Hugging Face's logo

TuWaveGod
/
pi05baseline-5000steps-batch4

Safetensors
Model card Files
xet
 Community
TuWaveGod commited on
Commit
83a9100
·
verified ·
1 Parent(s): 95f05ba

Upload folder using huggingface_hub

Browse files
Files changed (22) hide show
  1. checkpoints/checkpoint_step1000.safetensors +3 -0
  2. checkpoints/checkpoint_step1500.safetensors +3 -0
  3. checkpoints/checkpoint_step2000.safetensors +3 -0
  4. checkpoints/checkpoint_step2500.safetensors +3 -0
  5. checkpoints/checkpoint_step3000.safetensors +3 -0
  6. checkpoints/checkpoint_step3500.safetensors +3 -0
  7. checkpoints/checkpoint_step4000.safetensors +3 -0
  8. checkpoints/checkpoint_step4500.safetensors +3 -0
  9. checkpoints/checkpoint_step500.safetensors +3 -0
  10. checkpoints/checkpoint_step5000.safetensors +3 -0
  11. checkpoints/optimizer_step1000.pt +3 -0
  12. checkpoints/optimizer_step1500.pt +3 -0
  13. checkpoints/optimizer_step2000.pt +3 -0
  14. checkpoints/optimizer_step2500.pt +3 -0
  15. checkpoints/optimizer_step3000.pt +3 -0
  16. checkpoints/optimizer_step3500.pt +3 -0
  17. checkpoints/optimizer_step4000.pt +3 -0
  18. checkpoints/optimizer_step4500.pt +3 -0
  19. checkpoints/optimizer_step500.pt +3 -0
  20. checkpoints/optimizer_step5000.pt +3 -0
  21. final_model/model.safetensors +3 -0
  22. finetune_20260413_022538.log +180 -0
checkpoints/checkpoint_step1000.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c0974ccea58c52b3fce333cac4c82f3c709a4ec4ea37c24a183c7713df9e86c4
3
+ size 7233650408
checkpoints/checkpoint_step1500.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ea590879fac0dcdcdf3f242fe799915830f98a3cb2a469214cd491ac4b4e3a7e
3
+ size 7233650408
checkpoints/checkpoint_step2000.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ff7af87790dc36a235506bcb772d98eb4abccc2f7776d611878348ff779d100e
3
+ size 7233650408
checkpoints/checkpoint_step2500.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f257183abc026c698b1ca59fb9ab45817ea6f3d7247546e88a645be3639dc6d6
3
+ size 7233650408
checkpoints/checkpoint_step3000.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3594aca77c5c2ca9e53fda72577fc9ed30e8e422bdf9807cb13adedbcaa08b17
3
+ size 7233650408
checkpoints/checkpoint_step3500.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:eb1dba57882356cee7a2d38d93d7dffec334db6a37d4230fcf3659cf96babefd
3
+ size 7233650408
checkpoints/checkpoint_step4000.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:725420e8e8c797503a6895ad5f663222c70c6c92b10dab32980781cda82ede39
3
+ size 7233650408
checkpoints/checkpoint_step4500.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9d78bab153aebe1900f1278aa26aa7e915e253ec70f8e769d8913e7ca2bd0719
3
+ size 7233650408
checkpoints/checkpoint_step500.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b231950f9c80f66e850759b17e9d5445b77f6ed383cafb6438008a4efdc6c34a
3
+ size 7233650408
checkpoints/checkpoint_step5000.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:33a5758034a844a79910e3ef1be9349dbfb0901c3e974acf4ebc231501b198a1
3
+ size 7233650408
checkpoints/optimizer_step1000.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:428599b1d98ab639b60327b1c0d131d29d54563885f93d70e9db6353aecccf87
3
+ size 6602287841
checkpoints/optimizer_step1500.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fb3d96421d2e446faacf5a27a08945fa9d526090b6b6059327f83d508043af37
3
+ size 6602287841
checkpoints/optimizer_step2000.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d8c5fcbad3a313d429acca06e65ae65f837174e006e48942baef8bea1296540e
3
+ size 6602287841
checkpoints/optimizer_step2500.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:29242811b6f011cf50ddf714c54aa10ae2762e4dcea67c8ba6e0e9725c13174f
3
+ size 6602287841
checkpoints/optimizer_step3000.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a4bcdd1a091f7eefff6f0d681e24bba4805026f5e24e7f3f4be66e8190f7238c
3
+ size 6602287841
checkpoints/optimizer_step3500.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4738e20f1d406c8c6b6df57ff4bf2186795fb5e407869cb3cf1e45bee20462e3
3
+ size 6602287841
checkpoints/optimizer_step4000.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b01b0935335a8ed0f2549145586356b4ac025b2c881b4d75deaa77bfe07dee0a
3
+ size 6602287841
checkpoints/optimizer_step4500.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:406272907ce7a346b340034e3aad6edbcdee5421dca0feeb0b39fa032be895d8
3
+ size 6602287841
checkpoints/optimizer_step500.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ee039266e6f12044872101c0397ec4fdd895effdcf8e51166730c2d3fab26d01
3
+ size 6602285259
checkpoints/optimizer_step5000.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3127dc06fe4500830b0b2ef6d59ba7c4d67ba8b765eea8df09d02b41a014ca14
3
+ size 6602287841
final_model/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:33a5758034a844a79910e3ef1be9349dbfb0901c3e974acf4ebc231501b198a1
3
+ size 7233650408
finetune_20260413_022538.log ADDED
@@ -0,0 +1,180 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ 2026-04-13 02:25:47,641 - finetune_pi05 - INFO - ============================================================
2
+ 2026-04-13 02:25:47,641 - finetune_pi05 - INFO - PI05 Behavior Cloning Fine-tuning on LIBERO
3
+ 2026-04-13 02:25:47,641 - finetune_pi05 - INFO - World size: 8 GPUs
4
+ 2026-04-13 02:25:47,642 - finetune_pi05 - INFO - ============================================================
5
+ 2026-04-13 02:25:47,642 - finetune_pi05 - INFO - pi05_path: /SSD_DISK/users/wangyuang/Ctrl-World/openpi/ckpt/pi05_base_pytorch
6
+ 2026-04-13 02:25:47,642 - finetune_pi05 - INFO - train_suites: ['libero_spatial', 'libero_object', 'libero_goal']
7
+ 2026-04-13 02:25:47,642 - finetune_pi05 - INFO - train_demo_ratio: 0.2
8
+ 2026-04-13 02:25:47,642 - finetune_pi05 - INFO - batch_size: 4
9
+ 2026-04-13 02:25:47,642 - finetune_pi05 - INFO - num_workers: 4
10
+ 2026-04-13 02:25:47,642 - finetune_pi05 - INFO - total_steps: 5000
11
+ 2026-04-13 02:25:47,642 - finetune_pi05 - INFO - warmup_steps: 1000
12
+ 2026-04-13 02:25:47,643 - finetune_pi05 - INFO - peak_lr: 5e-05
13
+ 2026-04-13 02:25:47,643 - finetune_pi05 - INFO - min_lr: 1e-06
14
+ 2026-04-13 02:25:47,643 - finetune_pi05 - INFO - weight_decay: 0.01
15
+ 2026-04-13 02:25:47,643 - finetune_pi05 - INFO - grad_clip_norm: 1.0
16
+ 2026-04-13 02:25:47,643 - finetune_pi05 - INFO - save_steps: 500
17
+ 2026-04-13 02:25:47,643 - finetune_pi05 - INFO - log_interval: 50
18
+ 2026-04-13 02:25:47,643 - finetune_pi05 - INFO - wandb_project: pi05-libero-finetune
19
+ 2026-04-13 02:25:47,643 - finetune_pi05 - INFO - wandb_name: None
20
+ 2026-04-13 02:25:47,643 - finetune_pi05 - INFO - output_dir: output/pi05_finetune
21
+ 2026-04-13 02:25:47,643 - finetune_pi05 - INFO - device: cuda
22
+ 2026-04-13 02:25:47,644 - finetune_pi05 - INFO - seed: 42
23
+ 2026-04-13 02:25:47,644 - finetune_pi05 - INFO - Loading PI05-Base from /SSD_DISK/users/wangyuang/Ctrl-World/openpi/ckpt/pi05_base_pytorch ...
24
+ 2026-04-13 02:26:43,763 - finetune_pi05 - INFO - Total params: 3.62B, Trainable: 3.62B
25
+ 2026-04-13 02:26:44,468 - finetune_pi05 - INFO - Model wrapped with DDP on 8 GPUs
26
+ 2026-04-13 02:26:45,007 - finetune_pi05 - INFO - Optimizer: AdamW8bit (bitsandbytes)
27
+ 2026-04-13 02:26:45,008 - finetune_pi05 - INFO - Creating LIBERO training dataloader ...
28
+ 2026-04-13 02:26:45,008 - data.libero_dataset - INFO - [libero_spatial] Found 10 task files
29
+ 2026-04-13 02:26:45,013 - data.libero_dataset - INFO - [pick_up_the_black_bowl_between_the_plate_and_the_ramekin_and_place_it_on_the_plate_demo] train: 10/50 demos (ratio=0.2), 860 samples
30
+ 2026-04-13 02:26:45,017 - data.libero_dataset - INFO - [pick_up_the_black_bowl_from_table_center_and_place_it_on_the_plate_demo] train: 10/50 demos (ratio=0.2), 1025 samples
31
+ 2026-04-13 02:26:45,380 - data.libero_dataset - INFO - [pick_up_the_black_bowl_in_the_top_drawer_of_the_wooden_cabinet_and_place_it_on_the_plate_demo] train: 10/50 demos (ratio=0.2), 1397 samples
32
+ 2026-04-13 02:26:45,384 - data.libero_dataset - INFO - [pick_up_the_black_bowl_next_to_the_cookie_box_and_place_it_on_the_plate_demo] train: 10/50 demos (ratio=0.2), 1139 samples
33
+ 2026-04-13 02:26:45,388 - data.libero_dataset - INFO - [pick_up_the_black_bowl_next_to_the_plate_and_place_it_on_the_plate_demo] train: 10/50 demos (ratio=0.2), 1115 samples
34
+ 2026-04-13 02:26:45,392 - data.libero_dataset - INFO - [pick_up_the_black_bowl_next_to_the_ramekin_and_place_it_on_the_plate_demo] train: 10/50 demos (ratio=0.2), 1342 samples
35
+ 2026-04-13 02:26:45,396 - data.libero_dataset - INFO - [pick_up_the_black_bowl_on_the_cookie_box_and_place_it_on_the_plate_demo] train: 10/50 demos (ratio=0.2), 958 samples
36
+ 2026-04-13 02:26:45,400 - data.libero_dataset - INFO - [pick_up_the_black_bowl_on_the_ramekin_and_place_it_on_the_plate_demo] train: 10/50 demos (ratio=0.2), 1100 samples
37
+ 2026-04-13 02:26:45,404 - data.libero_dataset - INFO - [pick_up_the_black_bowl_on_the_stove_and_place_it_on_the_plate_demo] train: 10/50 demos (ratio=0.2), 1316 samples
38
+ 2026-04-13 02:26:45,408 - data.libero_dataset - INFO - [pick_up_the_black_bowl_on_the_wooden_cabinet_and_place_it_on_the_plate_demo] train: 10/50 demos (ratio=0.2), 1315 samples
39
+ 2026-04-13 02:26:45,409 - data.libero_dataset - INFO - [libero_spatial] train total samples: 11567
40
+ 2026-04-13 02:26:45,410 - data.libero_dataset - INFO - [libero_object] Found 10 task files
41
+ 2026-04-13 02:26:45,414 - data.libero_dataset - INFO - [pick_up_the_alphabet_soup_and_place_it_in_the_basket_demo] train: 10/50 demos (ratio=0.2), 1456 samples
42
+ 2026-04-13 02:26:45,418 - data.libero_dataset - INFO - [pick_up_the_bbq_sauce_and_place_it_in_the_basket_demo] train: 10/50 demos (ratio=0.2), 1419 samples
43
+ 2026-04-13 02:26:45,422 - data.libero_dataset - INFO - [pick_up_the_butter_and_place_it_in_the_basket_demo] train: 10/50 demos (ratio=0.2), 1462 samples
44
+ 2026-04-13 02:26:45,427 - data.libero_dataset - INFO - [pick_up_the_chocolate_pudding_and_place_it_in_the_basket_demo] train: 10/50 demos (ratio=0.2), 1460 samples
45
+ 2026-04-13 02:26:45,431 - data.libero_dataset - INFO - [pick_up_the_cream_cheese_and_place_it_in_the_basket_demo] train: 10/50 demos (ratio=0.2), 1402 samples
46
+ 2026-04-13 02:26:45,435 - data.libero_dataset - INFO - [pick_up_the_ketchup_and_place_it_in_the_basket_demo] train: 10/50 demos (ratio=0.2), 1515 samples
47
+ 2026-04-13 02:26:45,440 - data.libero_dataset - INFO - [pick_up_the_milk_and_place_it_in_the_basket_demo] train: 10/50 demos (ratio=0.2), 1380 samples
48
+ 2026-04-13 02:26:45,444 - data.libero_dataset - INFO - [pick_up_the_orange_juice_and_place_it_in_the_basket_demo] train: 10/50 demos (ratio=0.2), 1219 samples
49
+ 2026-04-13 02:26:45,448 - data.libero_dataset - INFO - [pick_up_the_salad_dressing_and_place_it_in_the_basket_demo] train: 10/50 demos (ratio=0.2), 1303 samples
50
+ 2026-04-13 02:26:45,452 - data.libero_dataset - INFO - [pick_up_the_tomato_sauce_and_place_it_in_the_basket_demo] train: 10/50 demos (ratio=0.2), 1354 samples
51
+ 2026-04-13 02:26:45,453 - data.libero_dataset - INFO - [libero_object] train total samples: 13970
52
+ 2026-04-13 02:26:45,453 - data.libero_dataset - INFO - [libero_goal] Found 10 task files
53
+ 2026-04-13 02:26:45,457 - data.libero_dataset - INFO - [open_the_middle_drawer_of_the_cabinet_demo] train: 10/50 demos (ratio=0.2), 1283 samples
54
+ 2026-04-13 02:26:45,462 - data.libero_dataset - INFO - [open_the_top_drawer_and_put_the_bowl_inside_demo] train: 10/50 demos (ratio=0.2), 1903 samples
55
+ 2026-04-13 02:26:45,466 - data.libero_dataset - INFO - [push_the_plate_to_the_front_of_the_stove_demo] train: 10/50 demos (ratio=0.2), 1380 samples
56
+ 2026-04-13 02:26:45,470 - data.libero_dataset - INFO - [put_the_bowl_on_the_plate_demo] train: 10/50 demos (ratio=0.2), 804 samples
57
+ 2026-04-13 02:26:45,474 - data.libero_dataset - INFO - [put_the_bowl_on_the_stove_demo] train: 10/50 demos (ratio=0.2), 902 samples
58
+ 2026-04-13 02:26:45,478 - data.libero_dataset - INFO - [put_the_bowl_on_top_of_the_cabinet_demo] train: 10/50 demos (ratio=0.2), 919 samples
59
+ 2026-04-13 02:26:45,482 - data.libero_dataset - INFO - [put_the_cream_cheese_in_the_bowl_demo] train: 10/50 demos (ratio=0.2), 897 samples
60
+ 2026-04-13 02:26:45,486 - data.libero_dataset - INFO - [put_the_wine_bottle_on_the_rack_demo] train: 10/50 demos (ratio=0.2), 1869 samples
61
+ 2026-04-13 02:26:45,490 - data.libero_dataset - INFO - [put_the_wine_bottle_on_top_of_the_cabinet_demo] train: 10/50 demos (ratio=0.2), 967 samples
62
+ 2026-04-13 02:26:45,494 - data.libero_dataset - INFO - [turn_on_the_stove_demo] train: 10/50 demos (ratio=0.2), 789 samples
63
+ 2026-04-13 02:26:45,495 - data.libero_dataset - INFO - [libero_goal] train total samples: 11713
64
+ 2026-04-13 02:26:45,495 - finetune_pi05 - INFO - Total samples: 37250
65
+ 2026-04-13 02:26:45,496 - finetune_pi05 - INFO - Per-GPU batch size: 4, effective batch: 32
66
+ 2026-04-13 02:26:45,496 - finetune_pi05 - INFO - Training batches per epoch (per GPU): 1164
67
+ 2026-04-13 02:26:45,501 - finetune_pi05 - INFO - Starting training for 5000 steps ...
68
+ 2026-04-13 02:28:40,161 - finetune_pi05 - INFO - Step 50/5000 | loss=0.6290 | lr=2.45e-06 | grad_norm=13.69 | speed=0.4 steps/s | ETA=3.2h
69
+ 2026-04-13 02:30:33,017 - finetune_pi05 - INFO - Step 100/5000 | loss=0.4917 | lr=4.95e-06 | grad_norm=3.16 | speed=0.4 steps/s | ETA=3.1h
70
+ 2026-04-13 02:32:25,679 - finetune_pi05 - INFO - Step 150/5000 | loss=0.3837 | lr=7.45e-06 | grad_norm=4.28 | speed=0.4 steps/s | ETA=3.1h
71
+ 2026-04-13 02:34:18,262 - finetune_pi05 - INFO - Step 200/5000 | loss=0.2913 | lr=9.95e-06 | grad_norm=1.85 | speed=0.4 steps/s | ETA=3.0h
72
+ 2026-04-13 02:36:10,911 - finetune_pi05 - INFO - Step 250/5000 | loss=0.2023 | lr=1.25e-05 | grad_norm=25.75 | speed=0.4 steps/s | ETA=3.0h
73
+ 2026-04-13 02:38:03,391 - finetune_pi05 - INFO - Step 300/5000 | loss=0.1875 | lr=1.50e-05 | grad_norm=1.91 | speed=0.4 steps/s | ETA=3.0h
74
+ 2026-04-13 02:39:56,039 - finetune_pi05 - INFO - Step 350/5000 | loss=0.1657 | lr=1.75e-05 | grad_norm=3.66 | speed=0.4 steps/s | ETA=2.9h
75
+ 2026-04-13 02:41:48,688 - finetune_pi05 - INFO - Step 400/5000 | loss=0.1214 | lr=2.00e-05 | grad_norm=1.73 | speed=0.4 steps/s | ETA=2.9h
76
+ 2026-04-13 02:43:41,508 - finetune_pi05 - INFO - Step 450/5000 | loss=0.1239 | lr=2.25e-05 | grad_norm=3.52 | speed=0.4 steps/s | ETA=2.9h
77
+ 2026-04-13 02:45:33,833 - finetune_pi05 - INFO - Step 500/5000 | loss=0.0918 | lr=2.49e-05 | grad_norm=2.78 | speed=0.4 steps/s | ETA=2.8h
78
+ 2026-04-13 02:45:54,561 - finetune_pi05 - INFO - Checkpoint saved at step 500 -> output/pi05_finetune/checkpoints
79
+ 2026-04-13 02:47:47,004 - finetune_pi05 - INFO - Step 550/5000 | loss=0.1047 | lr=2.75e-05 | grad_norm=1.16 | speed=0.4 steps/s | ETA=2.8h
80
+ 2026-04-13 02:49:39,594 - finetune_pi05 - INFO - Step 600/5000 | loss=0.0993 | lr=2.99e-05 | grad_norm=2.86 | speed=0.4 steps/s | ETA=2.8h
81
+ 2026-04-13 02:51:32,415 - finetune_pi05 - INFO - Step 650/5000 | loss=0.0821 | lr=3.25e-05 | grad_norm=0.66 | speed=0.4 steps/s | ETA=2.8h
82
+ 2026-04-13 02:53:25,043 - finetune_pi05 - INFO - Step 700/5000 | loss=0.0908 | lr=3.50e-05 | grad_norm=1.95 | speed=0.4 steps/s | ETA=2.7h
83
+ 2026-04-13 02:55:17,514 - finetune_pi05 - INFO - Step 750/5000 | loss=0.0965 | lr=3.75e-05 | grad_norm=3.91 | speed=0.4 steps/s | ETA=2.7h
84
+ 2026-04-13 02:57:09,851 - finetune_pi05 - INFO - Step 800/5000 | loss=0.0771 | lr=4.00e-05 | grad_norm=0.80 | speed=0.4 steps/s | ETA=2.7h
85
+ 2026-04-13 02:59:02,384 - finetune_pi05 - INFO - Step 850/5000 | loss=0.0856 | lr=4.25e-05 | grad_norm=3.56 | speed=0.4 steps/s | ETA=2.6h
86
+ 2026-04-13 03:00:55,311 - finetune_pi05 - INFO - Step 900/5000 | loss=0.0759 | lr=4.50e-05 | grad_norm=1.07 | speed=0.4 steps/s | ETA=2.6h
87
+ 2026-04-13 03:02:47,899 - finetune_pi05 - INFO - Step 950/5000 | loss=0.0799 | lr=4.75e-05 | grad_norm=0.85 | speed=0.4 steps/s | ETA=2.6h
88
+ 2026-04-13 03:04:40,008 - finetune_pi05 - INFO - Step 1000/5000 | loss=0.0706 | lr=5.00e-05 | grad_norm=1.27 | speed=0.4 steps/s | ETA=2.5h
89
+ 2026-04-13 03:05:06,959 - finetune_pi05 - INFO - Checkpoint saved at step 1000 -> output/pi05_finetune/checkpoints
90
+ 2026-04-13 03:06:59,574 - finetune_pi05 - INFO - Step 1050/5000 | loss=0.0675 | lr=5.00e-05 | grad_norm=0.90 | speed=0.4 steps/s | ETA=2.5h
91
+ 2026-04-13 03:08:51,656 - finetune_pi05 - INFO - Step 1100/5000 | loss=0.0712 | lr=4.99e-05 | grad_norm=1.10 | speed=0.4 steps/s | ETA=2.5h
92
+ 2026-04-13 03:10:44,475 - finetune_pi05 - INFO - Step 1150/5000 | loss=0.0713 | lr=4.98e-05 | grad_norm=0.63 | speed=0.4 steps/s | ETA=2.5h
93
+ 2026-04-13 03:12:36,982 - finetune_pi05 - INFO - Step 1200/5000 | loss=0.0638 | lr=4.97e-05 | grad_norm=0.64 | speed=0.4 steps/s | ETA=2.4h
94
+ 2026-04-13 03:14:28,485 - finetune_pi05 - INFO - Step 1250/5000 | loss=0.0648 | lr=4.95e-05 | grad_norm=0.30 | speed=0.4 steps/s | ETA=2.4h
95
+ 2026-04-13 03:16:20,259 - finetune_pi05 - INFO - Step 1300/5000 | loss=0.0571 | lr=4.93e-05 | grad_norm=0.39 | speed=0.4 steps/s | ETA=2.4h
96
+ 2026-04-13 03:18:12,169 - finetune_pi05 - INFO - Step 1350/5000 | loss=0.0555 | lr=4.91e-05 | grad_norm=0.82 | speed=0.4 steps/s | ETA=2.3h
97
+ 2026-04-13 03:20:04,242 - finetune_pi05 - INFO - Step 1400/5000 | loss=0.0491 | lr=4.88e-05 | grad_norm=1.07 | speed=0.4 steps/s | ETA=2.3h
98
+ 2026-04-13 03:21:55,991 - finetune_pi05 - INFO - Step 1450/5000 | loss=0.0456 | lr=4.85e-05 | grad_norm=0.95 | speed=0.4 steps/s | ETA=2.3h
99
+ 2026-04-13 03:23:48,179 - finetune_pi05 - INFO - Step 1500/5000 | loss=0.0500 | lr=4.81e-05 | grad_norm=0.41 | speed=0.4 steps/s | ETA=2.2h
100
+ 2026-04-13 03:24:07,206 - finetune_pi05 - INFO - Checkpoint saved at step 1500 -> output/pi05_finetune/checkpoints
101
+ 2026-04-13 03:25:59,382 - finetune_pi05 - INFO - Step 1550/5000 | loss=0.0494 | lr=4.78e-05 | grad_norm=0.41 | speed=0.4 steps/s | ETA=2.2h
102
+ 2026-04-13 03:27:51,937 - finetune_pi05 - INFO - Step 1600/5000 | loss=0.0564 | lr=4.73e-05 | grad_norm=0.58 | speed=0.4 steps/s | ETA=2.2h
103
+ 2026-04-13 03:29:44,397 - finetune_pi05 - INFO - Step 1650/5000 | loss=0.0475 | lr=4.69e-05 | grad_norm=0.62 | speed=0.4 steps/s | ETA=2.1h
104
+ 2026-04-13 03:31:37,010 - finetune_pi05 - INFO - Step 1700/5000 | loss=0.0493 | lr=4.64e-05 | grad_norm=0.61 | speed=0.4 steps/s | ETA=2.1h
105
+ 2026-04-13 03:33:29,390 - finetune_pi05 - INFO - Step 1750/5000 | loss=0.0409 | lr=4.59e-05 | grad_norm=0.39 | speed=0.4 steps/s | ETA=2.1h
106
+ 2026-04-13 03:35:22,218 - finetune_pi05 - INFO - Step 1800/5000 | loss=0.0470 | lr=4.53e-05 | grad_norm=0.75 | speed=0.4 steps/s | ETA=2.0h
107
+ 2026-04-13 03:37:14,779 - finetune_pi05 - INFO - Step 1850/5000 | loss=0.0349 | lr=4.48e-05 | grad_norm=0.29 | speed=0.4 steps/s | ETA=2.0h
108
+ 2026-04-13 03:39:06,357 - finetune_pi05 - INFO - Step 1900/5000 | loss=0.0499 | lr=4.41e-05 | grad_norm=0.37 | speed=0.4 steps/s | ETA=2.0h
109
+ 2026-04-13 03:40:58,913 - finetune_pi05 - INFO - Step 1950/5000 | loss=0.0441 | lr=4.35e-05 | grad_norm=0.57 | speed=0.4 steps/s | ETA=1.9h
110
+ 2026-04-13 03:42:51,413 - finetune_pi05 - INFO - Step 2000/5000 | loss=0.0396 | lr=4.28e-05 | grad_norm=0.28 | speed=0.4 steps/s | ETA=1.9h
111
+ 2026-04-13 03:43:10,625 - finetune_pi05 - INFO - Checkpoint saved at step 2000 -> output/pi05_finetune/checkpoints
112
+ 2026-04-13 03:45:03,592 - finetune_pi05 - INFO - Step 2050/5000 | loss=0.0436 | lr=4.21e-05 | grad_norm=0.43 | speed=0.4 steps/s | ETA=1.9h
113
+ 2026-04-13 03:46:56,187 - finetune_pi05 - INFO - Step 2100/5000 | loss=0.0449 | lr=4.14e-05 | grad_norm=0.52 | speed=0.4 steps/s | ETA=1.8h
114
+ 2026-04-13 03:48:48,855 - finetune_pi05 - INFO - Step 2150/5000 | loss=0.0391 | lr=4.07e-05 | grad_norm=0.33 | speed=0.4 steps/s | ETA=1.8h
115
+ 2026-04-13 03:50:41,376 - finetune_pi05 - INFO - Step 2200/5000 | loss=0.0352 | lr=3.99e-05 | grad_norm=0.66 | speed=0.4 steps/s | ETA=1.8h
116
+ 2026-04-13 03:52:34,090 - finetune_pi05 - INFO - Step 2250/5000 | loss=0.0414 | lr=3.91e-05 | grad_norm=0.69 | speed=0.4 steps/s | ETA=1.7h
117
+ 2026-04-13 03:54:26,486 - finetune_pi05 - INFO - Step 2300/5000 | loss=0.0288 | lr=3.83e-05 | grad_norm=0.23 | speed=0.4 steps/s | ETA=1.7h
118
+ 2026-04-13 03:56:18,363 - finetune_pi05 - INFO - Step 2350/5000 | loss=0.0357 | lr=3.75e-05 | grad_norm=0.52 | speed=0.4 steps/s | ETA=1.7h
119
+ 2026-04-13 03:58:10,738 - finetune_pi05 - INFO - Step 2400/5000 | loss=0.0304 | lr=3.66e-05 | grad_norm=0.26 | speed=0.4 steps/s | ETA=1.7h
120
+ 2026-04-13 04:00:03,572 - finetune_pi05 - INFO - Step 2450/5000 | loss=0.0287 | lr=3.58e-05 | grad_norm=0.17 | speed=0.4 steps/s | ETA=1.6h
121
+ 2026-04-13 04:01:55,795 - finetune_pi05 - INFO - Step 2500/5000 | loss=0.0302 | lr=3.49e-05 | grad_norm=0.69 | speed=0.4 steps/s | ETA=1.6h
122
+ 2026-04-13 04:02:14,854 - finetune_pi05 - INFO - Checkpoint saved at step 2500 -> output/pi05_finetune/checkpoints
123
+ 2026-04-13 04:04:07,825 - finetune_pi05 - INFO - Step 2550/5000 | loss=0.0318 | lr=3.40e-05 | grad_norm=0.36 | speed=0.4 steps/s | ETA=1.6h
124
+ 2026-04-13 04:06:00,500 - finetune_pi05 - INFO - Step 2600/5000 | loss=0.0285 | lr=3.31e-05 | grad_norm=0.17 | speed=0.4 steps/s | ETA=1.5h
125
+ 2026-04-13 04:07:52,887 - finetune_pi05 - INFO - Step 2650/5000 | loss=0.0250 | lr=3.22e-05 | grad_norm=0.72 | speed=0.4 steps/s | ETA=1.5h
126
+ 2026-04-13 04:09:45,167 - finetune_pi05 - INFO - Step 2700/5000 | loss=0.0280 | lr=3.12e-05 | grad_norm=0.20 | speed=0.4 steps/s | ETA=1.5h
127
+ 2026-04-13 04:11:37,584 - finetune_pi05 - INFO - Step 2750/5000 | loss=0.0290 | lr=3.03e-05 | grad_norm=0.32 | speed=0.4 steps/s | ETA=1.4h
128
+ 2026-04-13 04:13:29,801 - finetune_pi05 - INFO - Step 2800/5000 | loss=0.0248 | lr=2.94e-05 | grad_norm=0.72 | speed=0.4 steps/s | ETA=1.4h
129
+ 2026-04-13 04:15:22,355 - finetune_pi05 - INFO - Step 2850/5000 | loss=0.0224 | lr=2.84e-05 | grad_norm=0.37 | speed=0.4 steps/s | ETA=1.4h
130
+ 2026-04-13 04:17:13,837 - finetune_pi05 - INFO - Step 2900/5000 | loss=0.0287 | lr=2.74e-05 | grad_norm=0.28 | speed=0.4 steps/s | ETA=1.3h
131
+ 2026-04-13 04:19:05,441 - finetune_pi05 - INFO - Step 2950/5000 | loss=0.0250 | lr=2.65e-05 | grad_norm=0.25 | speed=0.4 steps/s | ETA=1.3h
132
+ 2026-04-13 04:20:58,092 - finetune_pi05 - INFO - Step 3000/5000 | loss=0.0259 | lr=2.55e-05 | grad_norm=0.40 | speed=0.4 steps/s | ETA=1.3h
133
+ 2026-04-13 04:21:16,990 - finetune_pi05 - INFO - Checkpoint saved at step 3000 -> output/pi05_finetune/checkpoints
134
+ 2026-04-13 04:23:09,227 - finetune_pi05 - INFO - Step 3050/5000 | loss=0.0213 | lr=2.46e-05 | grad_norm=0.20 | speed=0.4 steps/s | ETA=1.2h
135
+ 2026-04-13 04:25:01,728 - finetune_pi05 - INFO - Step 3100/5000 | loss=0.0331 | lr=2.36e-05 | grad_norm=2.89 | speed=0.4 steps/s | ETA=1.2h
136
+ 2026-04-13 04:26:54,247 - finetune_pi05 - INFO - Step 3150/5000 | loss=0.0227 | lr=2.26e-05 | grad_norm=0.45 | speed=0.4 steps/s | ETA=1.2h
137
+ 2026-04-13 04:28:46,667 - finetune_pi05 - INFO - Step 3200/5000 | loss=0.0225 | lr=2.17e-05 | grad_norm=0.80 | speed=0.4 steps/s | ETA=1.1h
138
+ 2026-04-13 04:30:39,074 - finetune_pi05 - INFO - Step 3250/5000 | loss=0.0249 | lr=2.07e-05 | grad_norm=0.15 | speed=0.4 steps/s | ETA=1.1h
139
+ 2026-04-13 04:32:31,522 - finetune_pi05 - INFO - Step 3300/5000 | loss=0.0230 | lr=1.98e-05 | grad_norm=0.38 | speed=0.4 steps/s | ETA=1.1h
140
+ 2026-04-13 04:34:23,728 - finetune_pi05 - INFO - Step 3350/5000 | loss=0.0225 | lr=1.89e-05 | grad_norm=0.21 | speed=0.4 steps/s | ETA=1.0h
141
+ 2026-04-13 04:36:15,728 - finetune_pi05 - INFO - Step 3400/5000 | loss=0.0226 | lr=1.79e-05 | grad_norm=0.22 | speed=0.4 steps/s | ETA=1.0h
142
+ 2026-04-13 04:38:08,024 - finetune_pi05 - INFO - Step 3450/5000 | loss=0.0277 | lr=1.70e-05 | grad_norm=0.32 | speed=0.4 steps/s | ETA=1.0h
143
+ 2026-04-13 04:40:00,020 - finetune_pi05 - INFO - Step 3500/5000 | loss=0.0306 | lr=1.61e-05 | grad_norm=4.00 | speed=0.4 steps/s | ETA=1.0h
144
+ 2026-04-13 04:40:19,197 - finetune_pi05 - INFO - Checkpoint saved at step 3500 -> output/pi05_finetune/checkpoints
145
+ 2026-04-13 04:42:11,856 - finetune_pi05 - INFO - Step 3550/5000 | loss=0.0214 | lr=1.53e-05 | grad_norm=0.43 | speed=0.4 steps/s | ETA=0.9h
146
+ 2026-04-13 04:44:04,615 - finetune_pi05 - INFO - Step 3600/5000 | loss=0.0280 | lr=1.44e-05 | grad_norm=0.39 | speed=0.4 steps/s | ETA=0.9h
147
+ 2026-04-13 04:45:57,552 - finetune_pi05 - INFO - Step 3650/5000 | loss=0.0161 | lr=1.35e-05 | grad_norm=0.20 | speed=0.4 steps/s | ETA=0.9h
148
+ 2026-04-13 04:47:50,298 - finetune_pi05 - INFO - Step 3700/5000 | loss=0.0217 | lr=1.27e-05 | grad_norm=0.36 | speed=0.4 steps/s | ETA=0.8h
149
+ 2026-04-13 04:49:42,969 - finetune_pi05 - INFO - Step 3750/5000 | loss=0.0222 | lr=1.19e-05 | grad_norm=0.33 | speed=0.4 steps/s | ETA=0.8h
150
+ 2026-04-13 04:51:36,347 - finetune_pi05 - INFO - Step 3800/5000 | loss=0.0197 | lr=1.11e-05 | grad_norm=0.43 | speed=0.4 steps/s | ETA=0.8h
151
+ 2026-04-13 04:53:28,226 - finetune_pi05 - INFO - Step 3850/5000 | loss=0.0214 | lr=1.03e-05 | grad_norm=0.34 | speed=0.4 steps/s | ETA=0.7h
152
+ 2026-04-13 04:55:20,951 - finetune_pi05 - INFO - Step 3900/5000 | loss=0.0243 | lr=9.60e-06 | grad_norm=0.33 | speed=0.4 steps/s | ETA=0.7h
153
+ 2026-04-13 04:57:13,282 - finetune_pi05 - INFO - Step 3950/5000 | loss=0.0203 | lr=8.88e-06 | grad_norm=0.17 | speed=0.4 steps/s | ETA=0.7h
154
+ 2026-04-13 04:59:06,195 - finetune_pi05 - INFO - Step 4000/5000 | loss=0.0203 | lr=8.19e-06 | grad_norm=0.42 | speed=0.4 steps/s | ETA=0.6h
155
+ 2026-04-13 04:59:25,423 - finetune_pi05 - INFO - Checkpoint saved at step 4000 -> output/pi05_finetune/checkpoints
156
+ 2026-04-13 05:01:18,431 - finetune_pi05 - INFO - Step 4050/5000 | loss=0.0185 | lr=7.52e-06 | grad_norm=0.17 | speed=0.4 steps/s | ETA=0.6h
157
+ 2026-04-13 05:03:11,013 - finetune_pi05 - INFO - Step 4100/5000 | loss=0.0228 | lr=6.88e-06 | grad_norm=0.27 | speed=0.4 steps/s | ETA=0.6h
158
+ 2026-04-13 05:05:03,860 - finetune_pi05 - INFO - Step 4150/5000 | loss=0.0209 | lr=6.27e-06 | grad_norm=0.59 | speed=0.4 steps/s | ETA=0.5h
159
+ 2026-04-13 05:06:56,130 - finetune_pi05 - INFO - Step 4200/5000 | loss=0.0220 | lr=5.69e-06 | grad_norm=0.36 | speed=0.4 steps/s | ETA=0.5h
160
+ 2026-04-13 05:08:48,444 - finetune_pi05 - INFO - Step 4250/5000 | loss=0.0222 | lr=5.14e-06 | grad_norm=0.24 | speed=0.4 steps/s | ETA=0.5h
161
+ 2026-04-13 05:10:41,510 - finetune_pi05 - INFO - Step 4300/5000 | loss=0.0223 | lr=4.62e-06 | grad_norm=2.53 | speed=0.4 steps/s | ETA=0.4h
162
+ 2026-04-13 05:12:33,865 - finetune_pi05 - INFO - Step 4350/5000 | loss=0.0202 | lr=4.13e-06 | grad_norm=0.25 | speed=0.4 steps/s | ETA=0.4h
163
+ 2026-04-13 05:14:26,631 - finetune_pi05 - INFO - Step 4400/5000 | loss=0.0184 | lr=3.68e-06 | grad_norm=0.20 | speed=0.4 steps/s | ETA=0.4h
164
+ 2026-04-13 05:16:19,700 - finetune_pi05 - INFO - Step 4450/5000 | loss=0.0191 | lr=3.26e-06 | grad_norm=0.27 | speed=0.4 steps/s | ETA=0.3h
165
+ 2026-04-13 05:18:11,639 - finetune_pi05 - INFO - Step 4500/5000 | loss=0.0167 | lr=2.87e-06 | grad_norm=0.26 | speed=0.4 steps/s | ETA=0.3h
166
+ 2026-04-13 05:18:30,720 - finetune_pi05 - INFO - Checkpoint saved at step 4500 -> output/pi05_finetune/checkpoints
167
+ 2026-04-13 05:20:22,760 - finetune_pi05 - INFO - Step 4550/5000 | loss=0.0208 | lr=2.52e-06 | grad_norm=0.29 | speed=0.4 steps/s | ETA=0.3h
168
+ 2026-04-13 05:22:14,814 - finetune_pi05 - INFO - Step 4600/5000 | loss=0.0180 | lr=2.21e-06 | grad_norm=0.17 | speed=0.4 steps/s | ETA=0.3h
169
+ 2026-04-13 05:24:06,857 - finetune_pi05 - INFO - Step 4650/5000 | loss=0.0184 | lr=1.93e-06 | grad_norm=0.33 | speed=0.4 steps/s | ETA=0.2h
170
+ 2026-04-13 05:25:59,568 - finetune_pi05 - INFO - Step 4700/5000 | loss=0.0221 | lr=1.68e-06 | grad_norm=0.28 | speed=0.4 steps/s | ETA=0.2h
171
+ 2026-04-13 05:27:52,958 - finetune_pi05 - INFO - Step 4750/5000 | loss=0.0194 | lr=1.47e-06 | grad_norm=0.22 | speed=0.4 steps/s | ETA=0.2h
172
+ 2026-04-13 05:29:45,795 - finetune_pi05 - INFO - Step 4800/5000 | loss=0.0202 | lr=1.30e-06 | grad_norm=0.23 | speed=0.4 steps/s | ETA=0.1h
173
+ 2026-04-13 05:31:39,334 - finetune_pi05 - INFO - Step 4850/5000 | loss=0.0210 | lr=1.17e-06 | grad_norm=0.52 | speed=0.4 steps/s | ETA=0.1h
174
+ 2026-04-13 05:33:32,474 - finetune_pi05 - INFO - Step 4900/5000 | loss=0.0203 | lr=1.08e-06 | grad_norm=0.29 | speed=0.4 steps/s | ETA=0.1h
175
+ 2026-04-13 05:35:25,771 - finetune_pi05 - INFO - Step 4950/5000 | loss=0.0185 | lr=1.02e-06 | grad_norm=0.19 | speed=0.4 steps/s | ETA=0.0h
176
+ 2026-04-13 05:37:18,545 - finetune_pi05 - INFO - Step 5000/5000 | loss=0.0232 | lr=1.00e-06 | grad_norm=0.25 | speed=0.4 steps/s | ETA=0.0h
177
+ 2026-04-13 05:37:37,347 - finetune_pi05 - INFO - Checkpoint saved at step 5000 -> output/pi05_finetune/checkpoints
178
+ 2026-04-13 05:37:37,465 - finetune_pi05 - INFO -
179
+ Training complete! Total time: 3.2h
180
+ 2026-04-13 05:37:49,154 - finetune_pi05 - INFO - Final model saved to output/pi05_finetune/final_model