🎯 Starting continuous training for 12 hours...
🤖 Autonomous evolution mode activated
🚀 Setting up training environment...
📊 GPU: NVIDIA H200 NVL
💾 GPU Memory: 139.8 GB
🤖 Autonomous evolution mode: ENABLED
📦 Loading model and tokenizer...
Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.

Loading checkpoint shards:   0%|          | 0/4 [00:00<?, ?it/s]
Loading checkpoint shards:  25%|██▌       | 1/4 [00:00<00:02,  1.21it/s]
Loading checkpoint shards:  50%|█████     | 2/4 [00:01<00:01,  1.23it/s]
Loading checkpoint shards:  75%|███████▌  | 3/4 [00:02<00:00,  1.26it/s]
Loading checkpoint shards: 100%|██████████| 4/4 [00:03<00:00,  1.29it/s]
Loading checkpoint shards: 100%|██████████| 4/4 [00:03<00:00,  1.27it/s]
✅ Model loaded: qwen2
✅ Tokenizer vocab size: 151665
📥 Loading Elizabeth corpus data...
✅ Loaded 3000 high-quality security-focused conversations
✅ Formatted 3000 training texts

Map:   0%|          | 0/3000 [00:00<?, ? examples/s]
Map:  67%|██████▋   | 2000/3000 [00:00<00:00, 16267.08 examples/s]
Map: 100%|██████████| 3000/3000 [00:00<00:00, 13459.72 examples/s]
✅ Tokenized dataset: 3000 examples
⚙️ Setting up training...
🔥 Starting training...
📈 Batch size: 4
📈 Gradient accumulation: 16
📈 Effective batch size: 64
⏰ Continuous training mode: 12 hours autonomous evolution

  0%|          | 0/92 [00:00<?, ?it/s]/home/x/.local/lib/python3.12/site-packages/torch/utils/checkpoint.py:460: UserWarning: torch.utils.checkpoint: please pass in use_reentrant=True or use_reentrant=False explicitly. The default value of use_reentrant will be updated to be False in the future. To maintain current behavior, pass use_reentrant=True. It is recommended that you use use_reentrant=False. Refer to docs for more details on the differences between the two variants.
  warnings.warn(

  1%|          | 1/92 [00:02<03:35,  2.37s/it]
  2%|▏         | 2/92 [00:04<03:16,  2.19s/it]
  3%|▎         | 3/92 [00:07<04:03,  2.73s/it]
  4%|▍         | 4/92 [00:10<03:55,  2.67s/it]
  5%|▌         | 5/92 [00:12<03:47,  2.62s/it]
  7%|▋         | 6/92 [00:14<03:22,  2.35s/it]
  8%|▊         | 7/92 [00:17<03:27,  2.44s/it]
  9%|▊         | 8/92 [00:19<03:25,  2.45s/it]
 10%|▉         | 9/92 [00:22<03:39,  2.65s/it]
 11%|█         | 10/92 [00:25<03:25,  2.51s/it]
                                               
{'loss': 2.6982, 'grad_norm': 34.75, 'learning_rate': 2e-05, 'epoch': 0.21}

 11%|█         | 10/92 [00:25<03:25,  2.51s/it]
 12%|█▏        | 11/92 [00:27<03:21,  2.49s/it]
 13%|█▎        | 12/92 [00:29<03:16,  2.46s/it]
 14%|█▍        | 13/92 [00:32<03:14,  2.46s/it]
 15%|█▌        | 14/92 [00:34<03:05,  2.37s/it]
 16%|█▋        | 15/92 [00:37<03:05,  2.41s/it]
 17%|█▋        | 16/92 [00:39<03:03,  2.41s/it]
 18%|█▊        | 17/92 [00:41<02:56,  2.35s/it]
 20%|█▉        | 18/92 [00:44<03:12,  2.60s/it]
 21%|██        | 19/92 [00:47<03:06,  2.55s/it]
 22%|██▏       | 20/92 [00:49<02:48,  2.33s/it]
                                               
{'loss': 0.5565, 'grad_norm': 9.25, 'learning_rate': 1.927502451102095e-05, 'epoch': 0.43}

 22%|██▏       | 20/92 [00:49<02:48,  2.33s/it]
 23%|██▎       | 21/92 [00:50<02:34,  2.18s/it]
 24%|██▍       | 22/92 [00:53<02:44,  2.35s/it]
 25%|██▌       | 23/92 [00:55<02:40,  2.32s/it]
 26%|██▌       | 24/92 [00:58<02:47,  2.47s/it]
 27%|██▋       | 25/92 [01:01<02:40,  2.39s/it]
 28%|██▊       | 26/92 [01:03<02:33,  2.32s/it]
 29%|██▉       | 27/92 [01:05<02:29,  2.29s/it]
 30%|███       | 28/92 [01:07<02:24,  2.26s/it]
 32%|███▏      | 29/92 [01:10<02:31,  2.41s/it]
 33%|███▎      | 30/92 [01:12<02:22,  2.30s/it]
                                               
{'loss': 0.1705, 'grad_norm': 4.59375, 'learning_rate': 1.720521593600787e-05, 'epoch': 0.64}

 33%|███▎      | 30/92 [01:12<02:22,  2.30s/it]
 34%|███▎      | 31/92 [01:15<02:34,  2.54s/it]
 35%|███▍      | 32/92 [01:17<02:31,  2.52s/it]
 36%|███▌      | 33/92 [01:20<02:23,  2.43s/it]
 37%|███▋      | 34/92 [01:22<02:15,  2.33s/it]
 38%|███▊      | 35/92 [01:24<02:15,  2.38s/it]
 39%|███▉      | 36/92 [01:28<02:30,  2.68s/it]
 40%|████      | 37/92 [01:31<02:38,  2.89s/it]
 41%|████▏     | 38/92 [01:33<02:22,  2.63s/it]
 42%|████▏     | 39/92 [01:36<02:18,  2.61s/it]
 43%|████▎     | 40/92 [01:38<02:14,  2.59s/it]
                                               
{'loss': 0.0486, 'grad_norm': 1.703125, 'learning_rate': 1.4090686371713403e-05, 'epoch': 0.85}

 43%|████▎     | 40/92 [01:38<02:14,  2.59s/it]
 45%|████▍     | 41/92 [01:41<02:13,  2.63s/it]
 46%|████▌     | 42/92 [01:43<02:09,  2.60s/it]
 47%|████▋     | 43/92 [01:47<02:20,  2.88s/it]
 48%|████▊     | 44/92 [01:49<02:12,  2.76s/it]
 49%|████▉     | 45/92 [01:52<02:09,  2.75s/it]
 50%|█████     | 46/92 [01:55<02:13,  2.90s/it]
 51%|█████     | 47/92 [01:58<02:04,  2.77s/it]
 52%|█████▏    | 48/92 [02:01<02:01,  2.76s/it]
 53%|█████▎    | 49/92 [02:03<01:52,  2.61s/it]
 54%|█████▍    | 50/92 [02:05<01:42,  2.44s/it]
                                               
{'loss': 0.0266, 'grad_norm': 0.8984375, 'learning_rate': 1.0383027336900356e-05, 'epoch': 1.07}

 54%|█████▍    | 50/92 [02:05<01:42,  2.44s/it]
 55%|█████▌    | 51/92 [02:07<01:36,  2.36s/it]
 57%|█████▋    | 52/92 [02:10<01:39,  2.48s/it]
 58%|█████▊    | 53/92 [02:13<01:39,  2.56s/it]
 59%|█████▊    | 54/92 [02:16<01:42,  2.70s/it]
 60%|█████▉    | 55/92 [02:18<01:33,  2.52s/it]
 61%|██████    | 56/92 [02:20<01:26,  2.40s/it]
 62%|██████▏   | 57/92 [02:23<01:31,  2.62s/it]
 63%|██████▎   | 58/92 [02:25<01:24,  2.48s/it]
 64%|██████▍   | 59/92 [02:28<01:21,  2.46s/it]
 65%|██████▌   | 60/92 [02:30<01:22,  2.58s/it]
                                               
{'loss': 0.0228, 'grad_norm': 0.54296875, 'learning_rate': 6.619831215914974e-06, 'epoch': 1.28}

 65%|██████▌   | 60/92 [02:30<01:22,  2.58s/it]
 66%|██████▋   | 61/92 [02:33<01:18,  2.54s/it]
 67%|██████▋   | 62/92 [02:36<01:18,  2.60s/it]
 68%|██████▊   | 63/92 [02:38<01:14,  2.55s/it]
 70%|██████▉   | 64/92 [02:40<01:08,  2.43s/it]
 71%|███████   | 65/92 [02:42<01:03,  2.37s/it]
 72%|███████▏  | 66/92 [02:44<00:57,  2.21s/it]
 73%|███████▎  | 67/92 [02:47<00:57,  2.30s/it]
 74%|███████▍  | 68/92 [02:49<00:58,  2.43s/it]
 75%|███████▌  | 69/92 [02:52<00:56,  2.45s/it]
 76%|███████▌  | 70/92 [02:55<00:56,  2.56s/it]
                                               
{'loss': 0.022, 'grad_norm': 0.474609375, 'learning_rate': 3.3467429983443477e-06, 'epoch': 1.49}

 76%|███████▌  | 70/92 [02:55<00:56,  2.56s/it]
 77%|███████▋  | 71/92 [02:57<00:54,  2.60s/it]
 78%|███████▊  | 72/92 [03:00<00:48,  2.44s/it]
 79%|███████▉  | 73/92 [03:02<00:47,  2.52s/it]
 80%|████████  | 74/92 [03:05<00:45,  2.53s/it]
 82%|████████▏ | 75/92 [03:07<00:41,  2.42s/it]
 83%|████████▎ | 76/92 [03:09<00:38,  2.38s/it]
 84%|████████▎ | 77/92 [03:12<00:37,  2.53s/it]
 85%|████████▍ | 78/92 [03:15<00:36,  2.60s/it]
 86%|████████▌ | 79/92 [03:18<00:35,  2.77s/it]
 87%|████████▋ | 80/92 [03:20<00:29,  2.46s/it]
                                               
{'loss': 0.0211, 'grad_norm': 0.55859375, 'learning_rate': 1.0383444303894453e-06, 'epoch': 1.71}

 87%|████████▋ | 80/92 [03:20<00:29,  2.46s/it]
 88%|████████▊ | 81/92 [03:22<00:25,  2.35s/it]
 89%|████████▉ | 82/92 [03:25<00:25,  2.57s/it]
 90%|█████████ | 83/92 [03:28<00:23,  2.63s/it]
 91%|█████████▏| 84/92 [03:31<00:21,  2.72s/it]
 92%|█████████▏| 85/92 [03:33<00:17,  2.46s/it]
 93%|█████████▎| 86/92 [03:35<00:14,  2.40s/it]
 95%|█████████▍| 87/92 [03:38<00:13,  2.64s/it]
 96%|█████████▌| 88/92 [03:40<00:09,  2.40s/it]
 97%|█████████▋| 89/92 [03:42<00:07,  2.39s/it]
 98%|█████████▊| 90/92 [03:44<00:04,  2.33s/it]
                                               
{'loss': 0.0215, 'grad_norm': 0.625, 'learning_rate': 2.9341988162595593e-08, 'epoch': 1.92}

 98%|█████████▊| 90/92 [03:44<00:04,  2.33s/it]
 99%|█████████▉| 91/92 [03:47<00:02,  2.41s/it]
100%|██████████| 92/92 [03:50<00:00,  2.54s/it]
                                               
{'train_runtime': 230.3452, 'train_samples_per_second': 26.048, 'train_steps_per_second': 0.399, 'train_loss': 0.3904266621836502, 'epoch': 1.96}

100%|██████████| 92/92 [03:50<00:00,  2.54s/it]
100%|██████████| 92/92 [03:50<00:00,  2.50s/it]
✅ Training completed in 0.07 hours
🎉 Training pipeline completed successfully!