Hugging Face's logo

Datasets:
hamidavatar
/
training-logs

Dataset card Files
xet
 Community
hamidavatar commited on
Commit
9919d6d
·
verified ·
1 Parent(s): b38d994

Upload training_20260312_235253_exit1.log with huggingface_hub

Browse files
Files changed (1) hide show
  1. training_20260312_235253_exit1.log +796 -0
training_20260312_235253_exit1.log ADDED
@@ -0,0 +1,796 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
0
  0%| | 0.00/44.7M [00:00<?, ?B/s]
1
  0%| | 0.00/44.7M [00:00<?, ?B/s]
2
  0%| | 0.00/44.7M [00:00<?, ?B/s]
3
  15%|█▌ | 6.75M/44.7M [00:00<00:00, 46.8MB/s]
4
  64%|██████▍ | 28.8M/44.7M [00:00<00:00, 101MB/s]
5
  19%|█▉ | 8.50M/44.7M [00:00<00:01, 28.8MB/s]
 
6
  39%|███▉ | 17.5M/44.7M [00:00<00:00, 48.0MB/s]
7
  58%|█████▊ | 26.0M/44.7M [00:00<00:00, 53.0MB/s]
8
  76%|███████▌ | 33.9M/44.7M [00:00<00:00, 58.3MB/s]
9
  15%|█▌ | 6.88M/44.7M [00:00<00:04, 9.60MB/s]
 
10
  24%|██▎ | 10.5M/44.7M [00:00<00:02, 13.1MB/s]
11
  0%| | 0.00/44.7M [00:00<?, ?B/s]
12
  5%|▌ | 2.25M/44.7M [00:00<00:01, 22.9MB/s]
13
  39%|███▉ | 17.6M/44.7M [00:01<00:01, 23.3MB/s][Avatar] Compiling policy with torch.compile (mode=reduce-overhead)...
 
14
  11%|█ | 4.75M/44.7M [00:00<00:01, 24.7MB/s]
15
  48%|████▊ | 21.5M/44.7M [00:01<00:01, 18.9MB/s]
16
  64%|██████▎ | 28.4M/44.7M [00:01<00:00, 27.0MB/s][Avatar] Compiling policy with torch.compile (mode=reduce-overhead)...
 
17
  72%|███████▏ | 32.4M/44.7M [00:01<00:00, 23.1MB/s]
18
  80%|███████▉ | 35.6M/44.7M [00:01<00:00, 21.9MB/s][Avatar] Policy compiled (first step will be slow)
 
19
  16%|█▌ | 7.12M/44.7M [00:01<00:07, 5.59MB/s]
20
  22%|██▏ | 9.62M/44.7M [00:01<00:04, 8.20MB/s]
21
  27%|██▋ | 11.9M/44.7M [00:01<00:04, 7.63MB/s][Avatar] Policy compiled (first step will be slow)
 
22
  33%|███▎ | 14.9M/44.7M [00:01<00:02, 10.7MB/s]
23
  86%|████████▌ | 38.4M/44.7M [00:02<00:00, 10.8MB/s]
24
  39%|███▉ | 17.6M/44.7M [00:01<00:02, 13.4MB/s]
25
  45%|████▌ | 20.1M/44.7M [00:01<00:01, 13.8MB/s]
26
  53%|█████▎ | 23.5M/44.7M [00:02<00:01, 17.8MB/s]
27
  59%|█████▉ | 26.4M/44.7M [00:02<00:01, 18.6MB/s]
28
  67%|██████▋ | 30.1M/44.7M [00:02<00:00, 23.1MB/s]
29
  82%|████████▏ | 36.8M/44.7M [00:02<00:00, 33.0MB/s]
30
  90%|█████████ | 40.4M/44.7M [00:03<00:00, 6.67MB/s]
 
31
  91%|█████████ | 40.5M/44.7M [00:02<00:00, 21.0MB/s]
32
  0%| | 0.00/44.7M [00:00<?, ?B/s]
33
  0%| | 0.00/44.7M [00:00<?, ?B/s]
34
  97%|█████████▋| 43.4M/44.7M [00:03<00:00, 16.2MB/s]
35
  1%| | 512k/44.7M [00:00<00:09, 4.68MB/s]
 
36
  2%|▏ | 896k/44.7M [00:00<00:16, 2.83MB/s]
37
  2%|▏ | 1.00M/44.7M [00:00<00:22, 2.02MB/s]
38
  3%|▎ | 1.38M/44.7M [00:00<00:27, 1.67MB/s]
39
  6%|▋ | 2.88M/44.7M [00:00<00:10, 4.33MB/s][Avatar] Compiling policy with torch.compile (mode=reduce-overhead)...
 
40
  12%|█▏ | 5.25M/44.7M [00:00<00:04, 8.67MB/s]
41
  17%|█▋ | 7.75M/44.7M [00:01<00:03, 12.7MB/s]
42
  0%| | 0.00/44.7M [00:00<?, ?B/s]
43
  0%| | 128k/44.7M [00:00<00:41, 1.11MB/s]
44
  21%|██▏ | 9.50M/44.7M [00:01<00:02, 12.4MB/s]
45
  11%|█ | 5.00M/44.7M [00:00<00:01, 27.5MB/s]
46
  3%|▎ | 1.25M/44.7M [00:01<00:56, 807kB/s]
47
  4%|▍ | 1.75M/44.7M [00:01<00:36, 1.24MB/s]
48
  19%|█▉ | 8.38M/44.7M [00:00<00:02, 14.8MB/s][Avatar] Policy compiled (first step will be slow)
 
49
  25%|██▍ | 11.0M/44.7M [00:01<00:05, 6.67MB/s]
50
  35%|███▍ | 15.5M/44.7M [00:00<00:01, 25.7MB/s]
51
  27%|██▋ | 12.1M/44.7M [00:01<00:04, 7.34MB/s]
52
  30%|███ | 13.5M/44.7M [00:01<00:03, 8.40MB/s]
53
  33%|███▎ | 14.8M/44.7M [00:02<00:03, 8.35MB/s]
54
  36%|███▌ | 16.0M/44.7M [00:02<00:03, 8.51MB/s]
55
  40%|████ | 17.9M/44.7M [00:02<00:02, 9.68MB/s]
56
  43%|████▎ | 19.2M/44.7M [00:02<00:02, 10.6MB/s]
57
  52%|█████▏ | 23.4M/44.7M [00:02<00:01, 13.3MB/s]
58
  58%|█████▊ | 26.1M/44.7M [00:02<00:01, 16.3MB/s]
59
  62%|██████▏ | 27.9M/44.7M [00:03<00:01, 15.9MB/s]
60
  67%|██████▋ | 30.0M/44.7M [00:03<00:00, 17.3MB/s]
61
  71%|███████▏ | 31.9M/44.7M [00:03<00:00, 16.0MB/s]
62
  75%|███████▌ | 33.5M/44.7M [00:03<00:00, 15.0MB/s]
63
  79%|███████��� | 35.2M/44.7M [00:03<00:00, 15.7MB/s]
64
  83%|████████▎ | 37.1M/44.7M [00:03<00:00, 16.7MB/s]
65
  87%|████████▋ | 38.9M/44.7M [00:03<00:00, 15.8MB/s]
66
  91%|█████████ | 40.5M/44.7M [00:03<00:00, 16.0MB/s]
67
  95%|█████████▍| 42.2M/44.7M [00:03<00:00, 16.6MB/s]
68
  98%|█████████▊| 43.9M/44.7M [00:04<00:00, 15.2MB/s]
 
 
 
 
69
  5%|▍ | 2.12M/44.7M [00:06<03:26, 216kB/s]
70
  6%|▌ | 2.62M/44.7M [00:06<02:11, 334kB/s][Avatar] Policy compiled (first step will be slow)
 
71
  8%|▊ | 3.38M/44.7M [00:06<01:14, 578kB/s]
72
  10%|▉ | 4.25M/44.7M [00:06<00:44, 960kB/s]
73
  13%|█▎ | 5.88M/44.7M [00:07<00:21, 1.92MB/s]
74
  17%|█▋ | 7.62M/44.7M [00:07<00:12, 3.20MB/s]
75
  20%|█▉ | 8.88M/44.7M [00:07<00:11, 3.21MB/s]
76
  22%|██▏ | 9.88M/44.7M [00:07<00:11, 3.27MB/s]
77
  27%|██▋ | 12.0M/44.7M [00:08<00:06, 5.26MB/s]
78
  42%|████▏ | 18.6M/44.7M [00:07<00:15, 1.80MB/s]
79
  30%|██▉ | 13.2M/44.7M [00:08<00:05, 5.57MB/s]
80
  49%|████▉ | 21.9M/44.7M [00:07<00:09, 2.48MB/s]
81
  55%|█████▍ | 24.4M/44.7M [00:07<00:06, 3.15MB/s]
82
  60%|█████▉ | 26.8M/44.7M [00:07<00:04, 3.96MB/s]
83
  32%|███▏ | 14.2M/44.7M [00:08<00:06, 4.67MB/s]
84
  35%|███▍ | 15.6M/44.7M [00:08<00:05, 5.54MB/s]
85
  37%|███▋ | 16.5M/44.7M [00:08<00:06, 4.72MB/s]
86
  39%|███▊ | 17.2M/44.7M [00:09<00:07, 4.08MB/s]
87
  65%|██████▍ | 29.0M/44.7M [00:08<00:04, 3.54MB/s]
88
  46%|████▌ | 20.6M/44.7M [00:09<00:03, 8.01MB/s]
89
  50%|█████ | 22.4M/44.7M [00:09<00:02, 9.56MB/s]
90
  69%|██████▉ | 30.9M/44.7M [00:08<00:03, 4.07MB/s]
91
  80%|████████ | 35.8M/44.7M [00:08<00:01, 7.02MB/s]
92
  88%|████████▊ | 39.1M/44.7M [00:08<00:00, 9.13MB/s]
93
  0%| | 0.00/44.7M [00:00<?, ?B/s]
94
  3%|▎ | 1.50M/44.7M [00:00<00:02, 15.7MB/s]
95
  53%|█████▎ | 23.8M/44.7M [00:10<00:05, 4.09MB/s]
96
  58%|█████▊ | 25.8M/44.7M [00:10<00:03, 5.64MB/s]
97
  60%|██████ | 27.0M/44.7M [00:10<00:02, 6.50MB/s]
98
  93%|█████████▎| 41.5M/44.7M [00:09<00:00, 5.74MB/s]
 
99
  66%|██████▋ | 29.6M/44.7M [00:11<00:02, 6.81MB/s]
100
  12%|█▏ | 5.38M/44.7M [00:01<00:13, 2.98MB/s]
101
  69%|██████▉ | 30.8M/44.7M [00:14<00:09, 1.53MB/s]
102
  71%|███████ | 31.6M/44.7M [00:14<00:08, 1.68MB/s]
103
  72%|███████▏ | 32.2M/44.7M [00:15<00:08, 1.58MB/s]
104
  73%|███████▎ | 32.8M/44.7M [00:15<00:07, 1.73MB/s][Avatar] Compiling policy with torch.compile (mode=reduce-overhead)...
 
105
  74%|███████▍ | 33.2M/44.7M [00:16<00:10, 1.13MB/s]
106
  75%|███████▌ | 33.6M/44.7M [00:16<00:10, 1.13MB/s][Avatar] Policy compiled (first step will be slow)
 
107
  76%|███████▋ | 34.1M/44.7M [00:17<00:09, 1.17MB/s]
108
  79%|███████▉ | 35.2M/44.7M [00:17<00:05, 1.90MB/s]
109
  83%|████████▎ | 37.0M/44.7M [00:17<00:02, 3.25MB/s]
110
  87%|████████▋ | 39.0M/44.7M [00:17<00:01, 5.00MB/s]
111
  92%|█████████▏| 40.9M/44.7M [00:17<00:00, 6.89MB/s]
112
  98%|█████████▊| 43.6M/44.7M [00:17<00:00, 10.2MB/s]
 
 
 
113
  15%|█▍ | 6.62M/44.7M [00:15<00:13, 2.98MB/s]
114
  15%|█▌ | 6.75M/44.7M [03:32<27:44, 23.9kB/s]
115
  17%|█▋ | 7.38M/44.7M [03:32<23:02, 28.3kB/s]
116
  18%|█▊ | 7.88M/44.7M [03:45<22:43, 28.3kB/s]
117
  18%|█▊ | 8.00M/44.7M [04:14<26:02, 24.6kB/s]
118
  18%|█▊ | 8.25M/44.7M [04:14<23:26, 27.1kB/s]
119
  20%|█▉ | 8.75M/44.7M [04:14<18:16, 34.3kB/s]
120
  20%|██ | 9.12M/44.7M [04:15<14:47, 42.0kB/s]
121
  21%|██▏ | 9.50M/44.7M [04:15<11:38, 52.8kB/s]
122
  21%|██▏ | 9.50M/44.7M [04:30<11:38, 52.8kB/s]
123
  22%|██▏ | 9.62M/44.7M [04:36<19:29, 31.4kB/s]
124
  22%|██▏ | 9.62M/44.7M [04:50<19:29, 31.4kB/s][rank6]:[E312 23:52:47.560128968 ProcessGroupNCCL.cpp:632] [Rank 6] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=2, OpType=ALLREDUCE, NumelIn=1, NumelOut=1, Timeout(ms)=600000) ran for 600082 milliseconds before timing out.
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ === Avatar Training (fire-and-forget) ===
2
+ Script: train_on_hf.py
3
+ Started: 2026-03-12 23:39:04 UTC
4
+
5
+ >>> [0/4] Raising file descriptor limit...
6
+ ulimit -n = 1048576
7
+ >>> [1/4] Setting environment...
8
+ HF_TOKEN set: 37 chars
9
+ WANDB_API_KEY set: 0 chars
10
+ >>> [2/4] Logging into HuggingFace...
11
+ The token has not been saved to the git credentials helper. Pass `add_to_git_credential=True` in this function directly or `--add-to-git-credential` if using via `hf`CLI if you want to set the git credential as well.
12
+ Token is valid (permission: write).
13
+ The token `beta-pc` has been saved to /root/.cache/huggingface/stored_tokens
14
+ Your token has been saved to /root/.cache/huggingface/token
15
+ Login successful.
16
+ Note: Environment variable`HF_TOKEN` is set and is the current active token independently from the token you've just configured.
17
+ ⚠️ Warning: 'huggingface-cli login' is deprecated. Use 'hf auth login' instead.
18
+ >>> [3/4] Pre-downloading pretrained model...
19
+ Downloading avatarrobotics/act-alpha-bot-right-resnet18-b12-20260312-1749...
20
+
21
+ Cached at /root/.cache/huggingface/hub/models--avatarrobotics--act-alpha-bot-right-resnet18-b12-20260312-1749/snapshots/67ded6ef397ecf93b5bc9a6df4cd4666ff74c534
22
+ GPUs: 8x NVIDIA A100-SXM4-80GB
23
+ >>> [4/4] Launching accelerate with 8 GPUs...
24
+ Script: train_on_hf.py
25
+ Train args: --dataset alpha_bot_right_afternoon --steps 400000 --batch-size 32 --lr 1e-4 --backbone resnet18 --preload --mixed-precision bf16 --compile --num-workers 8 --prefetch-factor 4 --save-freq 10000 --push-to-hub --pretrained-path avatarrobotics/act-alpha-bot-right-resnet18-b12-20260312-1749
26
+
27
+ The following values were not passed to `accelerate launch` and had defaults used instead:
28
+ `--num_machines` was set to a value of `1`
29
+ `--mixed_precision` was set to a value of `'no'`
30
+ `--dynamo_backend` was set to a value of `'no'`
31
+ To avoid this warning pass in values for each of the problematic parameters or run `accelerate config`.
32
+ Model will be pushed to: avatarrobotics/act-alpha-bot-right-resnet18-b256-20260312-2339
33
+ Model will be pushed to: avatarrobotics/act-alpha-bot-right-resnet18-b256-20260312-2339
34
+ [Avatar] Mixed precision: bf16
35
+
36
+ ============================================================
37
+ Training args:
38
+ mixed_precision=bf16
39
+ video_backend=pyav
40
+ prefetch_factor=4
41
+ preload=True
42
+ compile=True
43
+ backbone=resnet18
44
+ lerobot_train \
45
+ --dataset.repo_id=avatarrobotics/alpha-bot-right-2026-03-11-afternoon \
46
+ --dataset.video_backend=pyav \
47
+ --policy.repo_id=avatarrobotics/act-alpha-bot-right-resnet18-b256-20260312-2339 \
48
+ --policy.type=act \
49
+ --policy.chunk_size=100 \
50
+ --policy.n_action_steps=100 \
51
+ --policy.vision_backbone=resnet18 \
52
+ --policy.pretrained_backbone_weights=ResNet18_Weights.IMAGENET1K_V1 \
53
+ --policy.dim_model=512 \
54
+ --policy.n_heads=8 \
55
+ --policy.dim_feedforward=3200 \
56
+ --policy.n_encoder_layers=4 \
57
+ --policy.n_decoder_layers=1 \
58
+ --policy.use_vae=true \
59
+ --policy.latent_dim=32 \
60
+ --policy.n_vae_encoder_layers=4 \
61
+ --policy.dropout=0.1 \
62
+ --policy.kl_weight=10.0 \
63
+ --policy.optimizer_lr=0.0001 \
64
+ --policy.optimizer_weight_decay=1e-4 \
65
+ --policy.optimizer_lr_backbone=0.0001 \
66
+ --batch_size=32 \
67
+ --steps=400000 \
68
+ --save_freq=10000 \
69
+ --log_freq=200 \
70
+ --save_checkpoint=true \
71
+ --seed=1000 \
72
+ --num_workers=8 \
73
+ --output_dir=outputs/train/act-vacuum-pick \
74
+ --policy.pretrained_path=avatarrobotics/act-alpha-bot-right-resnet18-b12-20260312-1749 \
75
+ --policy.push_to_hub=true \
76
+ --policy.private=true
77
+ ============================================================
78
+
79
+ Model will be pushed to: avatarrobotics/act-alpha-bot-right-resnet18-b256-20260312-2339
80
+ [Avatar] Mixed precision: bf16
81
+
82
+ ============================================================
83
+ Training args:
84
+ mixed_precision=bf16
85
+ video_backend=pyav
86
+ prefetch_factor=4
87
+ preload=True
88
+ compile=True
89
+ backbone=resnet18
90
+ lerobot_train \
91
+ --dataset.repo_id=avatarrobotics/alpha-bot-right-2026-03-11-afternoon \
92
+ --dataset.video_backend=pyav \
93
+ --policy.repo_id=avatarrobotics/act-alpha-bot-right-resnet18-b256-20260312-2339 \
94
+ --policy.type=act \
95
+ --policy.chunk_size=100 \
96
+ --policy.n_action_steps=100 \
97
+ --policy.vision_backbone=resnet18 \
98
+ --policy.pretrained_backbone_weights=ResNet18_Weights.IMAGENET1K_V1 \
99
+ --policy.dim_model=512 \
100
+ --policy.n_heads=8 \
101
+ --policy.dim_feedforward=3200 \
102
+ --policy.n_encoder_layers=4 \
103
+ --policy.n_decoder_layers=1 \
104
+ --policy.use_vae=true \
105
+ --policy.latent_dim=32 \
106
+ --policy.n_vae_encoder_layers=4 \
107
+ --policy.dropout=0.1 \
108
+ --policy.kl_weight=10.0 \
109
+ --policy.optimizer_lr=0.0001 \
110
+ --policy.optimizer_weight_decay=1e-4 \
111
+ --policy.optimizer_lr_backbone=0.0001 \
112
+ --batch_size=32 \
113
+ --steps=400000 \
114
+ --save_freq=10000 \
115
+ --log_freq=200 \
116
+ --save_checkpoint=true \
117
+ --seed=1000 \
118
+ --num_workers=8 \
119
+ --output_dir=outputs/train/act-vacuum-pick \
120
+ --policy.pretrained_path=avatarrobotics/act-alpha-bot-right-resnet18-b12-20260312-1749 \
121
+ --policy.push_to_hub=true \
122
+ --policy.private=true
123
+ ============================================================
124
+
125
+ Model will be pushed to: avatarrobotics/act-alpha-bot-right-resnet18-b256-20260312-2339
126
+ [Avatar] Mixed precision: bf16
127
+
128
+ ============================================================
129
+ Training args:
130
+ mixed_precision=bf16
131
+ video_backend=pyav
132
+ prefetch_factor=4
133
+ preload=True
134
+ compile=True
135
+ backbone=resnet18
136
+ lerobot_train \
137
+ --dataset.repo_id=avatarrobotics/alpha-bot-right-2026-03-11-afternoon \
138
+ --dataset.video_backend=pyav \
139
+ --policy.repo_id=avatarrobotics/act-alpha-bot-right-resnet18-b256-20260312-2339 \
140
+ --policy.type=act \
141
+ --policy.chunk_size=100 \
142
+ --policy.n_action_steps=100 \
143
+ --policy.vision_backbone=resnet18 \
144
+ --policy.pretrained_backbone_weights=ResNet18_Weights.IMAGENET1K_V1 \
145
+ --policy.dim_model=512 \
146
+ --policy.n_heads=8 \
147
+ --policy.dim_feedforward=3200 \
148
+ --policy.n_encoder_layers=4 \
149
+ --policy.n_decoder_layers=1 \
150
+ --policy.use_vae=true \
151
+ --policy.latent_dim=32 \
152
+ --policy.n_vae_encoder_layers=4 \
153
+ --policy.dropout=0.1 \
154
+ --policy.kl_weight=10.0 \
155
+ --policy.optimizer_lr=0.0001 \
156
+ --policy.optimizer_weight_decay=1e-4 \
157
+ --policy.optimizer_lr_backbone=0.0001 \
158
+ --batch_size=32 \
159
+ --steps=400000 \
160
+ --save_freq=10000 \
161
+ --log_freq=200 \
162
+ --save_checkpoint=true \
163
+ --seed=1000 \
164
+ --num_workers=8 \
165
+ --output_dir=outputs/train/act-vacuum-pick \
166
+ --policy.pretrained_path=avatarrobotics/act-alpha-bot-right-resnet18-b12-20260312-1749 \
167
+ --policy.push_to_hub=true \
168
+ --policy.private=true
169
+ ============================================================
170
+
171
+ Model will be pushed to: avatarrobotics/act-alpha-bot-right-resnet18-b256-20260312-2339
172
+ [Avatar] Mixed precision: bf16
173
+
174
+ ============================================================
175
+ Training args:
176
+ Model will be pushed to: avatarrobotics/act-alpha-bot-right-resnet18-b256-20260312-2339
177
+ mixed_precision=bf16[Avatar] Mixed precision: bf16
178
+
179
+ video_backend=pyav
180
+ prefetch_factor=4
181
+ preload=True
182
+ compile=True
183
+ backbone=resnet18
184
+ lerobot_train \
185
+ --dataset.repo_id=avatarrobotics/alpha-bot-right-2026-03-11-afternoon \
186
+ --dataset.video_backend=pyav \
187
+ --policy.repo_id=avatarrobotics/act-alpha-bot-right-resnet18-b256-20260312-2339 \
188
+ --policy.type=act \
189
+ --policy.chunk_size=100 \
190
+ --policy.n_action_steps=100 \
191
+ --policy.vision_backbone=resnet18 \
192
+ --policy.pretrained_backbone_weights=ResNet18_Weights.IMAGENET1K_V1 \
193
+ --policy.dim_model=512 \
194
+ --policy.n_heads=8 \
195
+ --policy.dim_feedforward=3200 \
196
+ --policy.n_encoder_layers=4 \
197
+ --policy.n_decoder_layers=1 \
198
+ --policy.use_vae=true \
199
+ --policy.latent_dim=32 \
200
+ --policy.n_vae_encoder_layers=4 \
201
+ --policy.dropout=0.1 \
202
+ --policy.kl_weight=10.0 \
203
+ --policy.optimizer_lr=0.0001 \
204
+ --policy.optimizer_weight_decay=1e-4 \
205
+ --policy.optimizer_lr_backbone=0.0001 \
206
+ --batch_size=32 \
207
+ --steps=400000 \
208
+ --save_freq=10000 \
209
+ --log_freq=200 \
210
+ --save_checkpoint=true \
211
+ --seed=1000 \
212
+ --num_workers=8 \
213
+ --output_dir=outputs/train/act-vacuum-pick \
214
+ --policy.pretrained_path=avatarrobotics/act-alpha-bot-right-resnet18-b12-20260312-1749 \
215
+ --policy.push_to_hub=true \
216
+ --policy.private=true
217
+ ============================================================
218
+
219
+ ============================================================
220
+ Training args:
221
+
222
+ mixed_precision=bf16
223
+ video_backend=pyav
224
+ prefetch_factor=4
225
+ preload=True
226
+ compile=True
227
+ backbone=resnet18
228
+ lerobot_train \
229
+ --dataset.repo_id=avatarrobotics/alpha-bot-right-2026-03-11-afternoon \
230
+ --dataset.video_backend=pyav \
231
+ --policy.repo_id=avatarrobotics/act-alpha-bot-right-resnet18-b256-20260312-2339 \
232
+ --policy.type=act \
233
+ --policy.chunk_size=100 \
234
+ --policy.n_action_steps=100 \
235
+ --policy.vision_backbone=resnet18 \
236
+ --policy.pretrained_backbone_weights=ResNet18_Weights.IMAGENET1K_V1 \
237
+ --policy.dim_model=512 \
238
+ --policy.n_heads=8 \
239
+ --policy.dim_feedforward=3200 \
240
+ --policy.n_encoder_layers=4 \
241
+ --policy.n_decoder_layers=1 \
242
+ --policy.use_vae=true \
243
+ --policy.latent_dim=32 \
244
+ --policy.n_vae_encoder_layers=4 \
245
+ --policy.dropout=0.1 \
246
+ --policy.kl_weight=10.0 \
247
+ --policy.optimizer_lr=0.0001 \
248
+ --policy.optimizer_weight_decay=1e-4 \
249
+ --policy.optimizer_lr_backbone=0.0001 \
250
+ --batch_size=32 \
251
+ --steps=400000 \
252
+ --save_freq=10000 \
253
+ --log_freq=200 \
254
+ --save_checkpoint=true \
255
+ --seed=1000 \
256
+ --num_workers=8 \
257
+ --output_dir=outputs/train/act-vacuum-pick \
258
+ --policy.pretrained_path=avatarrobotics/act-alpha-bot-right-resnet18-b12-20260312-1749 \
259
+ --policy.push_to_hub=true \
260
+ --policy.private=true
261
+ ============================================================
262
+
263
+ Model will be pushed to: avatarrobotics/act-alpha-bot-right-resnet18-b256-20260312-2339
264
+ [Avatar] Mixed precision: bf16
265
+
266
+ ============================================================
267
+ Training args:
268
+ mixed_precision=bf16
269
+ video_backend=pyav
270
+ prefetch_factor=4
271
+ preload=True
272
+ compile=True
273
+ backbone=resnet18
274
+ lerobot_train \
275
+ --dataset.repo_id=avatarrobotics/alpha-bot-right-2026-03-11-afternoon \
276
+ --dataset.video_backend=pyav \
277
+ --policy.repo_id=avatarrobotics/act-alpha-bot-right-resnet18-b256-20260312-2339 \
278
+ --policy.type=act \
279
+ --policy.chunk_size=100 \
280
+ --policy.n_action_steps=100 \
281
+ --policy.vision_backbone=resnet18 \
282
+ --policy.pretrained_backbone_weights=ResNet18_Weights.IMAGENET1K_V1 \
283
+ --policy.dim_model=512 \
284
+ --policy.n_heads=8 \
285
+ --policy.dim_feedforward=3200 \
286
+ --policy.n_encoder_layers=4 \
287
+ --policy.n_decoder_layers=1 \
288
+ --policy.use_vae=true \
289
+ --policy.latent_dim=32 \
290
+ --policy.n_vae_encoder_layers=4 \
291
+ --policy.dropout=0.1 \
292
+ --policy.kl_weight=10.0 \
293
+ --policy.optimizer_lr=0.0001 \
294
+ --policy.optimizer_weight_decay=1e-4 \
295
+ --policy.optimizer_lr_backbone=0.0001 \
296
+ --batch_size=32 \
297
+ --steps=400000 \
298
+ --save_freq=10000 \
299
+ --log_freq=200 \
300
+ --save_checkpoint=true \
301
+ --seed=1000 \
302
+ --num_workers=8 \
303
+ --output_dir=outputs/train/act-vacuum-pick \
304
+ --policy.pretrained_path=avatarrobotics/act-alpha-bot-right-resnet18-b12-20260312-1749 \
305
+ --policy.push_to_hub=true \
306
+ --policy.private=true
307
+ ============================================================
308
+
309
+ Model will be pushed to: avatarrobotics/act-alpha-bot-right-resnet18-b256-20260312-2339
310
+ [Avatar] Mixed precision: bf16
311
+
312
+ ============================================================
313
+ Training args:
314
+ mixed_precision=bf16
315
+ video_backend=pyav
316
+ prefetch_factor=4
317
+ preload=True
318
+ compile=True
319
+ backbone=resnet18
320
+ lerobot_train \
321
+ --dataset.repo_id=avatarrobotics/alpha-bot-right-2026-03-11-afternoon \
322
+ --dataset.video_backend=pyav \
323
+ --policy.repo_id=avatarrobotics/act-alpha-bot-right-resnet18-b256-20260312-2339 \
324
+ --policy.type=act \
325
+ --policy.chunk_size=100 \
326
+ --policy.n_action_steps=100 \
327
+ --policy.vision_backbone=resnet18 \
328
+ --policy.pretrained_backbone_weights=ResNet18_Weights.IMAGENET1K_V1 \
329
+ --policy.dim_model=512 \
330
+ --policy.n_heads=8 \
331
+ --policy.dim_feedforward=3200 \
332
+ --policy.n_encoder_layers=4 \
333
+ --policy.n_decoder_layers=1 \
334
+ --policy.use_vae=true \
335
+ --policy.latent_dim=32 \
336
+ --policy.n_vae_encoder_layers=4 \
337
+ --policy.dropout=0.1 \
338
+ --policy.kl_weight=10.0 \
339
+ --policy.optimizer_lr=0.0001 \
340
+ --policy.optimizer_weight_decay=1e-4 \
341
+ --policy.optimizer_lr_backbone=0.0001 \
342
+ --batch_size=32 \
343
+ --steps=400000 \
344
+ --save_freq=10000 \
345
+ --log_freq=200 \
346
+ --save_checkpoint=true \
347
+ --seed=1000 \
348
+ --num_workers=8 \
349
+ --output_dir=outputs/train/act-vacuum-pick \
350
+ --policy.pretrained_path=avatarrobotics/act-alpha-bot-right-resnet18-b12-20260312-1749 \
351
+ --policy.push_to_hub=true \
352
+ --policy.private=true
353
+ ============================================================
354
+
355
+ [Avatar] Mixed precision: bf16
356
+
357
+ ============================================================
358
+ Training args:
359
+ mixed_precision=bf16
360
+ video_backend=pyav
361
+ prefetch_factor=4
362
+ preload=True
363
+ compile=True
364
+ backbone=resnet18
365
+ lerobot_train \
366
+ --dataset.repo_id=avatarrobotics/alpha-bot-right-2026-03-11-afternoon \
367
+ --dataset.video_backend=pyav \
368
+ --policy.repo_id=avatarrobotics/act-alpha-bot-right-resnet18-b256-20260312-2339 \
369
+ --policy.type=act \
370
+ --policy.chunk_size=100 \
371
+ --policy.n_action_steps=100 \
372
+ --policy.vision_backbone=resnet18 \
373
+ --policy.pretrained_backbone_weights=ResNet18_Weights.IMAGENET1K_V1 \
374
+ --policy.dim_model=512 \
375
+ --policy.n_heads=8 \
376
+ --policy.dim_feedforward=3200 \
377
+ --policy.n_encoder_layers=4 \
378
+ --policy.n_decoder_layers=1 \
379
+ --policy.use_vae=true \
380
+ --policy.latent_dim=32 \
381
+ --policy.n_vae_encoder_layers=4 \
382
+ --policy.dropout=0.1 \
383
+ --policy.kl_weight=10.0 \
384
+ --policy.optimizer_lr=0.0001 \
385
+ --policy.optimizer_weight_decay=1e-4 \
386
+ --policy.optimizer_lr_backbone=0.0001 \
387
+ --batch_size=32 \
388
+ --steps=400000 \
389
+ --save_freq=10000 \
390
+ --log_freq=200 \
391
+ --save_checkpoint=true \
392
+ --seed=1000 \
393
+ --num_workers=8 \
394
+ --output_dir=outputs/train/act-vacuum-pick \
395
+ --policy.pretrained_path=avatarrobotics/act-alpha-bot-right-resnet18-b12-20260312-1749 \
396
+ --policy.push_to_hub=true \
397
+ --policy.private=true
398
+ ============================================================
399
+
400
+ [Avatar] cudnn.benchmark=True, tf32=True
401
+ [Avatar] cudnn.benchmark=True, tf32=True
402
+ [Avatar] cudnn.benchmark=True, tf32=True
403
+ [Avatar] cudnn.benchmark=True, tf32=True
404
+ [Avatar] cudnn.benchmark=True, tf32=True
405
+ [Avatar] cudnn.benchmark=True, tf32=True
406
+ [Avatar] cudnn.benchmark=True, tf32=True
407
+ [Avatar] cudnn.benchmark=True, tf32=True
408
+ [Avatar] Checkpoint push enabled → avatarrobotics/act-alpha-bot-right-resnet18-b256-20260312-2339 (with per-step tags)[Avatar] Checkpoint push enabled → avatarrobotics/act-alpha-bot-right-resnet18-b256-20260312-2339 (with per-step tags)
409
+
410
+ [Avatar] torch.compile enabled (will compile on first forward pass)[Avatar] torch.compile enabled (will compile on first forward pass)
411
+
412
+ [Avatar] Overriding prefetch_factor=4[Avatar] Overriding prefetch_factor=4
413
+
414
+ [Avatar] Checkpoint push enabled → avatarrobotics/act-alpha-bot-right-resnet18-b256-20260312-2339 (with per-step tags)
415
+ [Avatar] torch.compile enabled (will compile on first forward pass)
416
+ [Avatar] Overriding prefetch_factor=4
417
+ [Avatar] Checkpoint push enabled → avatarrobotics/act-alpha-bot-right-resnet18-b256-20260312-2339 (with per-step tags)
418
+ [Avatar] torch.compile enabled (will compile on first forward pass)
419
+ [Avatar] Overriding prefetch_factor=4
420
+ [Avatar] Checkpoint push enabled → avatarrobotics/act-alpha-bot-right-resnet18-b256-20260312-2339 (with per-step tags)
421
+ [Avatar] torch.compile enabled (will compile on first forward pass)
422
+ [Avatar] Overriding prefetch_factor=4
423
+ [Avatar] Checkpoint push enabled → avatarrobotics/act-alpha-bot-right-resnet18-b256-20260312-2339 (with per-step tags)
424
+ [Avatar] torch.compile enabled (will compile on first forward pass)
425
+ [Avatar] Overriding prefetch_factor=4
426
+ [Avatar] Checkpoint push enabled → avatarrobotics/act-alpha-bot-right-resnet18-b256-20260312-2339 (with per-step tags)
427
+ [Avatar] torch.compile enabled (will compile on first forward pass)
428
+ [Avatar] Overriding prefetch_factor=4
429
+ [Avatar] Checkpoint push enabled → avatarrobotics/act-alpha-bot-right-resnet18-b256-20260312-2339 (with per-step tags)
430
+ [Avatar] torch.compile enabled (will compile on first forward pass)
431
+ [Avatar] Overriding prefetch_factor=4
432
+ WARNING:lerobot.configs.policies:Device 'None' is not available. Switching to 'cuda'.
433
+ WARNING:lerobot.configs.policies:Device 'None' is not available. Switching to 'cuda'.
434
+ WARNING:lerobot.configs.policies:Device 'None' is not available. Switching to 'cuda'.
435
+ WARNING:lerobot.configs.policies:Device 'None' is not available. Switching to 'cuda'.
436
+ WARNING:lerobot.configs.policies:Device 'None' is not available. Switching to 'cuda'.
437
+ WARNING:lerobot.configs.policies:Device 'None' is not available. Switching to 'cuda'.
438
+ WARNING:lerobot.configs.policies:Device 'None' is not available. Switching to 'cuda'.
439
+ WARNING:lerobot.configs.policies:Device 'None' is not available. Switching to 'cuda'.
440
+ INFO 2026-03-12 23:39:24 ot_train.py:195 {'batch_size': 32,
441
+ 'checkpoint_path': None,
442
+ 'dataset': {'episodes': None,
443
+ 'image_transforms': {'enable': False,
444
+ 'max_num_transforms': 3,
445
+ 'random_order': False,
446
+ 'tfs': {'affine': {'kwargs': {'degrees': [-5.0,
447
+ 5.0],
448
+ 'translate': [0.05,
449
+ 0.05]},
450
+ 'type': 'RandomAffine',
451
+ 'weight': 1.0},
452
+ 'brightness': {'kwargs': {'brightness': [0.8,
453
+ 1.2]},
454
+ 'type': 'ColorJitter',
455
+ 'weight': 1.0},
456
+ 'contrast': {'kwargs': {'contrast': [0.8,
457
+ 1.2]},
458
+ 'type': 'ColorJitter',
459
+ 'weight': 1.0},
460
+ 'hue': {'kwargs': {'hue': [-0.05,
461
+ 0.05]},
462
+ 'type': 'ColorJitter',
463
+ 'weight': 1.0},
464
+ 'saturation': {'kwargs': {'saturation': [0.5,
465
+ 1.5]},
466
+ 'type': 'ColorJitter',
467
+ 'weight': 1.0},
468
+ 'sharpness': {'kwargs': {'sharpness': [0.5,
469
+ 1.5]},
470
+ 'type': 'SharpnessJitter',
471
+ 'weight': 1.0}}},
472
+ 'repo_id': 'avatarrobotics/alpha-bot-right-2026-03-11-afternoon',
473
+ 'revision': None,
474
+ 'root': None,
475
+ 'streaming': False,
476
+ 'use_imagenet_stats': True,
477
+ 'video_backend': 'pyav'},
478
+ 'env': None,
479
+ 'eval': {'batch_size': 50, 'n_episodes': 50, 'use_async_envs': False},
480
+ 'eval_freq': 20000,
481
+ 'job_name': 'act',
482
+ 'log_freq': 200,
483
+ 'num_workers': 8,
484
+ 'optimizer': {'betas': [0.9, 0.999],
485
+ 'eps': 1e-08,
486
+ 'grad_clip_norm': 10.0,
487
+ 'lr': 0.0001,
488
+ 'type': 'adamw',
489
+ 'weight_decay': 0.0001},
490
+ 'output_dir': 'outputs/train/act-vacuum-pick',
491
+ 'peft': None,
492
+ 'policy': {'chunk_size': 100,
493
+ 'device': 'cuda',
494
+ 'dim_feedforward': 3200,
495
+ 'dim_model': 512,
496
+ 'dropout': 0.1,
497
+ 'feedforward_activation': 'relu',
498
+ 'input_features': {},
499
+ 'kl_weight': 10.0,
500
+ 'latent_dim': 32,
501
+ 'license': None,
502
+ 'n_action_steps': 100,
503
+ 'n_decoder_layers': 1,
504
+ 'n_encoder_layers': 4,
505
+ 'n_heads': 8,
506
+ 'n_obs_steps': 1,
507
+ 'n_vae_encoder_layers': 4,
508
+ 'normalization_mapping': {'ACTION': <NormalizationMode.MEAN_STD: 'MEAN_STD'>,
509
+ 'STATE': <NormalizationMode.MEAN_STD: 'MEAN_STD'>,
510
+ 'VISUAL': <NormalizationMode.MEAN_STD: 'MEAN_STD'>},
511
+ 'optimizer_lr': 0.0001,
512
+ 'optimizer_lr_backbone': 0.0001,
513
+ 'optimizer_weight_decay': 0.0001,
514
+ 'output_features': {},
515
+ 'pre_norm': False,
516
+ 'pretrained_backbone_weights': 'ResNet18_Weights.IMAGENET1K_V1',
517
+ 'pretrained_path': 'avatarrobotics/act-alpha-bot-right-resnet18-b12-20260312-1749',
518
+ 'private': True,
519
+ 'push_to_hub': True,
520
+ 'replace_final_stride_with_dilation': False,
521
+ 'repo_id': 'avatarrobotics/act-alpha-bot-right-resnet18-b256-20260312-2339',
522
+ 'tags': None,
523
+ 'temporal_ensemble_coeff': None,
524
+ 'type': 'act',
525
+ 'use_amp': False,
526
+ 'use_peft': False,
527
+ 'use_vae': True,
528
+ 'vision_backbone': 'resnet18'},
529
+ 'rabc_epsilon': 1e-06,
530
+ 'rabc_head_mode': 'sparse',
531
+ 'rabc_kappa': 0.01,
532
+ 'rabc_progress_path': None,
533
+ 'rename_map': {},
534
+ 'resume': False,
535
+ 'save_checkpoint': True,
536
+ 'save_freq': 10000,
537
+ 'scheduler': None,
538
+ 'seed': 1000,
539
+ 'steps': 400000,
540
+ 'tolerance_s': 0.0001,
541
+ 'use_policy_training_preset': True,
542
+ 'use_rabc': False,
543
+ 'wandb': {'disable_artifact': False,
544
+ 'enable': False,
545
+ 'entity': None,
546
+ 'mode': None,
547
+ 'notes': None,
548
+ 'project': 'lerobot',
549
+ 'run_id': None}}
550
+ INFO 2026-03-12 23:39:24 ot_train.py:203 Logs will be saved locally.
551
+ INFO 2026-03-12 23:39:24 ot_train.py:215 Creating dataset
552
+
553
+
554
+ [Avatar] Set video tolerance_s=0.02
555
+ [Avatar] Camera transforms: arm=right, target=480x480
556
+ [Avatar] Using /dev/shm (RAM-backed, 439GB free)
557
+ [Avatar] Pre-decoding videos to JPEG: /dev/shm/avatar_frames
558
+ Extracting 938 videos with ffmpeg...
559
+ [100/938] 0 errors, 4s
560
+ [200/938] 0 errors, 8s
561
+ [300/938] 0 errors, 11s
562
+ [400/938] 0 errors, 15s
563
+ [500/938] 0 errors, 19s
564
+ [600/938] 0 errors, 23s
565
+ [700/938] 0 errors, 28s
566
+ [800/938] 0 errors, 33s
567
+ [900/938] 0 errors, 38s
568
+ [938/938] 0 errors, 39s
569
+ Done: 938/938 videos OK, 270,246 JPEGs, 18.8GB, 41s
570
+ [Avatar] Patched decode_video_frames (disk cache)
571
+ [Avatar] Preload complete (44s total)
572
+ [Avatar] Rank 0: signaled 7 other ranks
573
+ [Avatar] Installed corrupt-frame retry handler
574
+ NCCL version 2.26.2+cuda12.2
575
+ INFO 2026-03-12 23:42:44 ot_train.py:233 Creating policy
576
+ [Avatar] Set video tolerance_s=0.02
577
+ [Avatar] Set video tolerance_s=0.02
578
+ [Avatar] Camera transforms: arm=right, target=480x480
579
+ [Avatar] Rank 1: waiting for preload...
580
+ [Avatar] Rank 1: preload ready (0s)
581
+ [Avatar] Patched decode_video_frames (disk cache)
582
+ [Avatar] Installed corrupt-frame retry handler
583
+ [Avatar] Camera transforms: arm=right, target=480x480
584
+ [Avatar] Rank 5: waiting for preload...
585
+ [Avatar] Rank 5: preload ready (0s)
586
+ [Avatar] Patched decode_video_frames (disk cache)
587
+ [Avatar] Installed corrupt-frame retry handler
588
+ [Avatar] Set video tolerance_s=0.02
589
+ [Avatar] Camera transforms: arm=right, target=480x480
590
+ [Avatar] Rank 2: waiting for preload...
591
+ [Avatar] Rank 2: preload ready (0s)
592
+ [Avatar] Patched decode_video_frames (disk cache)
593
+ [Avatar] Installed corrupt-frame retry handler
594
+ [Avatar] Set video tolerance_s=0.02
595
+ [Avatar] Camera transforms: arm=right, target=480x480
596
+ [Avatar] Rank 7: waiting for preload...
597
+ [Avatar] Rank 7: preload ready (0s)
598
+ [Avatar] Patched decode_video_frames (disk cache)
599
+ [Avatar] Installed corrupt-frame retry handler
600
+ [Avatar] Set video tolerance_s=0.02
601
+ [Avatar] Camera transforms: arm=right, target=480x480
602
+ [Avatar] Rank 3: waiting for preload...
603
+ [Avatar] Rank 3: preload ready (0s)
604
+ [Avatar] Patched decode_video_frames (disk cache)
605
+ [Avatar] Installed corrupt-frame retry handler
606
+ [Avatar] Set video tolerance_s=0.02
607
+ [Avatar] Camera transforms: arm=right, target=480x480
608
+ [Avatar] Rank 4: waiting for preload...
609
+ [Avatar] Rank 4: preload ready (0s)
610
+ [Avatar] Patched decode_video_frames (disk cache)
611
+ [Avatar] Installed corrupt-frame retry handler
612
+ [Avatar] Set video tolerance_s=0.02
613
+ [Avatar] Camera transforms: arm=right, target=480x480
614
+ [Avatar] Rank 6: waiting for preload...
615
+ [Avatar] Rank 6: preload ready (0s)
616
+ [Avatar] Patched decode_video_frames (disk cache)
617
+ [Avatar] Installed corrupt-frame retry handler
618
+ Downloading: "https://download.pytorch.org/models/resnet18-f37072fd.pth" to /root/.cache/torch/hub/checkpoints/resnet18-f37072fd.pth
619
+ Downloading: "https://download.pytorch.org/models/resnet18-f37072fd.pth" to /root/.cache/torch/hub/checkpoints/resnet18-f37072fd.pth
620
+ Downloading: "https://download.pytorch.org/models/resnet18-f37072fd.pth" to /root/.cache/torch/hub/checkpoints/resnet18-f37072fd.pth
621
+ Downloading: "https://download.pytorch.org/models/resnet18-f37072fd.pth" to /root/.cache/torch/hub/checkpoints/resnet18-f37072fd.pth
622
+ Downloading: "https://download.pytorch.org/models/resnet18-f37072fd.pth" to /root/.cache/torch/hub/checkpoints/resnet18-f37072fd.pth
623
+ Downloading: "https://download.pytorch.org/models/resnet18-f37072fd.pth" to /root/.cache/torch/hub/checkpoints/resnet18-f37072fd.pth
624
+ Downloading: "https://download.pytorch.org/models/resnet18-f37072fd.pth" to /root/.cache/torch/hub/checkpoints/resnet18-f37072fd.pth
625
+ Downloading: "https://download.pytorch.org/models/resnet18-f37072fd.pth" to /root/.cache/torch/hub/checkpoints/resnet18-f37072fd.pth
626
+
627
  0%| | 0.00/44.7M [00:00<?, ?B/s]
628
  0%| | 0.00/44.7M [00:00<?, ?B/s]
629
  0%| | 0.00/44.7M [00:00<?, ?B/s]
630
  15%|█▌ | 6.75M/44.7M [00:00<00:00, 46.8MB/s]
631
  64%|██████▍ | 28.8M/44.7M [00:00<00:00, 101MB/s]
632
  19%|█▉ | 8.50M/44.7M [00:00<00:01, 28.8MB/s]
633
+
634
  39%|███▉ | 17.5M/44.7M [00:00<00:00, 48.0MB/s]
635
  58%|█████▊ | 26.0M/44.7M [00:00<00:00, 53.0MB/s]
636
  76%|███████▌ | 33.9M/44.7M [00:00<00:00, 58.3MB/s]
637
  15%|█▌ | 6.88M/44.7M [00:00<00:04, 9.60MB/s]
638
+
639
  24%|██▎ | 10.5M/44.7M [00:00<00:02, 13.1MB/s]
640
  0%| | 0.00/44.7M [00:00<?, ?B/s]
641
  5%|▌ | 2.25M/44.7M [00:00<00:01, 22.9MB/s]
642
  39%|███▉ | 17.6M/44.7M [00:01<00:01, 23.3MB/s][Avatar] Compiling policy with torch.compile (mode=reduce-overhead)...
643
+
644
  11%|█ | 4.75M/44.7M [00:00<00:01, 24.7MB/s]
645
  48%|████▊ | 21.5M/44.7M [00:01<00:01, 18.9MB/s]
646
  64%|██████▎ | 28.4M/44.7M [00:01<00:00, 27.0MB/s][Avatar] Compiling policy with torch.compile (mode=reduce-overhead)...
647
+
648
  72%|███████▏ | 32.4M/44.7M [00:01<00:00, 23.1MB/s]
649
  80%|███████▉ | 35.6M/44.7M [00:01<00:00, 21.9MB/s][Avatar] Policy compiled (first step will be slow)
650
+
651
  16%|█▌ | 7.12M/44.7M [00:01<00:07, 5.59MB/s]
652
  22%|██▏ | 9.62M/44.7M [00:01<00:04, 8.20MB/s]
653
  27%|██▋ | 11.9M/44.7M [00:01<00:04, 7.63MB/s][Avatar] Policy compiled (first step will be slow)
654
+
655
  33%|███▎ | 14.9M/44.7M [00:01<00:02, 10.7MB/s]
656
  86%|████████▌ | 38.4M/44.7M [00:02<00:00, 10.8MB/s]
657
  39%|███▉ | 17.6M/44.7M [00:01<00:02, 13.4MB/s]
658
  45%|████▌ | 20.1M/44.7M [00:01<00:01, 13.8MB/s]
659
  53%|█████▎ | 23.5M/44.7M [00:02<00:01, 17.8MB/s]
660
  59%|█████▉ | 26.4M/44.7M [00:02<00:01, 18.6MB/s]
661
  67%|██████▋ | 30.1M/44.7M [00:02<00:00, 23.1MB/s]
662
  82%|████████▏ | 36.8M/44.7M [00:02<00:00, 33.0MB/s]
663
  90%|█████████ | 40.4M/44.7M [00:03<00:00, 6.67MB/s]
664
+
665
  91%|█████████ | 40.5M/44.7M [00:02<00:00, 21.0MB/s]
666
  0%| | 0.00/44.7M [00:00<?, ?B/s]
667
  0%| | 0.00/44.7M [00:00<?, ?B/s]
668
  97%|█████████▋| 43.4M/44.7M [00:03<00:00, 16.2MB/s]
669
  1%| | 512k/44.7M [00:00<00:09, 4.68MB/s]
670
+
671
  2%|▏ | 896k/44.7M [00:00<00:16, 2.83MB/s]
672
  2%|▏ | 1.00M/44.7M [00:00<00:22, 2.02MB/s]
673
  3%|▎ | 1.38M/44.7M [00:00<00:27, 1.67MB/s]
674
  6%|▋ | 2.88M/44.7M [00:00<00:10, 4.33MB/s][Avatar] Compiling policy with torch.compile (mode=reduce-overhead)...
675
+
676
  12%|█▏ | 5.25M/44.7M [00:00<00:04, 8.67MB/s]
677
  17%|█▋ | 7.75M/44.7M [00:01<00:03, 12.7MB/s]
678
  0%| | 0.00/44.7M [00:00<?, ?B/s]
679
  0%| | 128k/44.7M [00:00<00:41, 1.11MB/s]
680
  21%|██▏ | 9.50M/44.7M [00:01<00:02, 12.4MB/s]
681
  11%|█ | 5.00M/44.7M [00:00<00:01, 27.5MB/s]
682
  3%|▎ | 1.25M/44.7M [00:01<00:56, 807kB/s]
683
  4%|▍ | 1.75M/44.7M [00:01<00:36, 1.24MB/s]
684
  19%|█▉ | 8.38M/44.7M [00:00<00:02, 14.8MB/s][Avatar] Policy compiled (first step will be slow)
685
+
686
  25%|██▍ | 11.0M/44.7M [00:01<00:05, 6.67MB/s]
687
  35%|███▍ | 15.5M/44.7M [00:00<00:01, 25.7MB/s]
688
  27%|██▋ | 12.1M/44.7M [00:01<00:04, 7.34MB/s]
689
  30%|███ | 13.5M/44.7M [00:01<00:03, 8.40MB/s]
690
  33%|███▎ | 14.8M/44.7M [00:02<00:03, 8.35MB/s]
691
  36%|███▌ | 16.0M/44.7M [00:02<00:03, 8.51MB/s]
692
  40%|████ | 17.9M/44.7M [00:02<00:02, 9.68MB/s]
693
  43%|████▎ | 19.2M/44.7M [00:02<00:02, 10.6MB/s]
694
  52%|█████▏ | 23.4M/44.7M [00:02<00:01, 13.3MB/s]
695
  58%|█████▊ | 26.1M/44.7M [00:02<00:01, 16.3MB/s]
696
  62%|██████▏ | 27.9M/44.7M [00:03<00:01, 15.9MB/s]
697
  67%|██████▋ | 30.0M/44.7M [00:03<00:00, 17.3MB/s]
698
  71%|███████▏ | 31.9M/44.7M [00:03<00:00, 16.0MB/s]
699
  75%|███████▌ | 33.5M/44.7M [00:03<00:00, 15.0MB/s]
700
  79%|███████��� | 35.2M/44.7M [00:03<00:00, 15.7MB/s]
701
  83%|████████▎ | 37.1M/44.7M [00:03<00:00, 16.7MB/s]
702
  87%|████████▋ | 38.9M/44.7M [00:03<00:00, 15.8MB/s]
703
  91%|█████████ | 40.5M/44.7M [00:03<00:00, 16.0MB/s]
704
  95%|█████████▍| 42.2M/44.7M [00:03<00:00, 16.6MB/s]
705
  98%|█████████▊| 43.9M/44.7M [00:04<00:00, 15.2MB/s]
706
+ [Avatar] Compiling policy with torch.compile (mode=reduce-overhead)...
707
+ [Avatar] Compiling policy with torch.compile (mode=reduce-overhead)...
708
+ [Avatar] Policy compiled (first step will be slow)
709
+
710
  5%|▍ | 2.12M/44.7M [00:06<03:26, 216kB/s]
711
  6%|▌ | 2.62M/44.7M [00:06<02:11, 334kB/s][Avatar] Policy compiled (first step will be slow)
712
+
713
  8%|▊ | 3.38M/44.7M [00:06<01:14, 578kB/s]
714
  10%|▉ | 4.25M/44.7M [00:06<00:44, 960kB/s]
715
  13%|█▎ | 5.88M/44.7M [00:07<00:21, 1.92MB/s]
716
  17%|█▋ | 7.62M/44.7M [00:07<00:12, 3.20MB/s]
717
  20%|█▉ | 8.88M/44.7M [00:07<00:11, 3.21MB/s]
718
  22%|██▏ | 9.88M/44.7M [00:07<00:11, 3.27MB/s]
719
  27%|██▋ | 12.0M/44.7M [00:08<00:06, 5.26MB/s]
720
  42%|████▏ | 18.6M/44.7M [00:07<00:15, 1.80MB/s]
721
  30%|██▉ | 13.2M/44.7M [00:08<00:05, 5.57MB/s]
722
  49%|████▉ | 21.9M/44.7M [00:07<00:09, 2.48MB/s]
723
  55%|█████▍ | 24.4M/44.7M [00:07<00:06, 3.15MB/s]
724
  60%|█████▉ | 26.8M/44.7M [00:07<00:04, 3.96MB/s]
725
  32%|███▏ | 14.2M/44.7M [00:08<00:06, 4.67MB/s]
726
  35%|███▍ | 15.6M/44.7M [00:08<00:05, 5.54MB/s]
727
  37%|███▋ | 16.5M/44.7M [00:08<00:06, 4.72MB/s]
728
  39%|███▊ | 17.2M/44.7M [00:09<00:07, 4.08MB/s]
729
  65%|██████▍ | 29.0M/44.7M [00:08<00:04, 3.54MB/s]
730
  46%|████▌ | 20.6M/44.7M [00:09<00:03, 8.01MB/s]
731
  50%|█████ | 22.4M/44.7M [00:09<00:02, 9.56MB/s]
732
  69%|██████▉ | 30.9M/44.7M [00:08<00:03, 4.07MB/s]
733
  80%|████████ | 35.8M/44.7M [00:08<00:01, 7.02MB/s]
734
  88%|████████▊ | 39.1M/44.7M [00:08<00:00, 9.13MB/s]
735
  0%| | 0.00/44.7M [00:00<?, ?B/s]
736
  3%|▎ | 1.50M/44.7M [00:00<00:02, 15.7MB/s]
737
  53%|█████▎ | 23.8M/44.7M [00:10<00:05, 4.09MB/s]
738
  58%|█████▊ | 25.8M/44.7M [00:10<00:03, 5.64MB/s]
739
  60%|██████ | 27.0M/44.7M [00:10<00:02, 6.50MB/s]
740
  93%|█████████▎| 41.5M/44.7M [00:09<00:00, 5.74MB/s]
741
+
742
  66%|██████▋ | 29.6M/44.7M [00:11<00:02, 6.81MB/s]
743
  12%|█▏ | 5.38M/44.7M [00:01<00:13, 2.98MB/s]
744
  69%|██████▉ | 30.8M/44.7M [00:14<00:09, 1.53MB/s]
745
  71%|███████ | 31.6M/44.7M [00:14<00:08, 1.68MB/s]
746
  72%|███████▏ | 32.2M/44.7M [00:15<00:08, 1.58MB/s]
747
  73%|███████▎ | 32.8M/44.7M [00:15<00:07, 1.73MB/s][Avatar] Compiling policy with torch.compile (mode=reduce-overhead)...
748
+
749
  74%|███████▍ | 33.2M/44.7M [00:16<00:10, 1.13MB/s]
750
  75%|███████▌ | 33.6M/44.7M [00:16<00:10, 1.13MB/s][Avatar] Policy compiled (first step will be slow)
751
+
752
  76%|███████▋ | 34.1M/44.7M [00:17<00:09, 1.17MB/s]
753
  79%|███████▉ | 35.2M/44.7M [00:17<00:05, 1.90MB/s]
754
  83%|████████▎ | 37.0M/44.7M [00:17<00:02, 3.25MB/s]
755
  87%|████████▋ | 39.0M/44.7M [00:17<00:01, 5.00MB/s]
756
  92%|█████████▏| 40.9M/44.7M [00:17<00:00, 6.89MB/s]
757
  98%|█████████▊| 43.6M/44.7M [00:17<00:00, 10.2MB/s]
758
+ [Avatar] Compiling policy with torch.compile (mode=reduce-overhead)...
759
+ [Avatar] Policy compiled (first step will be slow)
760
+
761
  15%|█▍ | 6.62M/44.7M [00:15<00:13, 2.98MB/s]
762
  15%|█▌ | 6.75M/44.7M [03:32<27:44, 23.9kB/s]
763
  17%|█▋ | 7.38M/44.7M [03:32<23:02, 28.3kB/s]
764
  18%|█▊ | 7.88M/44.7M [03:45<22:43, 28.3kB/s]
765
  18%|█▊ | 8.00M/44.7M [04:14<26:02, 24.6kB/s]
766
  18%|█▊ | 8.25M/44.7M [04:14<23:26, 27.1kB/s]
767
  20%|█▉ | 8.75M/44.7M [04:14<18:16, 34.3kB/s]
768
  20%|██ | 9.12M/44.7M [04:15<14:47, 42.0kB/s]
769
  21%|██▏ | 9.50M/44.7M [04:15<11:38, 52.8kB/s]
770
  21%|██▏ | 9.50M/44.7M [04:30<11:38, 52.8kB/s]
771
  22%|██▏ | 9.62M/44.7M [04:36<19:29, 31.4kB/s]
772
  22%|██▏ | 9.62M/44.7M [04:50<19:29, 31.4kB/s][rank6]:[E312 23:52:47.560128968 ProcessGroupNCCL.cpp:632] [Rank 6] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=2, OpType=ALLREDUCE, NumelIn=1, NumelOut=1, Timeout(ms)=600000) ran for 600082 milliseconds before timing out.
773
+ [rank6]:[E312 23:52:47.560612504 ProcessGroupNCCL.cpp:2271] [PG ID 0 PG GUID 0(default_pg) Rank 6] failure detected by watchdog at work sequence id: 2 PG status: last enqueued work: 2, last completed work: 1
774
+ [rank6]:[E312 23:52:47.560626339 ProcessGroupNCCL.cpp:670] Stack trace of the failed collective not found, potentially because FlightRecorder is disabled. You can enable it by setting TORCH_NCCL_TRACE_BUFFER_SIZE to a non-zero value.
775
+ [rank6]:[E312 23:52:47.560674566 ProcessGroupNCCL.cpp:2106] [PG ID 0 PG GUID 0(default_pg) Rank 6] First PG on this rank to signal dumping.
776
+ [rank2]:[E312 23:52:47.027303888 ProcessGroupNCCL.cpp:632] [Rank 2] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=2, OpType=ALLREDUCE, NumelIn=1, NumelOut=1, Timeout(ms)=600000) ran for 600048 milliseconds before timing out.
777
+ [rank2]:[E312 23:52:47.027753041 ProcessGroupNCCL.cpp:2271] [PG ID 0 PG GUID 0(default_pg) Rank 2] failure detected by watchdog at work sequence id: 2 PG status: last enqueued work: 2, last completed work: 1
778
+ [rank2]:[E312 23:52:47.027765739 ProcessGroupNCCL.cpp:670] Stack trace of the failed collective not found, potentially because FlightRecorder is disabled. You can enable it by setting TORCH_NCCL_TRACE_BUFFER_SIZE to a non-zero value.
779
+ [rank2]:[E312 23:52:47.027820756 ProcessGroupNCCL.cpp:2106] [PG ID 0 PG GUID 0(default_pg) Rank 2] First PG on this rank to signal dumping.
780
+ [rank0]:[E312 23:52:47.125792343 ProcessGroupNCCL.cpp:1685] [PG ID 0 PG GUID 0(default_pg) Rank 0] Observed flight recorder dump signal from another rank via TCPStore.
781
+ [rank0]:[E312 23:52:47.125988125 ProcessGroupNCCL.cpp:1746] [PG ID 0 PG GUID 0(default_pg) Rank 0] Received a dump signal due to a collective timeout from rank 2 and we will try our best to dump the debug info. Last enqueued NCCL work: 2, last completed NCCL work: 1.This is most likely caused by incorrect usages of collectives, e.g., wrong sizes used across ranks, the order of collectives is not same for all ranks or the scheduled collective, for some reason, didn't run. Additionally, this can be caused by GIL deadlock or other reasons such as network errors or bugs in the communications library (e.g. NCCL), etc.
782
+ [rank0]:[E312 23:52:47.126227440 ProcessGroupNCCL.cpp:1536] [PG ID 0 PG GUID 0(default_pg) Rank 0] ProcessGroupNCCL preparing to dump debug info. Include stack trace: 1
783
+ [rank2]:[E312 23:52:48.331824803 ProcessGroupNCCL.cpp:1746] [PG ID 0 PG GUID 0(default_pg) Rank 2] Received a dump signal due to a collective timeout from this local rank and we will try our best to dump the debug info. Last enqueued NCCL work: 2, last completed NCCL work: 1.This is most likely caused by incorrect usages of collectives, e.g., wrong sizes used across ranks, the order of collectives is not same for all ranks or the scheduled collective, for some reason, didn't run. Additionally, this can be caused by GIL deadlock or other reasons such as network errors or bugs in the communications library (e.g. NCCL), etc.
784
+ [rank5]:[E312 23:52:48.331852659 ProcessGroupNCCL.cpp:1685] [PG ID 0 PG GUID 0(default_pg) Rank 5] Observed flight recorder dump signal from another rank via TCPStore.
785
+ [rank7]:[E312 23:52:48.331951518 ProcessGroupNCCL.cpp:1685] [PG ID 0 PG GUID 0(default_pg) Rank 7] Observed flight recorder dump signal from another rank via TCPStore.
786
+ [rank3]:[E312 23:52:48.331961172 ProcessGroupNCCL.cpp:1685] [PG ID 0 PG GUID 0(default_pg) Rank 3] Observed flight recorder dump signal from another rank via TCPStore.
787
+ [rank2]:[E312 23:52:48.332188719 ProcessGroupNCCL.cpp:1536] [PG ID 0 PG GUID 0(default_pg) Rank 2] ProcessGroupNCCL preparing to dump debug info. Include stack trace: 1
788
+ [rank7]:[E312 23:52:48.332206972 ProcessGroupNCCL.cpp:1746] [PG ID 0 PG GUID 0(default_pg) Rank 7] Received a dump signal due to a collective timeout from rank 2 and we will try our best to dump the debug info. Last enqueued NCCL work: 2, last completed NCCL work: 1.This is most likely caused by incorrect usages of collectives, e.g., wrong sizes used across ranks, the order of collectives is not same for all ranks or the scheduled collective, for some reason, didn't run. Additionally, this can be caused by GIL deadlock or other reasons such as network errors or bugs in the communications library (e.g. NCCL), etc.
789
+ [rank5]:[E312 23:52:48.332267062 ProcessGroupNCCL.cpp:1746] [PG ID 0 PG GUID 0(default_pg) Rank 5] Received a dump signal due to a collective timeout from rank 2 and we will try our best to dump the debug info. Last enqueued NCCL work: 1, last completed NCCL work: 1.This is most likely caused by incorrect usages of collectives, e.g., wrong sizes used across ranks, the order of collectives is not same for all ranks or the scheduled collective, for some reason, didn't run. Additionally, this can be caused by GIL deadlock or other reasons such as network errors or bugs in the communications library (e.g. NCCL), etc.
790
+ [rank3]:[E312 23:52:48.332276814 ProcessGroupNCCL.cpp:1746] [PG ID 0 PG GUID 0(default_pg) Rank 3] Received a dump signal due to a collective timeout from rank 2 and we will try our best to dump the debug info. Last enqueued NCCL work: 2, last completed NCCL work: 1.This is most likely caused by incorrect usages of collectives, e.g., wrong sizes used across ranks, the order of collectives is not same for all ranks or the scheduled collective, for some reason, didn't run. Additionally, this can be caused by GIL deadlock or other reasons such as network errors or bugs in the communications library (e.g. NCCL), etc.
791
+ [rank7]:[E312 23:52:48.332386105 ProcessGroupNCCL.cpp:1536] [PG ID 0 PG GUID 0(default_pg) Rank 7] ProcessGroupNCCL preparing to dump debug info. Include stack trace: 1
792
+ [rank3]:[E312 23:52:48.332513600 ProcessGroupNCCL.cpp:1536] [PG ID 0 PG GUID 0(default_pg) Rank 3] ProcessGroupNCCL preparing to dump debug info. Include stack trace: 1
793
+ [rank5]:[E312 23:52:48.332525100 ProcessGroupNCCL.cpp:1536] [PG ID 0 PG GUID 0(default_pg) Rank 5] ProcessGroupNCCL preparing to dump debug info. Include stack trace: 1
794
+ [rank6]:[E312 23:52:48.334694742 ProcessGroupNCCL.cpp:1746] [PG ID 0 PG GUID 0(default_pg) Rank 6] Received a dump signal due to a collective timeout from this local rank and we will try our best to dump the debug info. Last enqueued NCCL work: 2, last completed NCCL work: 1.This is most likely caused by incorrect usages of collectives, e.g., wrong sizes used across ranks, the order of collectives is not same for all ranks or the scheduled collective, for some reason, didn't run. Additionally, this can be caused by GIL deadlock or other reasons such as network errors or bugs in the communications library (e.g. NCCL), etc.
795
+ [rank1]:[E312 23:52:48.334735312 ProcessGroupNCCL.cpp:1685] [PG ID 0 PG GUID 0(default_pg) Rank 1] Observed flight recorder dump signal from another rank via TCPStore.
796
+ [rank4]:[E312 23:52:48.334809414 ProcessGroupNCCL.cpp:1685] [PG ID 0 PG GUID 0(default_pg) Rank 4] Observed flight recorder dump signal from another rank via TCPStore.
797
+ [rank1]:[E312 23:52:48.334879435 ProcessGroupNCCL.cpp:1746] [PG ID 0 PG GUID 0(default_pg) Rank 1] Received a dump signal due to a collective timeout from rank 2 and we will try our best to dump the debug info. Last enqueued NCCL work: 2, last completed NCCL work: 1.This is most likely caused by incorrect usages of collectives, e.g., wrong sizes used across ranks, the order of collectives is not same for all ranks or the scheduled collective, for some reason, didn't run. Additionally, this can be caused by GIL deadlock or other reasons such as network errors or bugs in the communications library (e.g. NCCL), etc.
798
+ [rank6]:[E312 23:52:48.334872716 ProcessGroupNCCL.cpp:1536] [PG ID 0 PG GUID 0(default_pg) Rank 6] ProcessGroupNCCL preparing to dump debug info. Include stack trace: 1
799
+ [rank4]:[E312 23:52:48.335026171 ProcessGroupNCCL.cpp:1746] [PG ID 0 PG GUID 0(default_pg) Rank 4] Received a dump signal due to a collective timeout from rank 2 and we will try our best to dump the debug info. Last enqueued NCCL work: 2, last completed NCCL work: 1.This is most likely caused by incorrect usages of collectives, e.g., wrong sizes used across ranks, the order of collectives is not same for all ranks or the scheduled collective, for some reason, didn't run. Additionally, this can be caused by GIL deadlock or other reasons such as network errors or bugs in the communications library (e.g. NCCL), etc.
800
+ [rank1]:[E312 23:52:48.335115822 ProcessGroupNCCL.cpp:1536] [PG ID 0 PG GUID 0(default_pg) Rank 1] ProcessGroupNCCL preparing to dump debug info. Include stack trace: 1
801
+ [rank4]:[E312 23:52:48.335325252 ProcessGroupNCCL.cpp:1536] [PG ID 0 PG GUID 0(default_pg) Rank 4] ProcessGroupNCCL preparing to dump debug info. Include stack trace: 1
802
+ INFO 2026-03-12 23:52:48 ot_train.py:288 Creating optimizer and scheduler
803
+ INFO 2026-03-12 23:52:48 ot_train.py:323 Output dir: outputs/train/act-vacuum-pick
804
+ INFO 2026-03-12 23:52:48 ot_train.py:330 cfg.steps=400000 (400K)
805
+ INFO 2026-03-12 23:52:48 ot_train.py:331 dataset.num_frames=134890 (135K)
806
+ INFO 2026-03-12 23:52:48 ot_train.py:332 dataset.num_episodes=469
807
+ INFO 2026-03-12 23:52:48 ot_train.py:335 Effective batch size: 32 x 8 = 256
808
+ INFO 2026-03-12 23:52:48 ot_train.py:336 num_learnable_params=51601288 (52M)
809
+ INFO 2026-03-12 23:52:48 ot_train.py:337 num_total_params=51601288 (52M)
810
+ [rank2]:[E312 23:52:48.861853434 ProcessGroupNCCL.cpp:684] [Rank 2] Some NCCL operations have failed or timed out. Due to the asynchronous nature of CUDA kernels, subsequent GPU operations might run on corrupted/incomplete data.
811
+ [rank2]:[E312 23:52:48.861876278 ProcessGroupNCCL.cpp:698] [Rank 2] To avoid data inconsistency, we are taking the entire process down.
812
+ [rank2]:[E312 23:52:48.863210237 ProcessGroupNCCL.cpp:1899] [PG ID 0 PG GUID 0(default_pg) Rank 2] Process group watchdog thread terminated with exception: [Rank 2] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=2, OpType=ALLREDUCE, NumelIn=1, NumelOut=1, Timeout(ms)=600000) ran for 600048 milliseconds before timing out.
813
+ Exception raised from checkTimeout at /pytorch/torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:635 (most recent call first):
814
+ frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> >) + 0x98 (0x790fc69785e8 in /usr/local/lib/python3.11/dist-packages/torch/lib/libc10.so)
815
+ frame #1: c10d::ProcessGroupNCCL::WorkNCCL::checkTimeout(std::optional<std::chrono::duration<long, std::ratio<1l, 1000l> > >) + 0x23d (0x790f713cba6d in /usr/local/lib/python3.11/dist-packages/torch/lib/libtorch_cuda.so)
816
+ frame #2: c10d::ProcessGroupNCCL::watchdogHandler() + 0xc80 (0x790f713cd7f0 in /usr/local/lib/python3.11/dist-packages/torch/lib/libtorch_cuda.so)
817
+ frame #3: c10d::ProcessGroupNCCL::ncclCommWatchdog() + 0x14d (0x790f713ceefd in /usr/local/lib/python3.11/dist-packages/torch/lib/libtorch_cuda.so)
818
+ frame #4: <unknown function> + 0xdc253 (0x790f611d8253 in /lib/x86_64-linux-gnu/libstdc++.so.6)
819
+ frame #5: <unknown function> + 0x94ac3 (0x790fc7850ac3 in /lib/x86_64-linux-gnu/libc.so.6)
820
+ frame #6: clone + 0x44 (0x790fc78e1a04 in /lib/x86_64-linux-gnu/libc.so.6)
821
+
822
+ terminate called after throwing an instance of 'c10::DistBackendError'
823
+ what(): [PG ID 0 PG GUID 0(default_pg) Rank 2] Process group watchdog thread terminated with exception: [Rank 2] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=2, OpType=ALLREDUCE, NumelIn=1, NumelOut=1, Timeout(ms)=600000) ran for 600048 milliseconds before timing out.
824
+ Exception raised from checkTimeout at /pytorch/torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:635 (most recent call first):
825
+ frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> >) + 0x98 (0x790fc69785e8 in /usr/local/lib/python3.11/dist-packages/torch/lib/libc10.so)
826
+ frame #1: c10d::ProcessGroupNCCL::WorkNCCL::checkTimeout(std::optional<std::chrono::duration<long, std::ratio<1l, 1000l> > >) + 0x23d (0x790f713cba6d in /usr/local/lib/python3.11/dist-packages/torch/lib/libtorch_cuda.so)
827
+ frame #2: c10d::ProcessGroupNCCL::watchdogHandler() + 0xc80 (0x790f713cd7f0 in /usr/local/lib/python3.11/dist-packages/torch/lib/libtorch_cuda.so)
828
+ frame #3: c10d::ProcessGroupNCCL::ncclCommWatchdog() + 0x14d (0x790f713ceefd in /usr/local/lib/python3.11/dist-packages/torch/lib/libtorch_cuda.so)
829
+ frame #4: <unknown function> + 0xdc253 (0x790f611d8253 in /lib/x86_64-linux-gnu/libstdc++.so.6)
830
+ frame #5: <unknown function> + 0x94ac3 (0x790fc7850ac3 in /lib/x86_64-linux-gnu/libc.so.6)
831
+ frame #6: clone + 0x44 (0x790fc78e1a04 in /lib/x86_64-linux-gnu/libc.so.6)
832
+
833
+ Exception raised from ncclCommWatchdog at /pytorch/torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1905 (most recent call first):
834
+ frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> >) + 0x98 (0x790fc69785e8 in /usr/local/lib/python3.11/dist-packages/torch/lib/libc10.so)
835
+ frame #1: <unknown function> + 0x11b4abe (0x790f7139dabe in /usr/local/lib/python3.11/dist-packages/torch/lib/libtorch_cuda.so)
836
+ frame #2: <unknown function> + 0xe07bed (0x790f70ff0bed in /usr/local/lib/python3.11/dist-packages/torch/lib/libtorch_cuda.so)
837
+ frame #3: <unknown function> + 0xdc253 (0x790f611d8253 in /lib/x86_64-linux-gnu/libstdc++.so.6)
838
+ frame #4: <unknown function> + 0x94ac3 (0x790fc7850ac3 in /lib/x86_64-linux-gnu/libc.so.6)
839
+ frame #5: clone + 0x44 (0x790fc78e1a04 in /lib/x86_64-linux-gnu/libc.so.6)
840
+
841
+ [rank6]:[E312 23:52:48.865788709 ProcessGroupNCCL.cpp:684] [Rank 6] Some NCCL operations have failed or timed out. Due to the asynchronous nature of CUDA kernels, subsequent GPU operations might run on corrupted/incomplete data.
842
+ [rank6]:[E312 23:52:48.865815177 ProcessGroupNCCL.cpp:698] [Rank 6] To avoid data inconsistency, we are taking the entire process down.
843
+ [rank6]:[E312 23:52:48.867172243 ProcessGroupNCCL.cpp:1899] [PG ID 0 PG GUID 0(default_pg) Rank 6] Process group watchdog thread terminated with exception: [Rank 6] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=2, OpType=ALLREDUCE, NumelIn=1, NumelOut=1, Timeout(ms)=600000) ran for 600082 milliseconds before timing out.
844
+ Exception raised from checkTimeout at /pytorch/torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:635 (most recent call first):
845
+ frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> >) + 0x98 (0x7b9ed57785e8 in /usr/local/lib/python3.11/dist-packages/torch/lib/libc10.so)
846
+ frame #1: c10d::ProcessGroupNCCL::WorkNCCL::checkTimeout(std::optional<std::chrono::duration<long, std::ratio<1l, 1000l> > >) + 0x23d (0x7b9e805cba6d in /usr/local/lib/python3.11/dist-packages/torch/lib/libtorch_cuda.so)
847
+ frame #2: c10d::ProcessGroupNCCL::watchdogHandler() + 0xc80 (0x7b9e805cd7f0 in /usr/local/lib/python3.11/dist-packages/torch/lib/libtorch_cuda.so)
848
+ frame #3: c10d::ProcessGroupNCCL::ncclCommWatchdog() + 0x14d (0x7b9e805ceefd in /usr/local/lib/python3.11/dist-packages/torch/lib/libtorch_cuda.so)
849
+ frame #4: <unknown function> + 0xdc253 (0x7b9e703d8253 in /lib/x86_64-linux-gnu/libstdc++.so.6)
850
+ frame #5: <unknown function> + 0x94ac3 (0x7b9ed690aac3 in /lib/x86_64-linux-gnu/libc.so.6)
851
+ frame #6: clone + 0x44 (0x7b9ed699ba04 in /lib/x86_64-linux-gnu/libc.so.6)
852
+
853
+ terminate called after throwing an instance of 'c10::DistBackendError'
854
+ what(): [PG ID 0 PG GUID 0(default_pg) Rank 6] Process group watchdog thread terminated with exception: [Rank 6] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=2, OpType=ALLREDUCE, NumelIn=1, NumelOut=1, Timeout(ms)=600000) ran for 600082 milliseconds before timing out.
855
+ Exception raised from checkTimeout at /pytorch/torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:635 (most recent call first):
856
+ frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> >) + 0x98 (0x7b9ed57785e8 in /usr/local/lib/python3.11/dist-packages/torch/lib/libc10.so)
857
+ frame #1: c10d::ProcessGroupNCCL::WorkNCCL::checkTimeout(std::optional<std::chrono::duration<long, std::ratio<1l, 1000l> > >) + 0x23d (0x7b9e805cba6d in /usr/local/lib/python3.11/dist-packages/torch/lib/libtorch_cuda.so)
858
+ frame #2: c10d::ProcessGroupNCCL::watchdogHandler() + 0xc80 (0x7b9e805cd7f0 in /usr/local/lib/python3.11/dist-packages/torch/lib/libtorch_cuda.so)
859
+ frame #3: c10d::ProcessGroupNCCL::ncclCommWatchdog() + 0x14d (0x7b9e805ceefd in /usr/local/lib/python3.11/dist-packages/torch/lib/libtorch_cuda.so)
860
+ frame #4: <unknown function> + 0xdc253 (0x7b9e703d8253 in /lib/x86_64-linux-gnu/libstdc++.so.6)
861
+ frame #5: <unknown function> + 0x94ac3 (0x7b9ed690aac3 in /lib/x86_64-linux-gnu/libc.so.6)
862
+ frame #6: clone + 0x44 (0x7b9ed699ba04 in /lib/x86_64-linux-gnu/libc.so.6)
863
+
864
+ Exception raised from ncclCommWatchdog at /pytorch/torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1905 (most recent call first):
865
+ frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> >) + 0x98 (0x7b9ed57785e8 in /usr/local/lib/python3.11/dist-packages/torch/lib/libc10.so)
866
+ frame #1: <unknown function> + 0x11b4abe (0x7b9e8059dabe in /usr/local/lib/python3.11/dist-packages/torch/lib/libtorch_cuda.so)
867
+ frame #2: <unknown function> + 0xe07bed (0x7b9e801f0bed in /usr/local/lib/python3.11/dist-packages/torch/lib/libtorch_cuda.so)
868
+ frame #3: <unknown function> + 0xdc253 (0x7b9e703d8253 in /lib/x86_64-linux-gnu/libstdc++.so.6)
869
+ frame #4: <unknown function> + 0x94ac3 (0x7b9ed690aac3 in /lib/x86_64-linux-gnu/libc.so.6)
870
+ frame #5: clone + 0x44 (0x7b9ed699ba04 in /lib/x86_64-linux-gnu/libc.so.6)
871
+
872
+ W0312 23:52:50.209000 2052 torch/distributed/elastic/multiprocessing/api.py:900] Sending process 2120 closing signal SIGTERM
873
+ W0312 23:52:50.211000 2052 torch/distributed/elastic/multiprocessing/api.py:900] Sending process 2121 closing signal SIGTERM
874
+ W0312 23:52:50.213000 2052 torch/distributed/elastic/multiprocessing/api.py:900] Sending process 2123 closing signal SIGTERM
875
+ W0312 23:52:50.214000 2052 torch/distributed/elastic/multiprocessing/api.py:900] Sending process 2124 closing signal SIGTERM
876
+ W0312 23:52:50.215000 2052 torch/distributed/elastic/multiprocessing/api.py:900] Sending process 2125 closing signal SIGTERM
877
+ W0312 23:52:50.216000 2052 torch/distributed/elastic/multiprocessing/api.py:900] Sending process 2127 closing signal SIGTERM
878
+ E0312 23:52:52.648000 2052 torch/distributed/elastic/multiprocessing/api.py:874] failed (exitcode: -6) local_rank: 2 (pid: 2122) of binary: /usr/bin/python
879
+ Traceback (most recent call last):
880
+ File "/usr/local/bin/accelerate", line 8, in <module>
881
+ sys.exit(main())
882
+ ^^^^^^
883
+ File "/usr/local/lib/python3.11/dist-packages/accelerate/commands/accelerate_cli.py", line 50, in main
884
+ args.func(args)
885
+ File "/usr/local/lib/python3.11/dist-packages/accelerate/commands/launch.py", line 1396, in launch_command
886
+ multi_gpu_launcher(args)
887
+ File "/usr/local/lib/python3.11/dist-packages/accelerate/commands/launch.py", line 1023, in multi_gpu_launcher
888
+ distrib_run.run(args)
889
+ File "/usr/local/lib/python3.11/dist-packages/torch/distributed/run.py", line 883, in run
890
+ elastic_launch(
891
+ File "/usr/local/lib/python3.11/dist-packages/torch/distributed/launcher/api.py", line 139, in __call__
892
+ return launch_agent(self._config, self._entrypoint, list(args))
893
+ ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
894
+ File "/usr/local/lib/python3.11/dist-packages/torch/distributed/launcher/api.py", line 270, in launch_agent
895
+ raise ChildFailedError(
896
+ torch.distributed.elastic.multiprocessing.errors.ChildFailedError:
897
+ =====================================================
898
+ /tmp/train_on_hf.py FAILED
899
+ -----------------------------------------------------
900
+ Failures:
901
+ [1]:
902
+ time : 2026-03-12_23:52:50
903
+ host : faa36c97b6c7
904
+ rank : 6 (local_rank: 6)
905
+ exitcode : -6 (pid: 2126)
906
+ error_file: <N/A>
907
+ traceback : Signal 6 (SIGABRT) received by PID 2126
908
+ -----------------------------------------------------
909
+ Root Cause (first observed failure):
910
+ [0]:
911
+ time : 2026-03-12_23:52:50
912
+ host : faa36c97b6c7
913
+ rank : 2 (local_rank: 2)
914
+ exitcode : -6 (pid: 2122)
915
+ error_file: <N/A>
916
+ traceback : Signal 6 (SIGABRT) received by PID 2122
917
+ =====================================================
918
+
919
+ >>> Training finished with exit code 1
920
+ Ended: 2026-03-12 23:52:53 UTC
921
+ >>> Saving training log to HuggingFace...