c96b5e579051e5a801512ed4e3d23ef053139b2fb42f7445f061aaa
3
+ size 506505672
taskA-unixcoder-focal/checkpoint-2200/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5bcdee09dd93cdd197e6ea1eee09b9979eaf49d3baada43d87bee36a080a9c8d
3
+ size 1007622411
taskA-unixcoder-focal/checkpoint-2200/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:15c0d4180bda455330407ed0bb58739c7ec87b45b8bb6d957f4d11a53fa4cc59
3
+ size 14645
taskA-unixcoder-focal/checkpoint-2200/scaler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:90bbc0bb51ada144c607a2d0e8dae7a41e3cca0fcbc5a9257404d9d17b7a3f9d
3
+ size 1383
taskA-unixcoder-focal/checkpoint-2200/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a7f7f10addafb465c38e374d351212daa821676c6da17d02f7fb0ffb64d30980
3
+ size 1977
taskA-unixcoder-focal/checkpoint-2200/special_tokens_map.json ADDED
@@ -0,0 +1,51 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": {
3
+ "content": "<s>",
4
+ "lstrip": false,
5
+ "normalized": true,
6
+ "rstrip": false,
7
+ "single_word": false
8
+ },
9
+ "cls_token": {
10
+ "content": "<s>",
11
+ "lstrip": false,
12
+ "normalized": true,
13
+ "rstrip": false,
14
+ "single_word": false
15
+ },
16
+ "eos_token": {
17
+ "content": "</s>",
18
+ "lstrip": false,
19
+ "normalized": true,
20
+ "rstrip": false,
21
+ "single_word": false
22
+ },
23
+ "mask_token": {
24
+ "content": "<mask>",
25
+ "lstrip": true,
26
+ "normalized": true,
27
+ "rstrip": false,
28
+ "single_word": false
29
+ },
30
+ "pad_token": {
31
+ "content": "<pad>",
32
+ "lstrip": false,
33
+ "normalized": true,
34
+ "rstrip": false,
35
+ "single_word": false
36
+ },
37
+ "sep_token": {
38
+ "content": "</s>",
39
+ "lstrip": false,
40
+ "normalized": true,
41
+ "rstrip": false,
42
+ "single_word": false
43
+ },
44
+ "unk_token": {
45
+ "content": "<unk>",
46
+ "lstrip": false,
47
+ "normalized": true,
48
+ "rstrip": false,
49
+ "single_word": false
50
+ }
51
+ }
taskA-unixcoder-focal/checkpoint-2200/tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
taskA-unixcoder-focal/checkpoint-2200/tokenizer_config.json ADDED
@@ -0,0 +1,58 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_prefix_space": false,
3
+ "added_tokens_decoder": {
4
+ "0": {
5
+ "content": "<s>",
6
+ "lstrip": false,
7
+ "normalized": true,
8
+ "rstrip": false,
9
+ "single_word": false,
10
+ "special": true
11
+ },
12
+ "1": {
13
+ "content": "<pad>",
14
+ "lstrip": false,
15
+ "normalized": true,
16
+ "rstrip": false,
17
+ "single_word": false,
18
+ "special": true
19
+ },
20
+ "2": {
21
+ "content": "</s>",
22
+ "lstrip": false,
23
+ "normalized": true,
24
+ "rstrip": false,
25
+ "single_word": false,
26
+ "special": true
27
+ },
28
+ "3": {
29
+ "content": "<unk>",
30
+ "lstrip": false,
31
+ "normalized": true,
32
+ "rstrip": false,
33
+ "single_word": false,
34
+ "special": true
35
+ },
36
+ "4": {
37
+ "content": "<mask>",
38
+ "lstrip": true,
39
+ "normalized": true,
40
+ "rstrip": false,
41
+ "single_word": false,
42
+ "special": true
43
+ }
44
+ },
45
+ "bos_token": "<s>",
46
+ "clean_up_tokenization_spaces": false,
47
+ "cls_token": "<s>",
48
+ "eos_token": "</s>",
49
+ "errors": "replace",
50
+ "extra_special_tokens": {},
51
+ "mask_token": "<mask>",
52
+ "model_max_length": 1000000000000000019884624838656,
53
+ "pad_token": "<pad>",
54
+ "sep_token": "</s>",
55
+ "tokenizer_class": "RobertaTokenizer",
56
+ "trim_offsets": true,
57
+ "unk_token": "<unk>"
58
+ }
taskA-unixcoder-focal/checkpoint-2200/trainer_state.json ADDED
@@ -0,0 +1,1869 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": 2100,
3
+ "best_metric": 0.9776543362648145,
4
+ "best_model_checkpoint": "taskA-unixcoder-focal/checkpoint-2100",
5
+ "epoch": 0.2691790040376851,
6
+ "eval_steps": 100,
7
+ "global_step": 2200,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 0.001223540927444023,
14
+ "grad_norm": 0.1898110806941986,
15
+ "learning_rate": 3.6734693877551025e-07,
16
+ "loss": 0.4422,
17
+ "step": 10
18
+ },
19
+ {
20
+ "epoch": 0.002447081854888046,
21
+ "grad_norm": 0.3350900411605835,
22
+ "learning_rate": 7.346938775510205e-07,
23
+ "loss": 0.4323,
24
+ "step": 20
25
+ },
26
+ {
27
+ "epoch": 0.003670622782332069,
28
+ "grad_norm": 0.3723449110984802,
29
+ "learning_rate": 1.142857142857143e-06,
30
+ "loss": 0.4362,
31
+ "step": 30
32
+ },
33
+ {
34
+ "epoch": 0.004894163709776092,
35
+ "grad_norm": 0.3299769461154938,
36
+ "learning_rate": 1.5510204081632654e-06,
37
+ "loss": 0.4191,
38
+ "step": 40
39
+ },
40
+ {
41
+ "epoch": 0.006117704637220115,
42
+ "grad_norm": 0.26681044697761536,
43
+ "learning_rate": 1.9591836734693877e-06,
44
+ "loss": 0.4289,
45
+ "step": 50
46
+ },
47
+ {
48
+ "epoch": 0.007341245564664138,
49
+ "grad_norm": 0.30233338475227356,
50
+ "learning_rate": 2.3673469387755103e-06,
51
+ "loss": 0.4083,
52
+ "step": 60
53
+ },
54
+ {
55
+ "epoch": 0.008564786492108161,
56
+ "grad_norm": 0.31031334400177,
57
+ "learning_rate": 2.775510204081633e-06,
58
+ "loss": 0.404,
59
+ "step": 70
60
+ },
61
+ {
62
+ "epoch": 0.009788327419552184,
63
+ "grad_norm": 0.2552475035190582,
64
+ "learning_rate": 3.183673469387755e-06,
65
+ "loss": 0.3889,
66
+ "step": 80
67
+ },
68
+ {
69
+ "epoch": 0.011011868346996207,
70
+ "grad_norm": 0.3131052553653717,
71
+ "learning_rate": 3.5918367346938777e-06,
72
+ "loss": 0.3813,
73
+ "step": 90
74
+ },
75
+ {
76
+ "epoch": 0.01223540927444023,
77
+ "grad_norm": 0.3377229571342468,
78
+ "learning_rate": 4.000000000000001e-06,
79
+ "loss": 0.3727,
80
+ "step": 100
81
+ },
82
+ {
83
+ "epoch": 0.01223540927444023,
84
+ "eval_accuracy": 0.6961,
85
+ "eval_f1_weighted": 0.6955947095331424,
86
+ "eval_loss": 0.15026023983955383,
87
+ "eval_macro_f1": 0.6959280508399847,
88
+ "eval_precision": 0.7019396279846728,
89
+ "eval_recall": 0.6961,
90
+ "eval_runtime": 789.7648,
91
+ "eval_samples_per_second": 126.62,
92
+ "eval_steps_per_second": 0.99,
93
+ "step": 100
94
+ },
95
+ {
96
+ "epoch": 0.013458950201884253,
97
+ "grad_norm": 0.3366774618625641,
98
+ "learning_rate": 4.408163265306123e-06,
99
+ "loss": 0.3671,
100
+ "step": 110
101
+ },
102
+ {
103
+ "epoch": 0.014682491129328276,
104
+ "grad_norm": 0.33491575717926025,
105
+ "learning_rate": 4.816326530612245e-06,
106
+ "loss": 0.3498,
107
+ "step": 120
108
+ },
109
+ {
110
+ "epoch": 0.0159060320567723,
111
+ "grad_norm": 0.4119006097316742,
112
+ "learning_rate": 5.2244897959183684e-06,
113
+ "loss": 0.3409,
114
+ "step": 130
115
+ },
116
+ {
117
+ "epoch": 0.017129572984216322,
118
+ "grad_norm": 0.2964725196361542,
119
+ "learning_rate": 5.63265306122449e-06,
120
+ "loss": 0.3433,
121
+ "step": 140
122
+ },
123
+ {
124
+ "epoch": 0.018353113911660345,
125
+ "grad_norm": 0.3075334429740906,
126
+ "learning_rate": 6.040816326530613e-06,
127
+ "loss": 0.3298,
128
+ "step": 150
129
+ },
130
+ {
131
+ "epoch": 0.019576654839104368,
132
+ "grad_norm": 0.33718687295913696,
133
+ "learning_rate": 6.4489795918367345e-06,
134
+ "loss": 0.3297,
135
+ "step": 160
136
+ },
137
+ {
138
+ "epoch": 0.02080019576654839,
139
+ "grad_norm": 0.2809618413448334,
140
+ "learning_rate": 6.857142857142858e-06,
141
+ "loss": 0.3275,
142
+ "step": 170
143
+ },
144
+ {
145
+ "epoch": 0.022023736693992414,
146
+ "grad_norm": 0.37805014848709106,
147
+ "learning_rate": 7.2653061224489805e-06,
148
+ "loss": 0.3201,
149
+ "step": 180
150
+ },
151
+ {
152
+ "epoch": 0.023247277621436437,
153
+ "grad_norm": 0.2940429151058197,
154
+ "learning_rate": 7.673469387755102e-06,
155
+ "loss": 0.3042,
156
+ "step": 190
157
+ },
158
+ {
159
+ "epoch": 0.02447081854888046,
160
+ "grad_norm": 0.3286060690879822,
161
+ "learning_rate": 8.081632653061225e-06,
162
+ "loss": 0.3071,
163
+ "step": 200
164
+ },
165
+ {
166
+ "epoch": 0.02447081854888046,
167
+ "eval_accuracy": 0.77807,
168
+ "eval_f1_weighted": 0.7781129395311969,
169
+ "eval_loss": 0.12153773754835129,
170
+ "eval_macro_f1": 0.7780651598452646,
171
+ "eval_precision": 0.7808306361972198,
172
+ "eval_recall": 0.77807,
173
+ "eval_runtime": 792.7157,
174
+ "eval_samples_per_second": 126.149,
175
+ "eval_steps_per_second": 0.986,
176
+ "step": 200
177
+ },
178
+ {
179
+ "epoch": 0.025694359476324483,
180
+ "grad_norm": 0.33857154846191406,
181
+ "learning_rate": 8.489795918367347e-06,
182
+ "loss": 0.2964,
183
+ "step": 210
184
+ },
185
+ {
186
+ "epoch": 0.026917900403768506,
187
+ "grad_norm": 0.2752092182636261,
188
+ "learning_rate": 8.89795918367347e-06,
189
+ "loss": 0.3,
190
+ "step": 220
191
+ },
192
+ {
193
+ "epoch": 0.02814144133121253,
194
+ "grad_norm": 0.4952544569969177,
195
+ "learning_rate": 9.306122448979593e-06,
196
+ "loss": 0.294,
197
+ "step": 230
198
+ },
199
+ {
200
+ "epoch": 0.029364982258656552,
201
+ "grad_norm": 0.37926042079925537,
202
+ "learning_rate": 9.714285714285715e-06,
203
+ "loss": 0.2964,
204
+ "step": 240
205
+ },
206
+ {
207
+ "epoch": 0.030588523186100575,
208
+ "grad_norm": 0.3222455680370331,
209
+ "learning_rate": 1.0122448979591836e-05,
210
+ "loss": 0.283,
211
+ "step": 250
212
+ },
213
+ {
214
+ "epoch": 0.0318120641135446,
215
+ "grad_norm": 0.2930658459663391,
216
+ "learning_rate": 1.053061224489796e-05,
217
+ "loss": 0.2793,
218
+ "step": 260
219
+ },
220
+ {
221
+ "epoch": 0.03303560504098862,
222
+ "grad_norm": 0.4313414692878723,
223
+ "learning_rate": 1.0938775510204083e-05,
224
+ "loss": 0.2728,
225
+ "step": 270
226
+ },
227
+ {
228
+ "epoch": 0.034259145968432644,
229
+ "grad_norm": 0.37351569533348083,
230
+ "learning_rate": 1.1346938775510206e-05,
231
+ "loss": 0.2752,
232
+ "step": 280
233
+ },
234
+ {
235
+ "epoch": 0.035482686895876664,
236
+ "grad_norm": 0.22772285342216492,
237
+ "learning_rate": 1.1755102040816328e-05,
238
+ "loss": 0.274,
239
+ "step": 290
240
+ },
241
+ {
242
+ "epoch": 0.03670622782332069,
243
+ "grad_norm": 0.15888753533363342,
244
+ "learning_rate": 1.2163265306122449e-05,
245
+ "loss": 0.264,
246
+ "step": 300
247
+ },
248
+ {
249
+ "epoch": 0.03670622782332069,
250
+ "eval_accuracy": 0.82326,
251
+ "eval_f1_weighted": 0.8227034206790359,
252
+ "eval_loss": 0.10234300792217255,
253
+ "eval_macro_f1": 0.8230100448030374,
254
+ "eval_precision": 0.8338333810964687,
255
+ "eval_recall": 0.82326,
256
+ "eval_runtime": 789.0168,
257
+ "eval_samples_per_second": 126.74,
258
+ "eval_steps_per_second": 0.991,
259
+ "step": 300
260
+ },
261
+ {
262
+ "epoch": 0.03792976875076471,
263
+ "grad_norm": 0.23560188710689545,
264
+ "learning_rate": 1.2571428571428572e-05,
265
+ "loss": 0.2753,
266
+ "step": 310
267
+ },
268
+ {
269
+ "epoch": 0.039153309678208736,
270
+ "grad_norm": 0.18699614703655243,
271
+ "learning_rate": 1.2979591836734696e-05,
272
+ "loss": 0.2565,
273
+ "step": 320
274
+ },
275
+ {
276
+ "epoch": 0.040376850605652756,
277
+ "grad_norm": 0.24035991728305817,
278
+ "learning_rate": 1.3387755102040818e-05,
279
+ "loss": 0.2716,
280
+ "step": 330
281
+ },
282
+ {
283
+ "epoch": 0.04160039153309678,
284
+ "grad_norm": 0.15309660136699677,
285
+ "learning_rate": 1.3795918367346941e-05,
286
+ "loss": 0.2515,
287
+ "step": 340
288
+ },
289
+ {
290
+ "epoch": 0.0428239324605408,
291
+ "grad_norm": 0.3382055461406708,
292
+ "learning_rate": 1.4204081632653062e-05,
293
+ "loss": 0.2597,
294
+ "step": 350
295
+ },
296
+ {
297
+ "epoch": 0.04404747338798483,
298
+ "grad_norm": 0.24439455568790436,
299
+ "learning_rate": 1.4612244897959185e-05,
300
+ "loss": 0.247,
301
+ "step": 360
302
+ },
303
+ {
304
+ "epoch": 0.04527101431542885,
305
+ "grad_norm": 0.26233720779418945,
306
+ "learning_rate": 1.5020408163265307e-05,
307
+ "loss": 0.2431,
308
+ "step": 370
309
+ },
310
+ {
311
+ "epoch": 0.046494555242872875,
312
+ "grad_norm": 0.20477451384067535,
313
+ "learning_rate": 1.5387755102040818e-05,
314
+ "loss": 0.2538,
315
+ "step": 380
316
+ },
317
+ {
318
+ "epoch": 0.047718096170316894,
319
+ "grad_norm": 0.2765439748764038,
320
+ "learning_rate": 1.5795918367346942e-05,
321
+ "loss": 0.2404,
322
+ "step": 390
323
+ },
324
+ {
325
+ "epoch": 0.04894163709776092,
326
+ "grad_norm": 0.13196279108524323,
327
+ "learning_rate": 1.6204081632653063e-05,
328
+ "loss": 0.2393,
329
+ "step": 400
330
+ },
331
+ {
332
+ "epoch": 0.04894163709776092,
333
+ "eval_accuracy": 0.84563,
334
+ "eval_f1_weighted": 0.8448128964509274,
335
+ "eval_loss": 0.09028046578168869,
336
+ "eval_macro_f1": 0.8451923512067675,
337
+ "eval_precision": 0.861024732271907,
338
+ "eval_recall": 0.84563,
339
+ "eval_runtime": 787.6067,
340
+ "eval_samples_per_second": 126.967,
341
+ "eval_steps_per_second": 0.993,
342
+ "step": 400
343
+ },
344
+ {
345
+ "epoch": 0.05016517802520494,
346
+ "grad_norm": 0.0924634337425232,
347
+ "learning_rate": 1.6612244897959184e-05,
348
+ "loss": 0.2331,
349
+ "step": 410
350
+ },
351
+ {
352
+ "epoch": 0.05138871895264897,
353
+ "grad_norm": 0.277784138917923,
354
+ "learning_rate": 1.7020408163265308e-05,
355
+ "loss": 0.2301,
356
+ "step": 420
357
+ },
358
+ {
359
+ "epoch": 0.052612259880092986,
360
+ "grad_norm": 0.32299959659576416,
361
+ "learning_rate": 1.742857142857143e-05,
362
+ "loss": 0.2218,
363
+ "step": 430
364
+ },
365
+ {
366
+ "epoch": 0.05383580080753701,
367
+ "grad_norm": 0.18170586228370667,
368
+ "learning_rate": 1.7836734693877553e-05,
369
+ "loss": 0.2318,
370
+ "step": 440
371
+ },
372
+ {
373
+ "epoch": 0.05505934173498103,
374
+ "grad_norm": 0.1520216166973114,
375
+ "learning_rate": 1.8244897959183674e-05,
376
+ "loss": 0.2136,
377
+ "step": 450
378
+ },
379
+ {
380
+ "epoch": 0.05628288266242506,
381
+ "grad_norm": 0.23452645540237427,
382
+ "learning_rate": 1.8653061224489795e-05,
383
+ "loss": 0.2219,
384
+ "step": 460
385
+ },
386
+ {
387
+ "epoch": 0.05750642358986908,
388
+ "grad_norm": 0.10036301612854004,
389
+ "learning_rate": 1.906122448979592e-05,
390
+ "loss": 0.2213,
391
+ "step": 470
392
+ },
393
+ {
394
+ "epoch": 0.058729964517313105,
395
+ "grad_norm": 0.24100220203399658,
396
+ "learning_rate": 1.9469387755102044e-05,
397
+ "loss": 0.2079,
398
+ "step": 480
399
+ },
400
+ {
401
+ "epoch": 0.059953505444757124,
402
+ "grad_norm": 0.28893032670021057,
403
+ "learning_rate": 1.9877551020408165e-05,
404
+ "loss": 0.1978,
405
+ "step": 490
406
+ },
407
+ {
408
+ "epoch": 0.06117704637220115,
409
+ "grad_norm": 0.22918537259101868,
410
+ "learning_rate": 1.9999836143766126e-05,
411
+ "loss": 0.202,
412
+ "step": 500
413
+ },
414
+ {
415
+ "epoch": 0.06117704637220115,
416
+ "eval_accuracy": 0.91628,
417
+ "eval_f1_weighted": 0.9163231793856368,
418
+ "eval_loss": 0.055908624082803726,
419
+ "eval_macro_f1": 0.9162187423616152,
420
+ "eval_precision": 0.9170320825159874,
421
+ "eval_recall": 0.91628,
422
+ "eval_runtime": 760.1138,
423
+ "eval_samples_per_second": 131.559,
424
+ "eval_steps_per_second": 1.029,
425
+ "step": 500
426
+ },
427
+ {
428
+ "epoch": 0.06240058729964517,
429
+ "grad_norm": 0.2930404841899872,
430
+ "learning_rate": 1.999903359554741e-05,
431
+ "loss": 0.1979,
432
+ "step": 510
433
+ },
434
+ {
435
+ "epoch": 0.0636241282270892,
436
+ "grad_norm": 0.0889664739370346,
437
+ "learning_rate": 1.9997562312908153e-05,
438
+ "loss": 0.1953,
439
+ "step": 520
440
+ },
441
+ {
442
+ "epoch": 0.06484766915453322,
443
+ "grad_norm": 0.1687588095664978,
444
+ "learning_rate": 1.9995422394247598e-05,
445
+ "loss": 0.1935,
446
+ "step": 530
447
+ },
448
+ {
449
+ "epoch": 0.06607121008197724,
450
+ "grad_norm": 0.2399725466966629,
451
+ "learning_rate": 1.9992613982683296e-05,
452
+ "loss": 0.1818,
453
+ "step": 540
454
+ },
455
+ {
456
+ "epoch": 0.06729475100942127,
457
+ "grad_norm": 0.2016785889863968,
458
+ "learning_rate": 1.9989137266041552e-05,
459
+ "loss": 0.1893,
460
+ "step": 550
461
+ },
462
+ {
463
+ "epoch": 0.06851829193686529,
464
+ "grad_norm": 0.17420713603496552,
465
+ "learning_rate": 1.9984992476844845e-05,
466
+ "loss": 0.1895,
467
+ "step": 560
468
+ },
469
+ {
470
+ "epoch": 0.06974183286430931,
471
+ "grad_norm": 0.2180609554052353,
472
+ "learning_rate": 1.998017989229629e-05,
473
+ "loss": 0.1836,
474
+ "step": 570
475
+ },
476
+ {
477
+ "epoch": 0.07096537379175333,
478
+ "grad_norm": 0.15563039481639862,
479
+ "learning_rate": 1.9974699834261074e-05,
480
+ "loss": 0.182,
481
+ "step": 580
482
+ },
483
+ {
484
+ "epoch": 0.07218891471919736,
485
+ "grad_norm": 0.25410088896751404,
486
+ "learning_rate": 1.996855266924497e-05,
487
+ "loss": 0.1862,
488
+ "step": 590
489
+ },
490
+ {
491
+ "epoch": 0.07341245564664138,
492
+ "grad_norm": 0.24206259846687317,
493
+ "learning_rate": 1.996173880836979e-05,
494
+ "loss": 0.1818,
495
+ "step": 600
496
+ },
497
+ {
498
+ "epoch": 0.07341245564664138,
499
+ "eval_accuracy": 0.91633,
500
+ "eval_f1_weighted": 0.9162437523349878,
501
+ "eval_loss": 0.05625538155436516,
502
+ "eval_macro_f1": 0.9163072974421806,
503
+ "eval_precision": 0.9233764457540103,
504
+ "eval_recall": 0.91633,
505
+ "eval_runtime": 759.6815,
506
+ "eval_samples_per_second": 131.634,
507
+ "eval_steps_per_second": 1.029,
508
+ "step": 600
509
+ },
510
+ {
511
+ "epoch": 0.0746359965740854,
512
+ "grad_norm": 0.27251410484313965,
513
+ "learning_rate": 1.9954258707345903e-05,
514
+ "loss": 0.1845,
515
+ "step": 610
516
+ },
517
+ {
518
+ "epoch": 0.07585953750152942,
519
+ "grad_norm": 0.1545839011669159,
520
+ "learning_rate": 1.9946112866441757e-05,
521
+ "loss": 0.1861,
522
+ "step": 620
523
+ },
524
+ {
525
+ "epoch": 0.07708307842897345,
526
+ "grad_norm": 0.2535873055458069,
527
+ "learning_rate": 1.9937301830450423e-05,
528
+ "loss": 0.1772,
529
+ "step": 630
530
+ },
531
+ {
532
+ "epoch": 0.07830661935641747,
533
+ "grad_norm": 0.2617948651313782,
534
+ "learning_rate": 1.992782618865315e-05,
535
+ "loss": 0.1711,
536
+ "step": 640
537
+ },
538
+ {
539
+ "epoch": 0.07953016028386149,
540
+ "grad_norm": 0.38676971197128296,
541
+ "learning_rate": 1.9917686574779972e-05,
542
+ "loss": 0.1748,
543
+ "step": 650
544
+ },
545
+ {
546
+ "epoch": 0.08075370121130551,
547
+ "grad_norm": 0.1511957049369812,
548
+ "learning_rate": 1.9906883666967296e-05,
549
+ "loss": 0.1825,
550
+ "step": 660
551
+ },
552
+ {
553
+ "epoch": 0.08197724213874955,
554
+ "grad_norm": 0.2525976002216339,
555
+ "learning_rate": 1.989541818771258e-05,
556
+ "loss": 0.1651,
557
+ "step": 670
558
+ },
559
+ {
560
+ "epoch": 0.08320078306619356,
561
+ "grad_norm": 0.25827622413635254,
562
+ "learning_rate": 1.9883290903825994e-05,
563
+ "loss": 0.1763,
564
+ "step": 680
565
+ },
566
+ {
567
+ "epoch": 0.08442432399363758,
568
+ "grad_norm": 0.24877820909023285,
569
+ "learning_rate": 1.9870502626379127e-05,
570
+ "loss": 0.1652,
571
+ "step": 690
572
+ },
573
+ {
574
+ "epoch": 0.0856478649210816,
575
+ "grad_norm": 0.278772234916687,
576
+ "learning_rate": 1.9857054210650776e-05,
577
+ "loss": 0.1651,
578
+ "step": 700
579
+ },
580
+ {
581
+ "epoch": 0.0856478649210816,
582
+ "eval_accuracy": 0.94905,
583
+ "eval_f1_weighted": 0.9490731118481598,
584
+ "eval_loss": 0.03884471580386162,
585
+ "eval_macro_f1": 0.9489981325237288,
586
+ "eval_precision": 0.9494795831024372,
587
+ "eval_recall": 0.94905,
588
+ "eval_runtime": 759.9337,
589
+ "eval_samples_per_second": 131.59,
590
+ "eval_steps_per_second": 1.029,
591
+ "step": 700
592
+ },
593
+ {
594
+ "epoch": 0.08687140584852564,
595
+ "grad_norm": 0.1895805150270462,
596
+ "learning_rate": 1.9842946556069705e-05,
597
+ "loss": 0.1628,
598
+ "step": 710
599
+ },
600
+ {
601
+ "epoch": 0.08809494677596966,
602
+ "grad_norm": 0.1171497106552124,
603
+ "learning_rate": 1.982818060615452e-05,
604
+ "loss": 0.1715,
605
+ "step": 720
606
+ },
607
+ {
608
+ "epoch": 0.08931848770341368,
609
+ "grad_norm": 0.22032728791236877,
610
+ "learning_rate": 1.9812757348450547e-05,
611
+ "loss": 0.1712,
612
+ "step": 730
613
+ },
614
+ {
615
+ "epoch": 0.0905420286308577,
616
+ "grad_norm": 0.20700488984584808,
617
+ "learning_rate": 1.9796677814463812e-05,
618
+ "loss": 0.1638,
619
+ "step": 740
620
+ },
621
+ {
622
+ "epoch": 0.09176556955830173,
623
+ "grad_norm": 0.24688613414764404,
624
+ "learning_rate": 1.977994307959202e-05,
625
+ "loss": 0.1647,
626
+ "step": 750
627
+ },
628
+ {
629
+ "epoch": 0.09298911048574575,
630
+ "grad_norm": 0.1886432021856308,
631
+ "learning_rate": 1.9762554263052655e-05,
632
+ "loss": 0.1734,
633
+ "step": 760
634
+ },
635
+ {
636
+ "epoch": 0.09421265141318977,
637
+ "grad_norm": 0.08407370001077652,
638
+ "learning_rate": 1.9744512527808127e-05,
639
+ "loss": 0.1755,
640
+ "step": 770
641
+ },
642
+ {
643
+ "epoch": 0.09543619234063379,
644
+ "grad_norm": 0.14160771667957306,
645
+ "learning_rate": 1.9725819080487976e-05,
646
+ "loss": 0.1749,
647
+ "step": 780
648
+ },
649
+ {
650
+ "epoch": 0.09665973326807782,
651
+ "grad_norm": 0.17038081586360931,
652
+ "learning_rate": 1.970647517130819e-05,
653
+ "loss": 0.1795,
654
+ "step": 790
655
+ },
656
+ {
657
+ "epoch": 0.09788327419552184,
658
+ "grad_norm": 0.24339532852172852,
659
+ "learning_rate": 1.968648209398759e-05,
660
+ "loss": 0.161,
661
+ "step": 800
662
+ },
663
+ {
664
+ "epoch": 0.09788327419552184,
665
+ "eval_accuracy": 0.95721,
666
+ "eval_f1_weighted": 0.9572295598206603,
667
+ "eval_loss": 0.03337462991476059,
668
+ "eval_macro_f1": 0.9571669846830866,
669
+ "eval_precision": 0.9576467364395861,
670
+ "eval_recall": 0.95721,
671
+ "eval_runtime": 759.9975,
672
+ "eval_samples_per_second": 131.579,
673
+ "eval_steps_per_second": 1.029,
674
+ "step": 800
675
+ },
676
+ {
677
+ "epoch": 0.09910681512296586,
678
+ "grad_norm": 0.16346058249473572,
679
+ "learning_rate": 1.9665841185661283e-05,
680
+ "loss": 0.1596,
681
+ "step": 810
682
+ },
683
+ {
684
+ "epoch": 0.10033035605040988,
685
+ "grad_norm": 0.25031423568725586,
686
+ "learning_rate": 1.964455382679128e-05,
687
+ "loss": 0.1662,
688
+ "step": 820
689
+ },
690
+ {
691
+ "epoch": 0.10155389697785391,
692
+ "grad_norm": 0.36678382754325867,
693
+ "learning_rate": 1.9622621441074113e-05,
694
+ "loss": 0.1552,
695
+ "step": 830
696
+ },
697
+ {
698
+ "epoch": 0.10277743790529793,
699
+ "grad_norm": 0.34810611605644226,
700
+ "learning_rate": 1.960004549534567e-05,
701
+ "loss": 0.1663,
702
+ "step": 840
703
+ },
704
+ {
705
+ "epoch": 0.10400097883274195,
706
+ "grad_norm": 0.15511688590049744,
707
+ "learning_rate": 1.957682749948307e-05,
708
+ "loss": 0.1571,
709
+ "step": 850
710
+ },
711
+ {
712
+ "epoch": 0.10522451976018597,
713
+ "grad_norm": 0.2082190364599228,
714
+ "learning_rate": 1.9552969006303666e-05,
715
+ "loss": 0.1497,
716
+ "step": 860
717
+ },
718
+ {
719
+ "epoch": 0.10644806068763,
720
+ "grad_norm": 0.14936988055706024,
721
+ "learning_rate": 1.9528471611461235e-05,
722
+ "loss": 0.1435,
723
+ "step": 870
724
+ },
725
+ {
726
+ "epoch": 0.10767160161507403,
727
+ "grad_norm": 0.252530574798584,
728
+ "learning_rate": 1.9503336953339223e-05,
729
+ "loss": 0.1536,
730
+ "step": 880
731
+ },
732
+ {
733
+ "epoch": 0.10889514254251804,
734
+ "grad_norm": 0.07253506034612656,
735
+ "learning_rate": 1.947756671294118e-05,
736
+ "loss": 0.1605,
737
+ "step": 890
738
+ },
739
+ {
740
+ "epoch": 0.11011868346996206,
741
+ "grad_norm": 0.09060605615377426,
742
+ "learning_rate": 1.9451162613778353e-05,
743
+ "loss": 0.1581,
744
+ "step": 900
745
+ },
746
+ {
747
+ "epoch": 0.11011868346996206,
748
+ "eval_accuracy": 0.95459,
749
+ "eval_f1_weighted": 0.9546135918699761,
750
+ "eval_loss": 0.03526490181684494,
751
+ "eval_macro_f1": 0.9545725664001194,
752
+ "eval_precision": 0.9559789658562307,
753
+ "eval_recall": 0.95459,
754
+ "eval_runtime": 759.7202,
755
+ "eval_samples_per_second": 131.627,
756
+ "eval_steps_per_second": 1.029,
757
+ "step": 900
758
+ },
759
+ {
760
+ "epoch": 0.1113422243974061,
761
+ "grad_norm": 0.21080073714256287,
762
+ "learning_rate": 1.942412642175439e-05,
763
+ "loss": 0.164,
764
+ "step": 910
765
+ },
766
+ {
767
+ "epoch": 0.11256576532485012,
768
+ "grad_norm": 0.22832269966602325,
769
+ "learning_rate": 1.939645994504726e-05,
770
+ "loss": 0.1595,
771
+ "step": 920
772
+ },
773
+ {
774
+ "epoch": 0.11378930625229414,
775
+ "grad_norm": 0.19044892489910126,
776
+ "learning_rate": 1.93681650339883e-05,
777
+ "loss": 0.1648,
778
+ "step": 930
779
+ },
780
+ {
781
+ "epoch": 0.11501284717973816,
782
+ "grad_norm": 0.2413518726825714,
783
+ "learning_rate": 1.9339243580938493e-05,
784
+ "loss": 0.155,
785
+ "step": 940
786
+ },
787
+ {
788
+ "epoch": 0.11623638810718219,
789
+ "grad_norm": 0.13218896090984344,
790
+ "learning_rate": 1.930969752016187e-05,
791
+ "loss": 0.1499,
792
+ "step": 950
793
+ },
794
+ {
795
+ "epoch": 0.11745992903462621,
796
+ "grad_norm": 0.23770441114902496,
797
+ "learning_rate": 1.927952882769619e-05,
798
+ "loss": 0.1555,
799
+ "step": 960
800
+ },
801
+ {
802
+ "epoch": 0.11868346996207023,
803
+ "grad_norm": 0.2193710207939148,
804
+ "learning_rate": 1.924873952122074e-05,
805
+ "loss": 0.1505,
806
+ "step": 970
807
+ },
808
+ {
809
+ "epoch": 0.11990701088951425,
810
+ "grad_norm": 0.1577174812555313,
811
+ "learning_rate": 1.9217331659921432e-05,
812
+ "loss": 0.1551,
813
+ "step": 980
814
+ },
815
+ {
816
+ "epoch": 0.12113055181695828,
817
+ "grad_norm": 0.1814967393875122,
818
+ "learning_rate": 1.9185307344353052e-05,
819
+ "loss": 0.1607,
820
+ "step": 990
821
+ },
822
+ {
823
+ "epoch": 0.1223540927444023,
824
+ "grad_norm": 0.27239707112312317,
825
+ "learning_rate": 1.91526687162988e-05,
826
+ "loss": 0.1506,
827
+ "step": 1000
828
+ },
829
+ {
830
+ "epoch": 0.1223540927444023,
831
+ "eval_accuracy": 0.96454,
832
+ "eval_f1_weighted": 0.9645510150201642,
833
+ "eval_loss": 0.028770776465535164,
834
+ "eval_macro_f1": 0.9644889377060355,
835
+ "eval_precision": 0.9646912114769961,
836
+ "eval_recall": 0.96454,
837
+ "eval_runtime": 759.5962,
838
+ "eval_samples_per_second": 131.649,
839
+ "eval_steps_per_second": 1.029,
840
+ "step": 1000
841
+ },
842
+ {
843
+ "epoch": 0.12357763367184632,
844
+ "grad_norm": 0.20511822402477264,
845
+ "learning_rate": 1.9119417958627022e-05,
846
+ "loss": 0.1503,
847
+ "step": 1010
848
+ },
849
+ {
850
+ "epoch": 0.12480117459929034,
851
+ "grad_norm": 0.2324151247739792,
852
+ "learning_rate": 1.9085557295145254e-05,
853
+ "loss": 0.1581,
854
+ "step": 1020
855
+ },
856
+ {
857
+ "epoch": 0.12602471552673436,
858
+ "grad_norm": 0.11019974201917648,
859
+ "learning_rate": 1.905108899045146e-05,
860
+ "loss": 0.1556,
861
+ "step": 1030
862
+ },
863
+ {
864
+ "epoch": 0.1272482564541784,
865
+ "grad_norm": 0.13370558619499207,
866
+ "learning_rate": 1.9016015349782594e-05,
867
+ "loss": 0.1469,
868
+ "step": 1040
869
+ },
870
+ {
871
+ "epoch": 0.12847179738162243,
872
+ "grad_norm": 0.11238931119441986,
873
+ "learning_rate": 1.8980338718860426e-05,
874
+ "loss": 0.155,
875
+ "step": 1050
876
+ },
877
+ {
878
+ "epoch": 0.12969533830906643,
879
+ "grad_norm": 0.2959650158882141,
880
+ "learning_rate": 1.894406148373466e-05,
881
+ "loss": 0.1478,
882
+ "step": 1060
883
+ },
884
+ {
885
+ "epoch": 0.13091887923651047,
886
+ "grad_norm": 0.14401015639305115,
887
+ "learning_rate": 1.890718607062334e-05,
888
+ "loss": 0.1558,
889
+ "step": 1070
890
+ },
891
+ {
892
+ "epoch": 0.13214242016395447,
893
+ "grad_norm": 0.2187667042016983,
894
+ "learning_rate": 1.8869714945750614e-05,
895
+ "loss": 0.1618,
896
+ "step": 1080
897
+ },
898
+ {
899
+ "epoch": 0.1333659610913985,
900
+ "grad_norm": 0.17045481503009796,
901
+ "learning_rate": 1.883165061518176e-05,
902
+ "loss": 0.1487,
903
+ "step": 1090
904
+ },
905
+ {
906
+ "epoch": 0.13458950201884254,
907
+ "grad_norm": 0.09508351981639862,
908
+ "learning_rate": 1.8792995624655605e-05,
909
+ "loss": 0.1461,
910
+ "step": 1100
911
+ },
912
+ {
913
+ "epoch": 0.13458950201884254,
914
+ "eval_accuracy": 0.95198,
915
+ "eval_f1_weighted": 0.9520023284947491,
916
+ "eval_loss": 0.04088817164301872,
917
+ "eval_macro_f1": 0.9519693367124435,
918
+ "eval_precision": 0.9538785423213092,
919
+ "eval_recall": 0.95198,
920
+ "eval_runtime": 759.8551,
921
+ "eval_samples_per_second": 131.604,
922
+ "eval_steps_per_second": 1.029,
923
+ "step": 1100
924
+ },
925
+ {
926
+ "epoch": 0.13581304294628654,
927
+ "grad_norm": 0.1880544275045395,
928
+ "learning_rate": 1.875375255941426e-05,
929
+ "loss": 0.1442,
930
+ "step": 1110
931
+ },
932
+ {
933
+ "epoch": 0.13703658387373058,
934
+ "grad_norm": 0.1614997684955597,
935
+ "learning_rate": 1.8713924044030212e-05,
936
+ "loss": 0.1622,
937
+ "step": 1120
938
+ },
939
+ {
940
+ "epoch": 0.1382601248011746,
941
+ "grad_norm": 0.10545000433921814,
942
+ "learning_rate": 1.86735127422308e-05,
943
+ "loss": 0.1531,
944
+ "step": 1130
945
+ },
946
+ {
947
+ "epoch": 0.13948366572861862,
948
+ "grad_norm": 0.2203921526670456,
949
+ "learning_rate": 1.8632521356720064e-05,
950
+ "loss": 0.1471,
951
+ "step": 1140
952
+ },
953
+ {
954
+ "epoch": 0.14070720665606265,
955
+ "grad_norm": 0.16150350868701935,
956
+ "learning_rate": 1.8590952628997994e-05,
957
+ "loss": 0.1436,
958
+ "step": 1150
959
+ },
960
+ {
961
+ "epoch": 0.14193074758350666,
962
+ "grad_norm": 0.1466459184885025,
963
+ "learning_rate": 1.8548809339177155e-05,
964
+ "loss": 0.1486,
965
+ "step": 1160
966
+ },
967
+ {
968
+ "epoch": 0.1431542885109507,
969
+ "grad_norm": 0.1858566850423813,
970
+ "learning_rate": 1.8506094305796786e-05,
971
+ "loss": 0.1485,
972
+ "step": 1170
973
+ },
974
+ {
975
+ "epoch": 0.14437782943839472,
976
+ "grad_norm": 0.21629975736141205,
977
+ "learning_rate": 1.846281038563428e-05,
978
+ "loss": 0.1558,
979
+ "step": 1180
980
+ },
981
+ {
982
+ "epoch": 0.14560137036583873,
983
+ "grad_norm": 0.12939126789569855,
984
+ "learning_rate": 1.841896047351412e-05,
985
+ "loss": 0.1388,
986
+ "step": 1190
987
+ },
988
+ {
989
+ "epoch": 0.14682491129328276,
990
+ "grad_norm": 0.12447889149188995,
991
+ "learning_rate": 1.8374547502114277e-05,
992
+ "loss": 0.1683,
993
+ "step": 1200
994
+ },
995
+ {
996
+ "epoch": 0.14682491129328276,
997
+ "eval_accuracy": 0.96627,
998
+ "eval_f1_weighted": 0.9662808443914441,
999
+ "eval_loss": 0.027206970378756523,
1000
+ "eval_macro_f1": 0.9662223734621376,
1001
+ "eval_precision": 0.966433068847901,
1002
+ "eval_recall": 0.96627,
1003
+ "eval_runtime": 661.8465,
1004
+ "eval_samples_per_second": 151.092,
1005
+ "eval_steps_per_second": 1.182,
1006
+ "step": 1200
1007
+ },
1008
+ {
1009
+ "epoch": 0.1480484522207268,
1010
+ "grad_norm": 0.07575482875108719,
1011
+ "learning_rate": 1.832957444177007e-05,
1012
+ "loss": 0.152,
1013
+ "step": 1210
1014
+ },
1015
+ {
1016
+ "epoch": 0.1492719931481708,
1017
+ "grad_norm": 0.22879591584205627,
1018
+ "learning_rate": 1.828404430027553e-05,
1019
+ "loss": 0.15,
1020
+ "step": 1220
1021
+ },
1022
+ {
1023
+ "epoch": 0.15049553407561483,
1024
+ "grad_norm": 0.25837475061416626,
1025
+ "learning_rate": 1.8237960122682207e-05,
1026
+ "loss": 0.1557,
1027
+ "step": 1230
1028
+ },
1029
+ {
1030
+ "epoch": 0.15171907500305884,
1031
+ "grad_norm": 0.1440538763999939,
1032
+ "learning_rate": 1.8191324991095527e-05,
1033
+ "loss": 0.154,
1034
+ "step": 1240
1035
+ },
1036
+ {
1037
+ "epoch": 0.15294261593050287,
1038
+ "grad_norm": 0.21304763853549957,
1039
+ "learning_rate": 1.814414202446869e-05,
1040
+ "loss": 0.1428,
1041
+ "step": 1250
1042
+ },
1043
+ {
1044
+ "epoch": 0.1541661568579469,
1045
+ "grad_norm": 0.2292693555355072,
1046
+ "learning_rate": 1.809641437839403e-05,
1047
+ "loss": 0.1441,
1048
+ "step": 1260
1049
+ },
1050
+ {
1051
+ "epoch": 0.1553896977853909,
1052
+ "grad_norm": 0.0891769602894783,
1053
+ "learning_rate": 1.8048145244892e-05,
1054
+ "loss": 0.1563,
1055
+ "step": 1270
1056
+ },
1057
+ {
1058
+ "epoch": 0.15661323871283495,
1059
+ "grad_norm": 0.13292840123176575,
1060
+ "learning_rate": 1.799933785219769e-05,
1061
+ "loss": 0.1418,
1062
+ "step": 1280
1063
+ },
1064
+ {
1065
+ "epoch": 0.15783677964027898,
1066
+ "grad_norm": 0.30031099915504456,
1067
+ "learning_rate": 1.7949995464544892e-05,
1068
+ "loss": 0.1543,
1069
+ "step": 1290
1070
+ },
1071
+ {
1072
+ "epoch": 0.15906032056772298,
1073
+ "grad_norm": 0.08812980353832245,
1074
+ "learning_rate": 1.790012138194783e-05,
1075
+ "loss": 0.1473,
1076
+ "step": 1300
1077
+ },
1078
+ {
1079
+ "epoch": 0.15906032056772298,
1080
+ "eval_accuracy": 0.96636,
1081
+ "eval_f1_weighted": 0.9663763998373325,
1082
+ "eval_loss": 0.02714269980788231,
1083
+ "eval_macro_f1": 0.9663303002945869,
1084
+ "eval_precision": 0.9669126928794434,
1085
+ "eval_recall": 0.96636,
1086
+ "eval_runtime": 661.9303,
1087
+ "eval_samples_per_second": 151.073,
1088
+ "eval_steps_per_second": 1.181,
1089
+ "step": 1300
1090
+ },
1091
+ {
1092
+ "epoch": 0.16028386149516702,
1093
+ "grad_norm": 0.1380453258752823,
1094
+ "learning_rate": 1.7849718939980426e-05,
1095
+ "loss": 0.1594,
1096
+ "step": 1310
1097
+ },
1098
+ {
1099
+ "epoch": 0.16150740242261102,
1100
+ "grad_norm": 0.08388533443212509,
1101
+ "learning_rate": 1.7798791509553224e-05,
1102
+ "loss": 0.1411,
1103
+ "step": 1320
1104
+ },
1105
+ {
1106
+ "epoch": 0.16273094335005506,
1107
+ "grad_norm": 0.2659275531768799,
1108
+ "learning_rate": 1.7747342496687957e-05,
1109
+ "loss": 0.1495,
1110
+ "step": 1330
1111
+ },
1112
+ {
1113
+ "epoch": 0.1639544842774991,
1114
+ "grad_norm": 0.275398313999176,
1115
+ "learning_rate": 1.7695375342289732e-05,
1116
+ "loss": 0.134,
1117
+ "step": 1340
1118
+ },
1119
+ {
1120
+ "epoch": 0.1651780252049431,
1121
+ "grad_norm": 0.15272289514541626,
1122
+ "learning_rate": 1.7642893521916926e-05,
1123
+ "loss": 0.1427,
1124
+ "step": 1350
1125
+ },
1126
+ {
1127
+ "epoch": 0.16640156613238713,
1128
+ "grad_norm": 0.21478359401226044,
1129
+ "learning_rate": 1.7589900545548706e-05,
1130
+ "loss": 0.1465,
1131
+ "step": 1360
1132
+ },
1133
+ {
1134
+ "epoch": 0.16762510705983116,
1135
+ "grad_norm": 0.2184278517961502,
1136
+ "learning_rate": 1.7536399957350326e-05,
1137
+ "loss": 0.1452,
1138
+ "step": 1370
1139
+ },
1140
+ {
1141
+ "epoch": 0.16884864798727517,
1142
+ "grad_norm": 0.23272652924060822,
1143
+ "learning_rate": 1.7482395335436052e-05,
1144
+ "loss": 0.142,
1145
+ "step": 1380
1146
+ },
1147
+ {
1148
+ "epoch": 0.1700721889147192,
1149
+ "grad_norm": 0.14892522990703583,
1150
+ "learning_rate": 1.7427890291629895e-05,
1151
+ "loss": 0.1465,
1152
+ "step": 1390
1153
+ },
1154
+ {
1155
+ "epoch": 0.1712957298421632,
1156
+ "grad_norm": 0.2066282033920288,
1157
+ "learning_rate": 1.737288847122402e-05,
1158
+ "loss": 0.1444,
1159
+ "step": 1400
1160
+ },
1161
+ {
1162
+ "epoch": 0.1712957298421632,
1163
+ "eval_accuracy": 0.95302,
1164
+ "eval_f1_weighted": 0.9530351521706546,
1165
+ "eval_loss": 0.03864819556474686,
1166
+ "eval_macro_f1": 0.9530165251021966,
1167
+ "eval_precision": 0.9557329367218795,
1168
+ "eval_recall": 0.95302,
1169
+ "eval_runtime": 661.7384,
1170
+ "eval_samples_per_second": 151.117,
1171
+ "eval_steps_per_second": 1.182,
1172
+ "step": 1400
1173
+ },
1174
+ {
1175
+ "epoch": 0.17251927076960724,
1176
+ "grad_norm": 0.2941270172595978,
1177
+ "learning_rate": 1.7317393552734977e-05,
1178
+ "loss": 0.136,
1179
+ "step": 1410
1180
+ },
1181
+ {
1182
+ "epoch": 0.17374281169705127,
1183
+ "grad_norm": 0.11988822370767593,
1184
+ "learning_rate": 1.7261409247657656e-05,
1185
+ "loss": 0.1551,
1186
+ "step": 1420
1187
+ },
1188
+ {
1189
+ "epoch": 0.17496635262449528,
1190
+ "grad_norm": 0.346236914396286,
1191
+ "learning_rate": 1.7204939300217084e-05,
1192
+ "loss": 0.1453,
1193
+ "step": 1430
1194
+ },
1195
+ {
1196
+ "epoch": 0.17618989355193931,
1197
+ "grad_norm": 0.15973429381847382,
1198
+ "learning_rate": 1.7147987487118006e-05,
1199
+ "loss": 0.1341,
1200
+ "step": 1440
1201
+ },
1202
+ {
1203
+ "epoch": 0.17741343447938335,
1204
+ "grad_norm": 0.2531983554363251,
1205
+ "learning_rate": 1.7090557617292287e-05,
1206
+ "loss": 0.1479,
1207
+ "step": 1450
1208
+ },
1209
+ {
1210
+ "epoch": 0.17863697540682735,
1211
+ "grad_norm": 0.10105066746473312,
1212
+ "learning_rate": 1.703265353164419e-05,
1213
+ "loss": 0.1459,
1214
+ "step": 1460
1215
+ },
1216
+ {
1217
+ "epoch": 0.1798605163342714,
1218
+ "grad_norm": 0.10651528090238571,
1219
+ "learning_rate": 1.6974279102793473e-05,
1220
+ "loss": 0.152,
1221
+ "step": 1470
1222
+ },
1223
+ {
1224
+ "epoch": 0.1810840572617154,
1225
+ "grad_norm": 0.2675159275531769,
1226
+ "learning_rate": 1.6915438234816423e-05,
1227
+ "loss": 0.1467,
1228
+ "step": 1480
1229
+ },
1230
+ {
1231
+ "epoch": 0.18230759818915943,
1232
+ "grad_norm": 0.09507659822702408,
1233
+ "learning_rate": 1.6856134862984718e-05,
1234
+ "loss": 0.1504,
1235
+ "step": 1490
1236
+ },
1237
+ {
1238
+ "epoch": 0.18353113911660346,
1239
+ "grad_norm": 0.13831426203250885,
1240
+ "learning_rate": 1.6796372953502252e-05,
1241
+ "loss": 0.151,
1242
+ "step": 1500
1243
+ },
1244
+ {
1245
+ "epoch": 0.18353113911660346,
1246
+ "eval_accuracy": 0.96716,
1247
+ "eval_f1_weighted": 0.9671757267424962,
1248
+ "eval_loss": 0.02643330581486225,
1249
+ "eval_macro_f1": 0.9671297823300075,
1250
+ "eval_precision": 0.9676731949933907,
1251
+ "eval_recall": 0.96716,
1252
+ "eval_runtime": 662.2285,
1253
+ "eval_samples_per_second": 151.005,
1254
+ "eval_steps_per_second": 1.181,
1255
+ "step": 1500
1256
+ },
1257
+ {
1258
+ "epoch": 0.18475468004404746,
1259
+ "grad_norm": 0.20989136397838593,
1260
+ "learning_rate": 1.673615650323988e-05,
1261
+ "loss": 0.1424,
1262
+ "step": 1510
1263
+ },
1264
+ {
1265
+ "epoch": 0.1859782209714915,
1266
+ "grad_norm": 0.13308484852313995,
1267
+ "learning_rate": 1.6675489539468094e-05,
1268
+ "loss": 0.1541,
1269
+ "step": 1520
1270
+ },
1271
+ {
1272
+ "epoch": 0.18720176189893553,
1273
+ "grad_norm": 0.3802664577960968,
1274
+ "learning_rate": 1.6614376119587685e-05,
1275
+ "loss": 0.1426,
1276
+ "step": 1530
1277
+ },
1278
+ {
1279
+ "epoch": 0.18842530282637954,
1280
+ "grad_norm": 0.254869282245636,
1281
+ "learning_rate": 1.6552820330858398e-05,
1282
+ "loss": 0.1407,
1283
+ "step": 1540
1284
+ },
1285
+ {
1286
+ "epoch": 0.18964884375382357,
1287
+ "grad_norm": 0.22702300548553467,
1288
+ "learning_rate": 1.6490826290125555e-05,
1289
+ "loss": 0.1404,
1290
+ "step": 1550
1291
+ },
1292
+ {
1293
+ "epoch": 0.19087238468126758,
1294
+ "grad_norm": 0.18147043883800507,
1295
+ "learning_rate": 1.6428398143544727e-05,
1296
+ "loss": 0.1398,
1297
+ "step": 1560
1298
+ },
1299
+ {
1300
+ "epoch": 0.1920959256087116,
1301
+ "grad_norm": 0.2697111964225769,
1302
+ "learning_rate": 1.636554006630446e-05,
1303
+ "loss": 0.1465,
1304
+ "step": 1570
1305
+ },
1306
+ {
1307
+ "epoch": 0.19331946653615564,
1308
+ "grad_norm": 0.22285781800746918,
1309
+ "learning_rate": 1.6302256262347e-05,
1310
+ "loss": 0.1398,
1311
+ "step": 1580
1312
+ },
1313
+ {
1314
+ "epoch": 0.19454300746359965,
1315
+ "grad_norm": 0.2551407814025879,
1316
+ "learning_rate": 1.6238550964087175e-05,
1317
+ "loss": 0.1347,
1318
+ "step": 1590
1319
+ },
1320
+ {
1321
+ "epoch": 0.19576654839104368,
1322
+ "grad_norm": 0.25052276253700256,
1323
+ "learning_rate": 1.61744284321293e-05,
1324
+ "loss": 0.1461,
1325
+ "step": 1600
1326
+ },
1327
+ {
1328
+ "epoch": 0.19576654839104368,
1329
+ "eval_accuracy": 0.96526,
1330
+ "eval_f1_weighted": 0.9652784309636304,
1331
+ "eval_loss": 0.02899531088769436,
1332
+ "eval_macro_f1": 0.9652431164406312,
1333
+ "eval_precision": 0.9664143950990177,
1334
+ "eval_recall": 0.96526,
1335
+ "eval_runtime": 662.3587,
1336
+ "eval_samples_per_second": 150.976,
1337
+ "eval_steps_per_second": 1.181,
1338
+ "step": 1600
1339
+ },
1340
+ {
1341
+ "epoch": 0.19699008931848772,
1342
+ "grad_norm": 0.09207922220230103,
1343
+ "learning_rate": 1.6109892954982242e-05,
1344
+ "loss": 0.1382,
1345
+ "step": 1610
1346
+ },
1347
+ {
1348
+ "epoch": 0.19821363024593172,
1349
+ "grad_norm": 0.2670917809009552,
1350
+ "learning_rate": 1.6044948848772605e-05,
1351
+ "loss": 0.142,
1352
+ "step": 1620
1353
+ },
1354
+ {
1355
+ "epoch": 0.19943717117337575,
1356
+ "grad_norm": 0.28024110198020935,
1357
+ "learning_rate": 1.597960045695607e-05,
1358
+ "loss": 0.1385,
1359
+ "step": 1630
1360
+ },
1361
+ {
1362
+ "epoch": 0.20066071210081976,
1363
+ "grad_norm": 0.15802879631519318,
1364
+ "learning_rate": 1.5913852150026895e-05,
1365
+ "loss": 0.1328,
1366
+ "step": 1640
1367
+ },
1368
+ {
1369
+ "epoch": 0.2018842530282638,
1370
+ "grad_norm": 0.12178096175193787,
1371
+ "learning_rate": 1.5847708325225618e-05,
1372
+ "loss": 0.1446,
1373
+ "step": 1650
1374
+ },
1375
+ {
1376
+ "epoch": 0.20310779395570783,
1377
+ "grad_norm": 0.22587499022483826,
1378
+ "learning_rate": 1.578117340624499e-05,
1379
+ "loss": 0.1406,
1380
+ "step": 1660
1381
+ },
1382
+ {
1383
+ "epoch": 0.20433133488315183,
1384
+ "grad_norm": 0.18149550259113312,
1385
+ "learning_rate": 1.5714251842934096e-05,
1386
+ "loss": 0.1459,
1387
+ "step": 1670
1388
+ },
1389
+ {
1390
+ "epoch": 0.20555487581059587,
1391
+ "grad_norm": 0.15662655234336853,
1392
+ "learning_rate": 1.5646948111000756e-05,
1393
+ "loss": 0.1344,
1394
+ "step": 1680
1395
+ },
1396
+ {
1397
+ "epoch": 0.2067784167380399,
1398
+ "grad_norm": 0.1990954577922821,
1399
+ "learning_rate": 1.5579266711712185e-05,
1400
+ "loss": 0.1384,
1401
+ "step": 1690
1402
+ },
1403
+ {
1404
+ "epoch": 0.2080019576654839,
1405
+ "grad_norm": 0.14277924597263336,
1406
+ "learning_rate": 1.551121217159397e-05,
1407
+ "loss": 0.1251,
1408
+ "step": 1700
1409
+ },
1410
+ {
1411
+ "epoch": 0.2080019576654839,
1412
+ "eval_accuracy": 0.97199,
1413
+ "eval_f1_weighted": 0.9719965673787706,
1414
+ "eval_loss": 0.025022055953741074,
1415
+ "eval_macro_f1": 0.9719444578845672,
1416
+ "eval_precision": 0.9720688283360513,
1417
+ "eval_recall": 0.97199,
1418
+ "eval_runtime": 661.2104,
1419
+ "eval_samples_per_second": 151.238,
1420
+ "eval_steps_per_second": 1.183,
1421
+ "step": 1700
1422
+ },
1423
+ {
1424
+ "epoch": 0.20922549859292794,
1425
+ "grad_norm": 0.13305141031742096,
1426
+ "learning_rate": 1.5442789042127306e-05,
1427
+ "loss": 0.1465,
1428
+ "step": 1710
1429
+ },
1430
+ {
1431
+ "epoch": 0.21044903952037194,
1432
+ "grad_norm": 0.2761312425136566,
1433
+ "learning_rate": 1.5374001899444624e-05,
1434
+ "loss": 0.1413,
1435
+ "step": 1720
1436
+ },
1437
+ {
1438
+ "epoch": 0.21167258044781598,
1439
+ "grad_norm": 0.11722280085086823,
1440
+ "learning_rate": 1.530485534402351e-05,
1441
+ "loss": 0.1337,
1442
+ "step": 1730
1443
+ },
1444
+ {
1445
+ "epoch": 0.21289612137526,
1446
+ "grad_norm": 0.0893188863992691,
1447
+ "learning_rate": 1.523535400037905e-05,
1448
+ "loss": 0.1458,
1449
+ "step": 1740
1450
+ },
1451
+ {
1452
+ "epoch": 0.21411966230270402,
1453
+ "grad_norm": 0.14586731791496277,
1454
+ "learning_rate": 1.5165502516754526e-05,
1455
+ "loss": 0.1389,
1456
+ "step": 1750
1457
+ },
1458
+ {
1459
+ "epoch": 0.21534320323014805,
1460
+ "grad_norm": 0.23379822075366974,
1461
+ "learning_rate": 1.509530556481056e-05,
1462
+ "loss": 0.1467,
1463
+ "step": 1760
1464
+ },
1465
+ {
1466
+ "epoch": 0.21656674415759208,
1467
+ "grad_norm": 0.12035554647445679,
1468
+ "learning_rate": 1.5024767839312642e-05,
1469
+ "loss": 0.1386,
1470
+ "step": 1770
1471
+ },
1472
+ {
1473
+ "epoch": 0.2177902850850361,
1474
+ "grad_norm": 0.1451006382703781,
1475
+ "learning_rate": 1.495389405781719e-05,
1476
+ "loss": 0.1435,
1477
+ "step": 1780
1478
+ },
1479
+ {
1480
+ "epoch": 0.21901382601248012,
1481
+ "grad_norm": 0.08014333248138428,
1482
+ "learning_rate": 1.4882688960355994e-05,
1483
+ "loss": 0.1491,
1484
+ "step": 1790
1485
+ },
1486
+ {
1487
+ "epoch": 0.22023736693992413,
1488
+ "grad_norm": 0.17279869318008423,
1489
+ "learning_rate": 1.4811157309119235e-05,
1490
+ "loss": 0.151,
1491
+ "step": 1800
1492
+ },
1493
+ {
1494
+ "epoch": 0.22023736693992413,
1495
+ "eval_accuracy": 0.96674,
1496
+ "eval_f1_weighted": 0.9667576374640046,
1497
+ "eval_loss": 0.02695448324084282,
1498
+ "eval_macro_f1": 0.9667205101995933,
1499
+ "eval_precision": 0.9677025757044254,
1500
+ "eval_recall": 0.96674,
1501
+ "eval_runtime": 661.5931,
1502
+ "eval_samples_per_second": 151.15,
1503
+ "eval_steps_per_second": 1.182,
1504
+ "step": 1800
1505
+ },
1506
+ {
1507
+ "epoch": 0.22146090786736816,
1508
+ "grad_norm": 0.32028236985206604,
1509
+ "learning_rate": 1.473930388813698e-05,
1510
+ "loss": 0.1342,
1511
+ "step": 1810
1512
+ },
1513
+ {
1514
+ "epoch": 0.2226844487948122,
1515
+ "grad_norm": 0.2838447690010071,
1516
+ "learning_rate": 1.466713350295922e-05,
1517
+ "loss": 0.1346,
1518
+ "step": 1820
1519
+ },
1520
+ {
1521
+ "epoch": 0.2239079897222562,
1522
+ "grad_norm": 0.2165336161851883,
1523
+ "learning_rate": 1.4594650980334492e-05,
1524
+ "loss": 0.1447,
1525
+ "step": 1830
1526
+ },
1527
+ {
1528
+ "epoch": 0.22513153064970023,
1529
+ "grad_norm": 0.1966007500886917,
1530
+ "learning_rate": 1.4521861167887043e-05,
1531
+ "loss": 0.1405,
1532
+ "step": 1840
1533
+ },
1534
+ {
1535
+ "epoch": 0.22635507157714427,
1536
+ "grad_norm": 0.16725124418735504,
1537
+ "learning_rate": 1.4448768933792644e-05,
1538
+ "loss": 0.1272,
1539
+ "step": 1850
1540
+ },
1541
+ {
1542
+ "epoch": 0.22757861250458827,
1543
+ "grad_norm": 0.22608086466789246,
1544
+ "learning_rate": 1.4375379166452991e-05,
1545
+ "loss": 0.1305,
1546
+ "step": 1860
1547
+ },
1548
+ {
1549
+ "epoch": 0.2288021534320323,
1550
+ "grad_norm": 0.24929635226726532,
1551
+ "learning_rate": 1.4301696774168782e-05,
1552
+ "loss": 0.1276,
1553
+ "step": 1870
1554
+ },
1555
+ {
1556
+ "epoch": 0.2300256943594763,
1557
+ "grad_norm": 0.0894479975104332,
1558
+ "learning_rate": 1.4227726684811439e-05,
1559
+ "loss": 0.1515,
1560
+ "step": 1880
1561
+ },
1562
+ {
1563
+ "epoch": 0.23124923528692035,
1564
+ "grad_norm": 0.31293338537216187,
1565
+ "learning_rate": 1.4153473845493531e-05,
1566
+ "loss": 0.1262,
1567
+ "step": 1890
1568
+ },
1569
+ {
1570
+ "epoch": 0.23247277621436438,
1571
+ "grad_norm": 0.34327465295791626,
1572
+ "learning_rate": 1.4078943222237933e-05,
1573
+ "loss": 0.1371,
1574
+ "step": 1900
1575
+ },
1576
+ {
1577
+ "epoch": 0.23247277621436438,
1578
+ "eval_accuracy": 0.97149,
1579
+ "eval_f1_weighted": 0.9715031773882226,
1580
+ "eval_loss": 0.02571570873260498,
1581
+ "eval_macro_f1": 0.9714618804747581,
1582
+ "eval_precision": 0.9719368831197501,
1583
+ "eval_recall": 0.97149,
1584
+ "eval_runtime": 660.9647,
1585
+ "eval_samples_per_second": 151.294,
1586
+ "eval_steps_per_second": 1.183,
1587
+ "step": 1900
1588
+ },
1589
+ {
1590
+ "epoch": 0.23369631714180839,
1591
+ "grad_norm": 0.07504434138536453,
1592
+ "learning_rate": 1.4004139799645669e-05,
1593
+ "loss": 0.1506,
1594
+ "step": 1910
1595
+ },
1596
+ {
1597
+ "epoch": 0.23491985806925242,
1598
+ "grad_norm": 0.2954120934009552,
1599
+ "learning_rate": 1.3929068580562561e-05,
1600
+ "loss": 0.1329,
1601
+ "step": 1920
1602
+ },
1603
+ {
1604
+ "epoch": 0.23614339899669645,
1605
+ "grad_norm": 0.17359323799610138,
1606
+ "learning_rate": 1.3853734585744642e-05,
1607
+ "loss": 0.1347,
1608
+ "step": 1930
1609
+ },
1610
+ {
1611
+ "epoch": 0.23736693992414046,
1612
+ "grad_norm": 0.15386463701725006,
1613
+ "learning_rate": 1.3778142853522352e-05,
1614
+ "loss": 0.1429,
1615
+ "step": 1940
1616
+ },
1617
+ {
1618
+ "epoch": 0.2385904808515845,
1619
+ "grad_norm": 0.2623615264892578,
1620
+ "learning_rate": 1.3702298439463589e-05,
1621
+ "loss": 0.1296,
1622
+ "step": 1950
1623
+ },
1624
+ {
1625
+ "epoch": 0.2398140217790285,
1626
+ "grad_norm": 0.12422071397304535,
1627
+ "learning_rate": 1.3626206416035589e-05,
1628
+ "loss": 0.1441,
1629
+ "step": 1960
1630
+ },
1631
+ {
1632
+ "epoch": 0.24103756270647253,
1633
+ "grad_norm": 0.14957550168037415,
1634
+ "learning_rate": 1.3549871872265672e-05,
1635
+ "loss": 0.1406,
1636
+ "step": 1970
1637
+ },
1638
+ {
1639
+ "epoch": 0.24226110363391656,
1640
+ "grad_norm": 0.11441356688737869,
1641
+ "learning_rate": 1.3473299913400905e-05,
1642
+ "loss": 0.1294,
1643
+ "step": 1980
1644
+ },
1645
+ {
1646
+ "epoch": 0.24348464456136057,
1647
+ "grad_norm": 0.19187535345554352,
1648
+ "learning_rate": 1.3396495660566651e-05,
1649
+ "loss": 0.1427,
1650
+ "step": 1990
1651
+ },
1652
+ {
1653
+ "epoch": 0.2447081854888046,
1654
+ "grad_norm": 0.10940376669168472,
1655
+ "learning_rate": 1.3319464250424068e-05,
1656
+ "loss": 0.1439,
1657
+ "step": 2000
1658
+ },
1659
+ {
1660
+ "epoch": 0.2447081854888046,
1661
+ "eval_accuracy": 0.97729,
1662
+ "eval_f1_weighted": 0.9772907740157332,
1663
+ "eval_loss": 0.01969166286289692,
1664
+ "eval_macro_f1": 0.9772431978486675,
1665
+ "eval_precision": 0.9772926495694174,
1666
+ "eval_recall": 0.97729,
1667
+ "eval_runtime": 660.9031,
1668
+ "eval_samples_per_second": 151.308,
1669
+ "eval_steps_per_second": 1.183,
1670
+ "step": 2000
1671
+ },
1672
+ {
1673
+ "epoch": 0.24593172641624864,
1674
+ "grad_norm": 0.07293081283569336,
1675
+ "learning_rate": 1.3242210834826575e-05,
1676
+ "loss": 0.1322,
1677
+ "step": 2010
1678
+ },
1679
+ {
1680
+ "epoch": 0.24715526734369264,
1681
+ "grad_norm": 0.1582045555114746,
1682
+ "learning_rate": 1.31647405804753e-05,
1683
+ "loss": 0.1273,
1684
+ "step": 2020
1685
+ },
1686
+ {
1687
+ "epoch": 0.24837880827113668,
1688
+ "grad_norm": 0.29221460223197937,
1689
+ "learning_rate": 1.3087058668573521e-05,
1690
+ "loss": 0.1303,
1691
+ "step": 2030
1692
+ },
1693
+ {
1694
+ "epoch": 0.24960234919858068,
1695
+ "grad_norm": 0.17140211164951324,
1696
+ "learning_rate": 1.3009170294480149e-05,
1697
+ "loss": 0.1394,
1698
+ "step": 2040
1699
+ },
1700
+ {
1701
+ "epoch": 0.25082589012602474,
1702
+ "grad_norm": 0.3539097011089325,
1703
+ "learning_rate": 1.2931080667362276e-05,
1704
+ "loss": 0.1331,
1705
+ "step": 2050
1706
+ },
1707
+ {
1708
+ "epoch": 0.2520494310534687,
1709
+ "grad_norm": 0.23371531069278717,
1710
+ "learning_rate": 1.2852795009846775e-05,
1711
+ "loss": 0.1238,
1712
+ "step": 2060
1713
+ },
1714
+ {
1715
+ "epoch": 0.25327297198091275,
1716
+ "grad_norm": 0.21888205409049988,
1717
+ "learning_rate": 1.2774318557671009e-05,
1718
+ "loss": 0.1509,
1719
+ "step": 2070
1720
+ },
1721
+ {
1722
+ "epoch": 0.2544965129083568,
1723
+ "grad_norm": 0.4899176359176636,
1724
+ "learning_rate": 1.2695656559332671e-05,
1725
+ "loss": 0.1461,
1726
+ "step": 2080
1727
+ },
1728
+ {
1729
+ "epoch": 0.2557200538358008,
1730
+ "grad_norm": 0.1515502780675888,
1731
+ "learning_rate": 1.261681427573877e-05,
1732
+ "loss": 0.1331,
1733
+ "step": 2090
1734
+ },
1735
+ {
1736
+ "epoch": 0.25694359476324485,
1737
+ "grad_norm": 0.13595114648342133,
1738
+ "learning_rate": 1.2537796979853764e-05,
1739
+ "loss": 0.1484,
1740
+ "step": 2100
1741
+ },
1742
+ {
1743
+ "epoch": 0.25694359476324485,
1744
+ "eval_accuracy": 0.97769,
1745
+ "eval_f1_weighted": 0.9776954901619872,
1746
+ "eval_loss": 0.019225435331463814,
1747
+ "eval_macro_f1": 0.9776543362648145,
1748
+ "eval_precision": 0.9777750172865488,
1749
+ "eval_recall": 0.97769,
1750
+ "eval_runtime": 660.729,
1751
+ "eval_samples_per_second": 151.348,
1752
+ "eval_steps_per_second": 1.184,
1753
+ "step": 2100
1754
+ },
1755
+ {
1756
+ "epoch": 0.25816713569068883,
1757
+ "grad_norm": 0.14403671026229858,
1758
+ "learning_rate": 1.2458609956346927e-05,
1759
+ "loss": 0.1386,
1760
+ "step": 2110
1761
+ },
1762
+ {
1763
+ "epoch": 0.25939067661813286,
1764
+ "grad_norm": 0.28055140376091003,
1765
+ "learning_rate": 1.23792585012389e-05,
1766
+ "loss": 0.1331,
1767
+ "step": 2120
1768
+ },
1769
+ {
1770
+ "epoch": 0.2606142175455769,
1771
+ "grad_norm": 0.15281730890274048,
1772
+ "learning_rate": 1.2299747921547485e-05,
1773
+ "loss": 0.1428,
1774
+ "step": 2130
1775
+ },
1776
+ {
1777
+ "epoch": 0.26183775847302093,
1778
+ "grad_norm": 0.2775639593601227,
1779
+ "learning_rate": 1.2220083534932731e-05,
1780
+ "loss": 0.1292,
1781
+ "step": 2140
1782
+ },
1783
+ {
1784
+ "epoch": 0.26306129940046497,
1785
+ "grad_norm": 0.23231762647628784,
1786
+ "learning_rate": 1.2140270669341272e-05,
1787
+ "loss": 0.1278,
1788
+ "step": 2150
1789
+ },
1790
+ {
1791
+ "epoch": 0.26428484032790894,
1792
+ "grad_norm": 0.15349626541137695,
1793
+ "learning_rate": 1.2060314662650008e-05,
1794
+ "loss": 0.1289,
1795
+ "step": 2160
1796
+ },
1797
+ {
1798
+ "epoch": 0.265508381255353,
1799
+ "grad_norm": 0.270783394575119,
1800
+ "learning_rate": 1.1980220862309097e-05,
1801
+ "loss": 0.1353,
1802
+ "step": 2170
1803
+ },
1804
+ {
1805
+ "epoch": 0.266731922182797,
1806
+ "grad_norm": 0.25189876556396484,
1807
+ "learning_rate": 1.1899994624984324e-05,
1808
+ "loss": 0.128,
1809
+ "step": 2180
1810
+ },
1811
+ {
1812
+ "epoch": 0.26795546311024104,
1813
+ "grad_norm": 0.3517930805683136,
1814
+ "learning_rate": 1.1819641316198856e-05,
1815
+ "loss": 0.1368,
1816
+ "step": 2190
1817
+ },
1818
+ {
1819
+ "epoch": 0.2691790040376851,
1820
+ "grad_norm": 0.1853000968694687,
1821
+ "learning_rate": 1.173916630997437e-05,
1822
+ "loss": 0.1261,
1823
+ "step": 2200
1824
+ },
1825
+ {
1826
+ "epoch": 0.2691790040376851,
1827
+ "eval_accuracy": 0.96989,
1828
+ "eval_f1_weighted": 0.9699058894297731,
1829
+ "eval_loss": 0.027551453560590744,
1830
+ "eval_macro_f1": 0.9698711242322156,
1831
+ "eval_precision": 0.97078292839073,
1832
+ "eval_recall": 0.96989,
1833
+ "eval_runtime": 660.4203,
1834
+ "eval_samples_per_second": 151.419,
1835
+ "eval_steps_per_second": 1.184,
1836
+ "step": 2200
1837
+ }
1838
+ ],
1839
+ "logging_steps": 10,
1840
+ "max_steps": 8173,
1841
+ "num_input_tokens_seen": 0,
1842
+ "num_train_epochs": 1,
1843
+ "save_steps": 100,
1844
+ "stateful_callbacks": {
1845
+ "EarlyStoppingCallback": {
1846
+ "args": {
1847
+ "early_stopping_patience": 3,
1848
+ "early_stopping_threshold": 0.0
1849
+ },
1850
+ "attributes": {
1851
+ "early_stopping_patience_counter": 1
1852
+ }
1853
+ },
1854
+ "TrainerControl": {
1855
+ "args": {
1856
+ "should_epoch_stop": false,
1857
+ "should_evaluate": false,
1858
+ "should_log": false,
1859
+ "should_save": true,
1860
+ "should_training_stop": false
1861
+ },
1862
+ "attributes": {}
1863
+ }
1864
+ },
1865
+ "total_flos": 3.734350177304909e+16,
1866
+ "train_batch_size": 64,
1867
+ "trial_name": null,
1868
+ "trial_params": null
1869
+ }
taskA-unixcoder-focal/checkpoint-2200/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:34af5433a3002793dc6233ee444a7f3918508304d82b9638eca4b9f53208c9bf
3
+ size 5841
taskA-unixcoder-focal/checkpoint-2200/vocab.json ADDED
The diff for this file is too large to render. See raw diff