| 2026-06-25 10:31:54,806 [INFO] Seed: 42 |
| 2026-06-25 10:31:54,807 [INFO] Using CUDA: NVIDIA L40 |
| 2026-06-25 10:31:54,809 [INFO] Enabled TF32 matmul/cudnn for CUDA |
| 2026-06-25 10:31:54,978 [INFO] HTTP Request: HEAD https://huggingface.co/Qwen/Qwen3-Embedding-0.6B/resolve/main/config.json "HTTP/1.1 307 Temporary Redirect" |
| 2026-06-25 10:31:54,985 [INFO] HTTP Request: HEAD https://huggingface.co/api/resolve-cache/models/Qwen/Qwen3-Embedding-0.6B/97b0c614be4d77ee51c0cef4e5f07c00f9eb65b3/config.json "HTTP/1.1 200 OK" |
| 2026-06-25 10:31:55,122 [INFO] HTTP Request: HEAD https://huggingface.co/Qwen/Qwen3-Embedding-0.6B/resolve/main/tokenizer_config.json "HTTP/1.1 307 Temporary Redirect" |
| 2026-06-25 10:31:55,127 [INFO] HTTP Request: HEAD https://huggingface.co/api/resolve-cache/models/Qwen/Qwen3-Embedding-0.6B/97b0c614be4d77ee51c0cef4e5f07c00f9eb65b3/tokenizer_config.json "HTTP/1.1 200 OK" |
| 2026-06-25 10:31:55,267 [INFO] HTTP Request: GET https://huggingface.co/api/models/Qwen/Qwen3-Embedding-0.6B/tree/main/additional_chat_templates?recursive=false&expand=false "HTTP/1.1 404 Not Found" |
| 2026-06-25 10:31:55,408 [INFO] HTTP Request: GET https://huggingface.co/api/models/Qwen/Qwen3-Embedding-0.6B/tree/main?recursive=true&expand=false "HTTP/1.1 200 OK" |
| 2026-06-25 10:31:56,294 [INFO] HTTP Request: GET https://huggingface.co/api/models/Qwen/Qwen3-Embedding-0.6B "HTTP/1.1 200 OK" |
| 2026-06-25 10:31:56,466 [INFO] HTTP Request: HEAD https://huggingface.co/Qwen/Qwen3-Embedding-0.6B/resolve/main/config.json "HTTP/1.1 307 Temporary Redirect" |
| 2026-06-25 10:31:56,472 [INFO] HTTP Request: HEAD https://huggingface.co/api/resolve-cache/models/Qwen/Qwen3-Embedding-0.6B/97b0c614be4d77ee51c0cef4e5f07c00f9eb65b3/config.json "HTTP/1.1 200 OK" |
| 2026-06-25 10:31:56,718 [INFO] HTTP Request: HEAD https://huggingface.co/Qwen/Qwen3-Embedding-0.6B/resolve/main/config.json "HTTP/1.1 307 Temporary Redirect" |
| 2026-06-25 10:31:56,723 [INFO] HTTP Request: HEAD https://huggingface.co/api/resolve-cache/models/Qwen/Qwen3-Embedding-0.6B/97b0c614be4d77ee51c0cef4e5f07c00f9eb65b3/config.json "HTTP/1.1 200 OK" |
| 2026-06-25 10:32:03,703 [INFO] LLM2Seq Model Summary |
| ================================================== |
| Encoder: Qwen/Qwen3-Embedding-0.6B |
| Trainable: True |
| LoRA: True |
| Adaptor: gated_residual |
| Layer Fusion: True |
| Fusion Type: tokenwise |
| Salience Gate: True |
| EncStack: True |
| Global Tokens: 16 |
| Decoder: |
| Layers: 8 |
| Hidden Size: 1024 |
| Heads: 16 |
| FFN Size: 4096 |
| ================================================== |
| Total params: 1,000,232,449 |
| Trainable params: 404,455,937 |
| Frozen params: 595,776,512 |
| 2026-06-25 10:32:04,225 [INFO] Enabled gradient checkpointing for encoder |
| 2026-06-25 10:32:04,583 [INFO] Resuming from checkpoint: runs/phase2_tiep_downloaded/checkpoints/phase2_lora_encoder_qwen_wikilingua_tiep/best.pt |
| 2026-06-25 10:32:05,899 [INFO] Checkpoint load allowed 310 missing keys for stage=phase2_lora_encoder_qwen_wikilingua_tiep context=resume, prefixes=('encoder.',) |
| 2026-06-25 10:32:05,900 [INFO] Reset global step because training.reset_global_step_on_resume=true |
| 2026-06-25 10:32:05,900 [INFO] Reset best eval loss because training.reset_best_eval_loss_on_resume=true |
| 2026-06-25 10:32:05,900 [INFO] Resumed at step 0 |
| 2026-06-25 10:32:06,053 [INFO] HTTP Request: POST https://huggingface.co/api/repos/create "HTTP/1.1 409 Conflict" |
| 2026-06-25 10:32:06,054 [INFO] Hugging Face uploads enabled: repo=BienKieu/llm2seq-wikilingua path=checkpoints/phase2_lora_encoder_qwen_wikilingua |
| 2026-06-25 10:32:06,054 [INFO] Train file: llm2seq/data/processed/train.jsonl (13,999 examples) |
| 2026-06-25 10:32:06,054 [INFO] Eval file: llm2seq/data/processed/validation.jsonl (1,680 examples) |
| 2026-06-25 10:32:06,054 [INFO] Starting training for 3 epochs (1,314 optimizer steps)... |
| 2026-06-25 10:32:06,054 [INFO] Batch size: 8 |
| 2026-06-25 10:32:06,054 [INFO] Gradient accumulation: 4 |
| 2026-06-25 10:32:06,054 [INFO] Effective batch size: 32 |
| 2026-06-25 10:32:06,054 [INFO] Steps per epoch: 438 |
| 2026-06-25 10:32:06,054 [INFO] Warmup steps: 65 |
| 2026-06-25 10:32:06,054 [INFO] Eval every: 438 optimizer steps |
| 2026-06-25 10:32:06,054 [INFO] Save every: 438 optimizer steps |
| 2026-06-25 10:32:06,054 [INFO] Mixed precision: fp16=False, bf16=True |
| 2026-06-25 10:32:06,054 [INFO] Initial optimizer LRs: adaptor=0.00e+00, decoder=0.00e+00, encoder=0.00e+00 |
| 2026-06-25 10:32:41,904 [INFO] Step 5/1314 | Loss: 1.2223 | CE: 1.2223 | LR: 3.85e-06 | Epoch: 1/3 |
| 2026-06-25 10:33:16,946 [INFO] Step 10/1314 | Loss: 1.2231 | CE: 1.2231 | LR: 7.69e-06 | Epoch: 1/3 |
| 2026-06-25 10:33:44,101 [INFO] Step 15/1314 | Loss: 1.3032 | CE: 1.3032 | LR: 1.15e-05 | Epoch: 1/3 |
| 2026-06-25 10:34:22,677 [INFO] Step 20/1314 | Loss: 1.3171 | CE: 1.3171 | LR: 1.54e-05 | Epoch: 1/3 |
| 2026-06-25 10:34:56,571 [INFO] Step 25/1314 | Loss: 1.2319 | CE: 1.2319 | LR: 1.92e-05 | Epoch: 1/3 |
| 2026-06-25 10:35:22,393 [INFO] Step 30/1314 | Loss: 1.2137 | CE: 1.2137 | LR: 2.31e-05 | Epoch: 1/3 |
| 2026-06-25 10:35:51,027 [INFO] Step 35/1314 | Loss: 1.1899 | CE: 1.1899 | LR: 2.69e-05 | Epoch: 1/3 |
| 2026-06-25 10:36:22,248 [INFO] Step 40/1314 | Loss: 1.1941 | CE: 1.1941 | LR: 3.08e-05 | Epoch: 1/3 |
| 2026-06-25 10:36:59,226 [INFO] Step 45/1314 | Loss: 1.1875 | CE: 1.1875 | LR: 3.46e-05 | Epoch: 1/3 |
| 2026-06-25 10:37:29,177 [INFO] Step 50/1314 | Loss: 1.1912 | CE: 1.1912 | LR: 3.85e-05 | Epoch: 1/3 |
| 2026-06-25 10:37:58,487 [INFO] Step 55/1314 | Loss: 1.1377 | CE: 1.1377 | LR: 4.23e-05 | Epoch: 1/3 |
| 2026-06-25 10:38:31,660 [INFO] Step 60/1314 | Loss: 1.1947 | CE: 1.1947 | LR: 4.62e-05 | Epoch: 1/3 |
| 2026-06-25 10:39:04,631 [INFO] Step 65/1314 | Loss: 1.0759 | CE: 1.0759 | LR: 5.00e-05 | Epoch: 1/3 |
| 2026-06-25 10:39:36,933 [INFO] Step 70/1314 | Loss: 1.1080 | CE: 1.1080 | LR: 5.00e-05 | Epoch: 1/3 |
| 2026-06-25 10:40:09,737 [INFO] Step 75/1314 | Loss: 1.0679 | CE: 1.0679 | LR: 5.00e-05 | Epoch: 1/3 |
| 2026-06-25 10:40:48,307 [INFO] Step 80/1314 | Loss: 1.1077 | CE: 1.1077 | LR: 5.00e-05 | Epoch: 1/3 |
| 2026-06-25 10:41:16,842 [INFO] Step 85/1314 | Loss: 1.0349 | CE: 1.0349 | LR: 5.00e-05 | Epoch: 1/3 |
| 2026-06-25 10:41:45,076 [INFO] Step 90/1314 | Loss: 1.0159 | CE: 1.0159 | LR: 5.00e-05 | Epoch: 1/3 |
| 2026-06-25 10:42:14,643 [INFO] Step 95/1314 | Loss: 1.0613 | CE: 1.0613 | LR: 4.99e-05 | Epoch: 1/3 |
| 2026-06-25 10:42:47,866 [INFO] Step 100/1314 | Loss: 1.0359 | CE: 1.0359 | LR: 4.99e-05 | Epoch: 1/3 |
| 2026-06-25 10:43:15,478 [INFO] Step 105/1314 | Loss: 1.1110 | CE: 1.1110 | LR: 4.99e-05 | Epoch: 1/3 |
| 2026-06-25 10:43:48,100 [INFO] Step 110/1314 | Loss: 1.0767 | CE: 1.0767 | LR: 4.98e-05 | Epoch: 1/3 |
| 2026-06-25 10:44:29,951 [INFO] Step 115/1314 | Loss: 1.0397 | CE: 1.0397 | LR: 4.98e-05 | Epoch: 1/3 |
| 2026-06-25 10:45:01,874 [INFO] Step 120/1314 | Loss: 1.0350 | CE: 1.0350 | LR: 4.98e-05 | Epoch: 1/3 |
| 2026-06-25 10:45:40,430 [INFO] Step 125/1314 | Loss: 1.0430 | CE: 1.0430 | LR: 4.97e-05 | Epoch: 1/3 |
| 2026-06-25 10:46:15,711 [INFO] Step 130/1314 | Loss: 1.0818 | CE: 1.0818 | LR: 4.97e-05 | Epoch: 1/3 |
| 2026-06-25 10:46:50,142 [INFO] Step 135/1314 | Loss: 1.0795 | CE: 1.0795 | LR: 4.96e-05 | Epoch: 1/3 |
| 2026-06-25 10:47:27,193 [INFO] Step 140/1314 | Loss: 1.0102 | CE: 1.0102 | LR: 4.96e-05 | Epoch: 1/3 |
| 2026-06-25 10:48:00,345 [INFO] Step 145/1314 | Loss: 1.0572 | CE: 1.0572 | LR: 4.95e-05 | Epoch: 1/3 |
| 2026-06-25 10:48:29,126 [INFO] Step 150/1314 | Loss: 0.9785 | CE: 0.9785 | LR: 4.94e-05 | Epoch: 1/3 |
| 2026-06-25 10:48:57,146 [INFO] Step 155/1314 | Loss: 1.1187 | CE: 1.1187 | LR: 4.94e-05 | Epoch: 1/3 |
| 2026-06-25 10:49:31,067 [INFO] Step 160/1314 | Loss: 1.0280 | CE: 1.0280 | LR: 4.93e-05 | Epoch: 1/3 |
| 2026-06-25 10:50:06,022 [INFO] Step 165/1314 | Loss: 1.0586 | CE: 1.0586 | LR: 4.92e-05 | Epoch: 1/3 |
| 2026-06-25 10:50:43,846 [INFO] Step 170/1314 | Loss: 0.9697 | CE: 0.9697 | LR: 4.91e-05 | Epoch: 1/3 |
| 2026-06-25 10:51:20,430 [INFO] Step 175/1314 | Loss: 1.1142 | CE: 1.1142 | LR: 4.90e-05 | Epoch: 1/3 |
| 2026-06-25 10:51:51,949 [INFO] Step 180/1314 | Loss: 1.0135 | CE: 1.0135 | LR: 4.90e-05 | Epoch: 1/3 |
| 2026-06-25 10:52:25,900 [INFO] Step 185/1314 | Loss: 1.0268 | CE: 1.0268 | LR: 4.89e-05 | Epoch: 1/3 |
| 2026-06-25 10:53:00,497 [INFO] Step 190/1314 | Loss: 1.0583 | CE: 1.0583 | LR: 4.88e-05 | Epoch: 1/3 |
| 2026-06-25 10:53:32,810 [INFO] Step 195/1314 | Loss: 1.0427 | CE: 1.0427 | LR: 4.87e-05 | Epoch: 1/3 |
| 2026-06-25 10:54:04,050 [INFO] Step 200/1314 | Loss: 1.0206 | CE: 1.0206 | LR: 4.86e-05 | Epoch: 1/3 |
| 2026-06-25 10:54:37,920 [INFO] Step 205/1314 | Loss: 1.0674 | CE: 1.0674 | LR: 4.85e-05 | Epoch: 1/3 |
| 2026-06-25 10:55:07,659 [INFO] Step 210/1314 | Loss: 1.0228 | CE: 1.0228 | LR: 4.84e-05 | Epoch: 1/3 |
| 2026-06-25 10:55:41,771 [INFO] Step 215/1314 | Loss: 1.0172 | CE: 1.0172 | LR: 4.82e-05 | Epoch: 1/3 |
| 2026-06-25 10:56:07,025 [INFO] Step 220/1314 | Loss: 0.9537 | CE: 0.9537 | LR: 4.81e-05 | Epoch: 1/3 |
| 2026-06-25 10:56:38,713 [INFO] Step 225/1314 | Loss: 1.0152 | CE: 1.0152 | LR: 4.80e-05 | Epoch: 1/3 |
| 2026-06-25 10:57:09,733 [INFO] Step 230/1314 | Loss: 1.0707 | CE: 1.0707 | LR: 4.79e-05 | Epoch: 1/3 |
| 2026-06-25 10:57:38,965 [INFO] Step 235/1314 | Loss: 0.9801 | CE: 0.9801 | LR: 4.77e-05 | Epoch: 1/3 |
| 2026-06-25 10:58:08,484 [INFO] Step 240/1314 | Loss: 1.0786 | CE: 1.0786 | LR: 4.76e-05 | Epoch: 1/3 |
| 2026-06-25 10:58:46,080 [INFO] Step 245/1314 | Loss: 1.1201 | CE: 1.1201 | LR: 4.75e-05 | Epoch: 1/3 |
| 2026-06-25 10:59:18,588 [INFO] Step 250/1314 | Loss: 1.0053 | CE: 1.0053 | LR: 4.73e-05 | Epoch: 1/3 |
| 2026-06-25 10:59:45,682 [INFO] Step 255/1314 | Loss: 0.9513 | CE: 0.9513 | LR: 4.72e-05 | Epoch: 1/3 |
| 2026-06-25 11:00:22,191 [INFO] Step 260/1314 | Loss: 1.0657 | CE: 1.0657 | LR: 4.71e-05 | Epoch: 1/3 |
| 2026-06-25 11:00:52,383 [INFO] Step 265/1314 | Loss: 1.1124 | CE: 1.1124 | LR: 4.69e-05 | Epoch: 1/3 |
| 2026-06-25 11:01:22,514 [INFO] Step 270/1314 | Loss: 1.0609 | CE: 1.0609 | LR: 4.67e-05 | Epoch: 1/3 |
| 2026-06-25 11:01:58,894 [INFO] Step 275/1314 | Loss: 1.0561 | CE: 1.0561 | LR: 4.66e-05 | Epoch: 1/3 |
| 2026-06-25 11:02:26,944 [INFO] Step 280/1314 | Loss: 1.0024 | CE: 1.0024 | LR: 4.64e-05 | Epoch: 1/3 |
| 2026-06-25 11:03:05,372 [INFO] Step 285/1314 | Loss: 1.1537 | CE: 1.1537 | LR: 4.63e-05 | Epoch: 1/3 |
| 2026-06-25 11:03:38,677 [INFO] Step 290/1314 | Loss: 1.1088 | CE: 1.1088 | LR: 4.61e-05 | Epoch: 1/3 |
| 2026-06-25 11:04:14,901 [INFO] Step 295/1314 | Loss: 1.1205 | CE: 1.1205 | LR: 4.59e-05 | Epoch: 1/3 |
| 2026-06-25 11:04:44,498 [INFO] Step 300/1314 | Loss: 1.0576 | CE: 1.0576 | LR: 4.58e-05 | Epoch: 1/3 |
| 2026-06-25 11:05:20,185 [INFO] Step 305/1314 | Loss: 1.1157 | CE: 1.1157 | LR: 4.56e-05 | Epoch: 1/3 |
| 2026-06-25 11:05:56,243 [INFO] Step 310/1314 | Loss: 1.0391 | CE: 1.0391 | LR: 4.54e-05 | Epoch: 1/3 |
| 2026-06-25 11:06:28,420 [INFO] Step 315/1314 | Loss: 1.0499 | CE: 1.0499 | LR: 4.52e-05 | Epoch: 1/3 |
| 2026-06-25 11:07:08,849 [INFO] Step 320/1314 | Loss: 1.0931 | CE: 1.0931 | LR: 4.50e-05 | Epoch: 1/3 |
| 2026-06-25 11:07:40,086 [INFO] Step 325/1314 | Loss: 1.0764 | CE: 1.0764 | LR: 4.48e-05 | Epoch: 1/3 |
| 2026-06-25 11:08:10,644 [INFO] Step 330/1314 | Loss: 1.0893 | CE: 1.0893 | LR: 4.46e-05 | Epoch: 1/3 |
| 2026-06-25 11:08:43,802 [INFO] Step 335/1314 | Loss: 1.0452 | CE: 1.0452 | LR: 4.45e-05 | Epoch: 1/3 |
| 2026-06-25 11:09:12,556 [INFO] Step 340/1314 | Loss: 1.0591 | CE: 1.0591 | LR: 4.43e-05 | Epoch: 1/3 |
| 2026-06-25 11:09:42,671 [INFO] Step 345/1314 | Loss: 1.0268 | CE: 1.0268 | LR: 4.41e-05 | Epoch: 1/3 |
| 2026-06-25 11:10:12,869 [INFO] Step 350/1314 | Loss: 1.0086 | CE: 1.0086 | LR: 4.38e-05 | Epoch: 1/3 |
| 2026-06-25 11:10:44,284 [INFO] Step 355/1314 | Loss: 1.1038 | CE: 1.1038 | LR: 4.36e-05 | Epoch: 1/3 |
| 2026-06-25 11:11:26,431 [INFO] Step 360/1314 | Loss: 1.0343 | CE: 1.0343 | LR: 4.34e-05 | Epoch: 1/3 |
| 2026-06-25 11:12:06,155 [INFO] Step 365/1314 | Loss: 1.1276 | CE: 1.1276 | LR: 4.32e-05 | Epoch: 1/3 |
| 2026-06-25 11:12:36,958 [INFO] Step 370/1314 | Loss: 1.0585 | CE: 1.0585 | LR: 4.30e-05 | Epoch: 1/3 |
| 2026-06-25 11:13:09,953 [INFO] Step 375/1314 | Loss: 1.0722 | CE: 1.0722 | LR: 4.28e-05 | Epoch: 1/3 |
| 2026-06-25 11:13:44,379 [INFO] Step 380/1314 | Loss: 1.0973 | CE: 1.0973 | LR: 4.26e-05 | Epoch: 1/3 |
| 2026-06-25 11:14:16,246 [INFO] Step 385/1314 | Loss: 1.0209 | CE: 1.0209 | LR: 4.23e-05 | Epoch: 1/3 |
| 2026-06-25 11:14:44,668 [INFO] Step 390/1314 | Loss: 1.1053 | CE: 1.1053 | LR: 4.21e-05 | Epoch: 1/3 |
| 2026-06-25 11:15:20,714 [INFO] Step 395/1314 | Loss: 1.0722 | CE: 1.0722 | LR: 4.19e-05 | Epoch: 1/3 |
| 2026-06-25 11:15:57,706 [INFO] Step 400/1314 | Loss: 1.0380 | CE: 1.0380 | LR: 4.16e-05 | Epoch: 1/3 |
| 2026-06-25 11:16:30,295 [INFO] Step 405/1314 | Loss: 1.0338 | CE: 1.0338 | LR: 4.14e-05 | Epoch: 1/3 |
| 2026-06-25 11:17:00,328 [INFO] Step 410/1314 | Loss: 1.0287 | CE: 1.0287 | LR: 4.12e-05 | Epoch: 1/3 |
| 2026-06-25 11:17:39,405 [INFO] Step 415/1314 | Loss: 1.0614 | CE: 1.0614 | LR: 4.09e-05 | Epoch: 1/3 |
| 2026-06-25 11:18:14,834 [INFO] Step 420/1314 | Loss: 1.1105 | CE: 1.1105 | LR: 4.07e-05 | Epoch: 1/3 |
| 2026-06-25 11:18:46,171 [INFO] Step 425/1314 | Loss: 1.0736 | CE: 1.0736 | LR: 4.04e-05 | Epoch: 1/3 |
| 2026-06-25 11:19:19,486 [INFO] Step 430/1314 | Loss: 1.1955 | CE: 1.1955 | LR: 4.02e-05 | Epoch: 1/3 |
| 2026-06-25 11:19:58,086 [INFO] Step 435/1314 | Loss: 1.2491 | CE: 1.2491 | LR: 3.99e-05 | Epoch: 1/3 |
| 2026-06-25 11:21:28,962 [INFO] Step 438 | Eval Loss: 2.7103 |
| 2026-06-25 11:21:30,927 [INFO] Saved trainable-only checkpoint to runs/phase2_lora_encoder_qwen_wikilingua/best.pt (575 model tensors, stores_base_encoder_weights=False) |
| 2026-06-25 11:21:32,374 [INFO] New best eval loss: 2.7103 |
| 2026-06-25 11:21:34,327 [INFO] Saved trainable-only checkpoint to runs/phase2_lora_encoder_qwen_wikilingua/epoch_1.pt (575 model tensors, stores_base_encoder_weights=False) |
| 2026-06-25 11:21:35,799 [INFO] Saved epoch 1 checkpoint: runs/phase2_lora_encoder_qwen_wikilingua/epoch_1.pt |
| 2026-06-25 11:21:37,239 [INFO] Uploading to HF: runs/phase2_lora_encoder_qwen_wikilingua/epoch_1.pt -> BienKieu/llm2seq-wikilingua/checkpoints/phase2_lora_encoder_qwen_wikilingua/epochs/epoch_001_step_438.pt |
| 2026-06-25 11:21:37,435 [INFO] HTTP Request: POST https://huggingface.co/api/models/BienKieu/llm2seq-wikilingua/preupload/main "HTTP/1.1 200 OK" |
| 2026-06-25 11:21:56,167 [INFO] HTTP Request: POST https://huggingface.co/api/models/BienKieu/llm2seq-wikilingua/commit/main "HTTP/1.1 200 OK" |
| 2026-06-25 11:21:56,169 [INFO] Uploading to HF: runs/phase2_lora_encoder_qwen_wikilingua/config.yaml -> BienKieu/llm2seq-wikilingua/checkpoints/phase2_lora_encoder_qwen_wikilingua/config.yaml |
| 2026-06-25 11:21:56,329 [INFO] HTTP Request: POST https://huggingface.co/api/models/BienKieu/llm2seq-wikilingua/preupload/main "HTTP/1.1 200 OK" |
| 2026-06-25 11:21:56,729 [INFO] HTTP Request: POST https://huggingface.co/api/models/BienKieu/llm2seq-wikilingua/commit/main "HTTP/1.1 200 OK" |
| 2026-06-25 11:21:56,730 [INFO] Uploading to HF: runs/phase2_lora_encoder_qwen_wikilingua/train.log -> BienKieu/llm2seq-wikilingua/checkpoints/phase2_lora_encoder_qwen_wikilingua/train.log |
| 2026-06-25 11:21:56,892 [INFO] HTTP Request: POST https://huggingface.co/api/models/BienKieu/llm2seq-wikilingua/preupload/main "HTTP/1.1 200 OK" |
| 2026-06-25 11:21:57,342 [INFO] HTTP Request: POST https://huggingface.co/api/models/BienKieu/llm2seq-wikilingua/commit/main "HTTP/1.1 200 OK" |
| 2026-06-25 11:21:57,343 [INFO] Uploading to HF: runs/phase2_lora_encoder_qwen_wikilingua/checkpoint_manifest.json -> BienKieu/llm2seq-wikilingua/checkpoints/phase2_lora_encoder_qwen_wikilingua/checkpoint_manifest.json |
| 2026-06-25 11:21:57,504 [INFO] HTTP Request: POST https://huggingface.co/api/models/BienKieu/llm2seq-wikilingua/preupload/main "HTTP/1.1 200 OK" |
| 2026-06-25 11:21:58,259 [INFO] HTTP Request: POST https://huggingface.co/api/models/BienKieu/llm2seq-wikilingua/commit/main "HTTP/1.1 200 OK" |
| 2026-06-25 11:21:59,629 [INFO] Uploading to HF: runs/phase2_lora_encoder_qwen_wikilingua/best.pt -> BienKieu/llm2seq-wikilingua/checkpoints/phase2_lora_encoder_qwen_wikilingua/best.pt |
| 2026-06-25 11:21:59,790 [INFO] HTTP Request: POST https://huggingface.co/api/models/BienKieu/llm2seq-wikilingua/preupload/main "HTTP/1.1 200 OK" |
| 2026-06-25 11:22:10,863 [INFO] HTTP Request: POST https://huggingface.co/api/models/BienKieu/llm2seq-wikilingua/commit/main "HTTP/1.1 200 OK" |
| 2026-06-25 11:22:22,989 [INFO] Step 440/1314 | Loss: 1.1373 | CE: 1.1373 | LR: 3.97e-05 | Epoch: 2/3 |
| 2026-06-25 11:23:00,666 [INFO] Step 445/1314 | Loss: 1.0627 | CE: 1.0627 | LR: 3.94e-05 | Epoch: 2/3 |
| 2026-06-25 11:23:36,805 [INFO] Step 450/1314 | Loss: 1.0317 | CE: 1.0317 | LR: 3.92e-05 | Epoch: 2/3 |
| 2026-06-25 11:24:11,324 [INFO] Step 455/1314 | Loss: 1.1101 | CE: 1.1101 | LR: 3.89e-05 | Epoch: 2/3 |
| 2026-06-25 11:24:44,013 [INFO] Step 460/1314 | Loss: 1.0420 | CE: 1.0420 | LR: 3.86e-05 | Epoch: 2/3 |
| 2026-06-25 11:25:10,177 [INFO] Step 465/1314 | Loss: 1.0539 | CE: 1.0539 | LR: 3.84e-05 | Epoch: 2/3 |
| 2026-06-25 11:25:38,622 [INFO] Step 470/1314 | Loss: 1.0029 | CE: 1.0029 | LR: 3.81e-05 | Epoch: 2/3 |
| 2026-06-25 11:26:13,050 [INFO] Step 475/1314 | Loss: 1.0971 | CE: 1.0971 | LR: 3.78e-05 | Epoch: 2/3 |
| 2026-06-25 11:26:49,906 [INFO] Step 480/1314 | Loss: 1.1059 | CE: 1.1059 | LR: 3.76e-05 | Epoch: 2/3 |
| 2026-06-25 11:27:23,125 [INFO] Step 485/1314 | Loss: 1.1175 | CE: 1.1175 | LR: 3.73e-05 | Epoch: 2/3 |
| 2026-06-25 11:28:09,800 [INFO] Step 490/1314 | Loss: 1.0665 | CE: 1.0665 | LR: 3.70e-05 | Epoch: 2/3 |
| 2026-06-25 11:28:46,523 [INFO] Step 495/1314 | Loss: 1.1268 | CE: 1.1268 | LR: 3.67e-05 | Epoch: 2/3 |
| 2026-06-25 11:29:26,340 [INFO] Step 500/1314 | Loss: 1.0398 | CE: 1.0398 | LR: 3.65e-05 | Epoch: 2/3 |
| 2026-06-25 11:29:59,894 [INFO] Step 505/1314 | Loss: 1.0177 | CE: 1.0177 | LR: 3.62e-05 | Epoch: 2/3 |
| 2026-06-25 11:30:33,137 [INFO] Step 510/1314 | Loss: 1.0822 | CE: 1.0822 | LR: 3.59e-05 | Epoch: 2/3 |
| 2026-06-25 11:31:02,108 [INFO] Step 515/1314 | Loss: 1.0694 | CE: 1.0694 | LR: 3.56e-05 | Epoch: 2/3 |
| 2026-06-25 11:31:30,859 [INFO] Step 520/1314 | Loss: 1.0420 | CE: 1.0420 | LR: 3.53e-05 | Epoch: 2/3 |
| 2026-06-25 11:32:10,074 [INFO] Step 525/1314 | Loss: 1.0666 | CE: 1.0666 | LR: 3.51e-05 | Epoch: 2/3 |
| 2026-06-25 11:32:49,511 [INFO] Step 530/1314 | Loss: 0.9933 | CE: 0.9933 | LR: 3.48e-05 | Epoch: 2/3 |
| 2026-06-25 11:33:23,376 [INFO] Step 535/1314 | Loss: 1.0828 | CE: 1.0828 | LR: 3.45e-05 | Epoch: 2/3 |
| 2026-06-25 11:33:59,622 [INFO] Step 540/1314 | Loss: 1.0853 | CE: 1.0853 | LR: 3.42e-05 | Epoch: 2/3 |
| 2026-06-25 11:34:28,630 [INFO] Step 545/1314 | Loss: 1.0582 | CE: 1.0582 | LR: 3.39e-05 | Epoch: 2/3 |
| 2026-06-25 11:34:59,039 [INFO] Step 550/1314 | Loss: 1.0644 | CE: 1.0644 | LR: 3.36e-05 | Epoch: 2/3 |
| 2026-06-25 11:35:29,342 [INFO] Step 555/1314 | Loss: 1.0484 | CE: 1.0484 | LR: 3.33e-05 | Epoch: 2/3 |
| 2026-06-25 11:35:59,010 [INFO] Step 560/1314 | Loss: 1.0344 | CE: 1.0344 | LR: 3.30e-05 | Epoch: 2/3 |
| 2026-06-25 11:36:33,413 [INFO] Step 565/1314 | Loss: 1.0662 | CE: 1.0662 | LR: 3.27e-05 | Epoch: 2/3 |
| 2026-06-25 11:37:07,867 [INFO] Step 570/1314 | Loss: 1.0819 | CE: 1.0819 | LR: 3.24e-05 | Epoch: 2/3 |
| 2026-06-25 11:37:36,715 [INFO] Step 575/1314 | Loss: 1.0807 | CE: 1.0807 | LR: 3.21e-05 | Epoch: 2/3 |
| 2026-06-25 11:38:08,730 [INFO] Step 580/1314 | Loss: 1.0103 | CE: 1.0103 | LR: 3.18e-05 | Epoch: 2/3 |
| 2026-06-25 11:38:42,467 [INFO] Step 585/1314 | Loss: 1.0551 | CE: 1.0551 | LR: 3.15e-05 | Epoch: 2/3 |
| 2026-06-25 11:39:17,746 [INFO] Step 590/1314 | Loss: 1.0011 | CE: 1.0011 | LR: 3.12e-05 | Epoch: 2/3 |
| 2026-06-25 11:39:48,370 [INFO] Step 595/1314 | Loss: 1.1177 | CE: 1.1177 | LR: 3.09e-05 | Epoch: 2/3 |
| 2026-06-25 11:40:16,234 [INFO] Step 600/1314 | Loss: 1.0691 | CE: 1.0691 | LR: 3.06e-05 | Epoch: 2/3 |
| 2026-06-25 11:40:47,645 [INFO] Step 605/1314 | Loss: 1.0620 | CE: 1.0620 | LR: 3.03e-05 | Epoch: 2/3 |
| 2026-06-25 11:41:23,667 [INFO] Step 610/1314 | Loss: 1.1213 | CE: 1.1213 | LR: 3.00e-05 | Epoch: 2/3 |
| 2026-06-25 11:41:58,274 [INFO] Step 615/1314 | Loss: 1.0418 | CE: 1.0418 | LR: 2.97e-05 | Epoch: 2/3 |
| 2026-06-25 11:42:30,819 [INFO] Step 620/1314 | Loss: 0.9987 | CE: 0.9987 | LR: 2.93e-05 | Epoch: 2/3 |
| 2026-06-25 11:43:10,444 [INFO] Step 625/1314 | Loss: 1.0275 | CE: 1.0275 | LR: 2.90e-05 | Epoch: 2/3 |
| 2026-06-25 11:43:48,720 [INFO] Step 630/1314 | Loss: 1.0322 | CE: 1.0322 | LR: 2.87e-05 | Epoch: 2/3 |
| 2026-06-25 11:44:20,325 [INFO] Step 635/1314 | Loss: 1.0403 | CE: 1.0403 | LR: 2.84e-05 | Epoch: 2/3 |
| 2026-06-25 11:44:52,928 [INFO] Step 640/1314 | Loss: 1.0615 | CE: 1.0615 | LR: 2.81e-05 | Epoch: 2/3 |
| 2026-06-25 11:45:23,782 [INFO] Step 645/1314 | Loss: 1.0507 | CE: 1.0507 | LR: 2.78e-05 | Epoch: 2/3 |
| 2026-06-25 11:45:49,461 [INFO] Step 650/1314 | Loss: 1.0374 | CE: 1.0374 | LR: 2.75e-05 | Epoch: 2/3 |
| 2026-06-25 11:46:28,862 [INFO] Step 655/1314 | Loss: 1.1245 | CE: 1.1245 | LR: 2.72e-05 | Epoch: 2/3 |
| 2026-06-25 11:47:00,926 [INFO] Step 660/1314 | Loss: 1.1091 | CE: 1.1091 | LR: 2.69e-05 | Epoch: 2/3 |
| 2026-06-25 11:47:40,563 [INFO] Step 665/1314 | Loss: 1.1080 | CE: 1.1080 | LR: 2.65e-05 | Epoch: 2/3 |
| 2026-06-25 11:48:16,175 [INFO] Step 670/1314 | Loss: 1.0421 | CE: 1.0421 | LR: 2.62e-05 | Epoch: 2/3 |
| 2026-06-25 11:48:48,974 [INFO] Step 675/1314 | Loss: 1.0845 | CE: 1.0845 | LR: 2.59e-05 | Epoch: 2/3 |
| 2026-06-25 11:49:19,025 [INFO] Step 680/1314 | Loss: 1.0744 | CE: 1.0744 | LR: 2.56e-05 | Epoch: 2/3 |
| 2026-06-25 11:49:54,045 [INFO] Step 685/1314 | Loss: 1.0929 | CE: 1.0929 | LR: 2.53e-05 | Epoch: 2/3 |
| 2026-06-25 11:50:32,245 [INFO] Step 690/1314 | Loss: 1.0504 | CE: 1.0504 | LR: 2.50e-05 | Epoch: 2/3 |
| 2026-06-25 11:50:59,745 [INFO] Step 695/1314 | Loss: 1.0641 | CE: 1.0641 | LR: 2.47e-05 | Epoch: 2/3 |
| 2026-06-25 11:51:31,481 [INFO] Step 700/1314 | Loss: 1.0638 | CE: 1.0638 | LR: 2.43e-05 | Epoch: 2/3 |
| 2026-06-25 11:52:01,443 [INFO] Step 705/1314 | Loss: 1.0610 | CE: 1.0610 | LR: 2.40e-05 | Epoch: 2/3 |
| 2026-06-25 11:52:37,183 [INFO] Step 710/1314 | Loss: 1.0305 | CE: 1.0305 | LR: 2.37e-05 | Epoch: 2/3 |
| 2026-06-25 11:53:07,651 [INFO] Step 715/1314 | Loss: 1.0960 | CE: 1.0960 | LR: 2.34e-05 | Epoch: 2/3 |
| 2026-06-25 11:53:38,333 [INFO] Step 720/1314 | Loss: 1.0960 | CE: 1.0960 | LR: 2.31e-05 | Epoch: 2/3 |
| 2026-06-25 11:54:08,442 [INFO] Step 725/1314 | Loss: 1.0478 | CE: 1.0478 | LR: 2.28e-05 | Epoch: 2/3 |
| 2026-06-25 11:54:45,093 [INFO] Step 730/1314 | Loss: 1.0634 | CE: 1.0634 | LR: 2.25e-05 | Epoch: 2/3 |
| 2026-06-25 11:55:17,035 [INFO] Step 735/1314 | Loss: 1.0822 | CE: 1.0822 | LR: 2.21e-05 | Epoch: 2/3 |
| 2026-06-25 11:55:52,783 [INFO] Step 740/1314 | Loss: 1.0862 | CE: 1.0862 | LR: 2.18e-05 | Epoch: 2/3 |
| 2026-06-25 11:56:30,194 [INFO] Step 745/1314 | Loss: 1.1552 | CE: 1.1552 | LR: 2.15e-05 | Epoch: 2/3 |
| 2026-06-25 11:57:03,076 [INFO] Step 750/1314 | Loss: 1.0590 | CE: 1.0590 | LR: 2.12e-05 | Epoch: 2/3 |
| 2026-06-25 11:57:30,641 [INFO] Step 755/1314 | Loss: 1.1114 | CE: 1.1114 | LR: 2.09e-05 | Epoch: 2/3 |
| 2026-06-25 11:58:00,587 [INFO] Step 760/1314 | Loss: 1.0799 | CE: 1.0799 | LR: 2.06e-05 | Epoch: 2/3 |
| 2026-06-25 11:58:32,882 [INFO] Step 765/1314 | Loss: 1.0982 | CE: 1.0982 | LR: 2.03e-05 | Epoch: 2/3 |
| 2026-06-25 11:59:15,517 [INFO] Step 770/1314 | Loss: 1.1101 | CE: 1.1101 | LR: 2.00e-05 | Epoch: 2/3 |
| 2026-06-25 11:59:46,893 [INFO] Step 775/1314 | Loss: 1.0595 | CE: 1.0595 | LR: 1.97e-05 | Epoch: 2/3 |
| 2026-06-25 12:00:20,963 [INFO] Step 780/1314 | Loss: 1.0053 | CE: 1.0053 | LR: 1.94e-05 | Epoch: 2/3 |
| 2026-06-25 12:00:52,437 [INFO] Step 785/1314 | Loss: 1.0409 | CE: 1.0409 | LR: 1.91e-05 | Epoch: 2/3 |
| 2026-06-25 12:01:20,396 [INFO] Step 790/1314 | Loss: 1.0445 | CE: 1.0445 | LR: 1.87e-05 | Epoch: 2/3 |
| 2026-06-25 12:01:55,634 [INFO] Step 795/1314 | Loss: 1.0188 | CE: 1.0188 | LR: 1.84e-05 | Epoch: 2/3 |
| 2026-06-25 12:02:23,686 [INFO] Step 800/1314 | Loss: 1.0184 | CE: 1.0184 | LR: 1.81e-05 | Epoch: 2/3 |
| 2026-06-25 12:03:01,948 [INFO] Step 805/1314 | Loss: 1.1393 | CE: 1.1393 | LR: 1.78e-05 | Epoch: 2/3 |
| 2026-06-25 12:03:36,546 [INFO] Step 810/1314 | Loss: 1.0584 | CE: 1.0584 | LR: 1.75e-05 | Epoch: 2/3 |
| 2026-06-25 12:04:09,705 [INFO] Step 815/1314 | Loss: 1.0828 | CE: 1.0828 | LR: 1.72e-05 | Epoch: 2/3 |
| 2026-06-25 12:04:42,813 [INFO] Step 820/1314 | Loss: 1.0408 | CE: 1.0408 | LR: 1.69e-05 | Epoch: 2/3 |
| 2026-06-25 12:05:16,418 [INFO] Step 825/1314 | Loss: 1.0599 | CE: 1.0599 | LR: 1.66e-05 | Epoch: 2/3 |
| 2026-06-25 12:05:49,931 [INFO] Step 830/1314 | Loss: 1.0699 | CE: 1.0699 | LR: 1.63e-05 | Epoch: 2/3 |
| 2026-06-25 12:06:22,098 [INFO] Step 835/1314 | Loss: 1.0250 | CE: 1.0250 | LR: 1.61e-05 | Epoch: 2/3 |
| 2026-06-25 12:06:50,391 [INFO] Step 840/1314 | Loss: 1.0462 | CE: 1.0462 | LR: 1.58e-05 | Epoch: 2/3 |
| 2026-06-25 12:07:18,054 [INFO] Step 845/1314 | Loss: 1.0315 | CE: 1.0315 | LR: 1.55e-05 | Epoch: 2/3 |
| 2026-06-25 12:07:50,366 [INFO] Step 850/1314 | Loss: 1.0482 | CE: 1.0482 | LR: 1.52e-05 | Epoch: 2/3 |
| 2026-06-25 12:08:26,112 [INFO] Step 855/1314 | Loss: 1.0981 | CE: 1.0981 | LR: 1.49e-05 | Epoch: 2/3 |
| 2026-06-25 12:08:56,033 [INFO] Step 860/1314 | Loss: 1.0842 | CE: 1.0842 | LR: 1.46e-05 | Epoch: 2/3 |
| 2026-06-25 12:09:31,636 [INFO] Step 865/1314 | Loss: 1.0539 | CE: 1.0539 | LR: 1.43e-05 | Epoch: 2/3 |
| 2026-06-25 12:10:16,406 [INFO] Step 870/1314 | Loss: 1.1842 | CE: 1.1842 | LR: 1.40e-05 | Epoch: 2/3 |
| 2026-06-25 12:10:46,870 [INFO] Step 875/1314 | Loss: 1.0815 | CE: 1.0815 | LR: 1.38e-05 | Epoch: 3/3 |
| 2026-06-25 12:12:06,395 [INFO] Step 876 | Eval Loss: 2.7204 |
| 2026-06-25 12:17:07,150 [INFO] Seed: 42 |
| 2026-06-25 12:17:07,152 [INFO] Using CUDA: NVIDIA L40 |
| 2026-06-25 12:17:07,152 [INFO] Enabled TF32 matmul/cudnn for CUDA |
| 2026-06-25 12:17:07,314 [INFO] HTTP Request: HEAD https://huggingface.co/Qwen/Qwen3-Embedding-0.6B/resolve/main/config.json "HTTP/1.1 307 Temporary Redirect" |
| 2026-06-25 12:17:07,320 [INFO] HTTP Request: HEAD https://huggingface.co/api/resolve-cache/models/Qwen/Qwen3-Embedding-0.6B/97b0c614be4d77ee51c0cef4e5f07c00f9eb65b3/config.json "HTTP/1.1 200 OK" |
| 2026-06-25 12:17:07,504 [INFO] HTTP Request: HEAD https://huggingface.co/Qwen/Qwen3-Embedding-0.6B/resolve/main/tokenizer_config.json "HTTP/1.1 307 Temporary Redirect" |
| 2026-06-25 12:17:07,510 [INFO] HTTP Request: HEAD https://huggingface.co/api/resolve-cache/models/Qwen/Qwen3-Embedding-0.6B/97b0c614be4d77ee51c0cef4e5f07c00f9eb65b3/tokenizer_config.json "HTTP/1.1 200 OK" |
| 2026-06-25 12:17:07,646 [INFO] HTTP Request: GET https://huggingface.co/api/models/Qwen/Qwen3-Embedding-0.6B/tree/main/additional_chat_templates?recursive=false&expand=false "HTTP/1.1 404 Not Found" |
| 2026-06-25 12:17:07,783 [INFO] HTTP Request: GET https://huggingface.co/api/models/Qwen/Qwen3-Embedding-0.6B/tree/main?recursive=true&expand=false "HTTP/1.1 200 OK" |
| 2026-06-25 12:17:08,499 [INFO] HTTP Request: GET https://huggingface.co/api/models/Qwen/Qwen3-Embedding-0.6B "HTTP/1.1 200 OK" |
| 2026-06-25 12:17:08,661 [INFO] HTTP Request: HEAD https://huggingface.co/Qwen/Qwen3-Embedding-0.6B/resolve/main/config.json "HTTP/1.1 307 Temporary Redirect" |
| 2026-06-25 12:17:08,667 [INFO] HTTP Request: HEAD https://huggingface.co/api/resolve-cache/models/Qwen/Qwen3-Embedding-0.6B/97b0c614be4d77ee51c0cef4e5f07c00f9eb65b3/config.json "HTTP/1.1 200 OK" |
| 2026-06-25 12:17:08,903 [INFO] HTTP Request: HEAD https://huggingface.co/Qwen/Qwen3-Embedding-0.6B/resolve/main/config.json "HTTP/1.1 307 Temporary Redirect" |
| 2026-06-25 12:17:08,910 [INFO] HTTP Request: HEAD https://huggingface.co/api/resolve-cache/models/Qwen/Qwen3-Embedding-0.6B/97b0c614be4d77ee51c0cef4e5f07c00f9eb65b3/config.json "HTTP/1.1 200 OK" |
| 2026-06-25 12:17:15,747 [INFO] LLM2Seq Model Summary |
| ================================================== |
| Encoder: Qwen/Qwen3-Embedding-0.6B |
| Trainable: True |
| LoRA: True |
| Adaptor: gated_residual |
| Layer Fusion: True |
| Fusion Type: tokenwise |
| Salience Gate: True |
| EncStack: True |
| Global Tokens: 16 |
| Decoder: |
| Layers: 8 |
| Hidden Size: 1024 |
| Heads: 16 |
| FFN Size: 4096 |
| ================================================== |
| Total params: 1,000,232,449 |
| Trainable params: 404,455,937 |
| Frozen params: 595,776,512 |
| 2026-06-25 12:17:16,423 [INFO] Enabled gradient checkpointing for encoder |
| 2026-06-25 12:17:16,777 [INFO] Resuming from checkpoint: runs/phase2_lora_encoder_qwen_wikilingua/best.pt |
| 2026-06-25 12:17:18,029 [INFO] Checkpoint load allowed 310 missing keys for stage=phase2_lora_encoder_qwen_wikilingua_tiep context=resume, prefixes=('encoder.',) |
| 2026-06-25 12:17:18,030 [INFO] Reset global step because training.reset_global_step_on_resume=true |
| 2026-06-25 12:17:18,030 [INFO] Reset best eval loss because training.reset_best_eval_loss_on_resume=true |
| 2026-06-25 12:17:18,030 [INFO] Resumed at step 0 |
| 2026-06-25 12:17:18,185 [INFO] HTTP Request: POST https://huggingface.co/api/repos/create "HTTP/1.1 409 Conflict" |
| 2026-06-25 12:17:18,186 [INFO] Hugging Face uploads enabled: repo=BienKieu/llm2seq-wikilingua path=checkpoints/phase2_lora_encoder_qwen_wikilingua |
| 2026-06-25 12:17:18,186 [INFO] Train file: llm2seq/data/processed/train.jsonl (13,999 examples) |
| 2026-06-25 12:17:18,186 [INFO] Eval file: llm2seq/data/processed/validation.jsonl (1,680 examples) |
| 2026-06-25 12:17:18,186 [INFO] Starting training for 2 epochs (876 optimizer steps)... |
| 2026-06-25 12:17:18,186 [INFO] Batch size: 8 |
| 2026-06-25 12:17:18,186 [INFO] Gradient accumulation: 4 |
| 2026-06-25 12:17:18,186 [INFO] Effective batch size: 32 |
| 2026-06-25 12:17:18,186 [INFO] Steps per epoch: 438 |
| 2026-06-25 12:17:18,186 [INFO] Warmup steps: 43 |
| 2026-06-25 12:17:18,186 [INFO] Eval every: 438 optimizer steps |
| 2026-06-25 12:17:18,186 [INFO] Save every: 438 optimizer steps |
| 2026-06-25 12:17:18,186 [INFO] Mixed precision: fp16=False, bf16=True |
| 2026-06-25 12:17:18,187 [INFO] Initial optimizer LRs: adaptor=0.00e+00, decoder=0.00e+00, encoder=0.00e+00 |
| 2026-06-25 12:17:53,833 [INFO] Step 5/876 | Loss: 1.1469 | CE: 1.1469 | LR: 5.81e-06 | Epoch: 1/2 |
| 2026-06-25 12:18:28,554 [INFO] Step 10/876 | Loss: 1.1246 | CE: 1.1246 | LR: 1.16e-05 | Epoch: 1/2 |
| 2026-06-25 12:18:55,190 [INFO] Step 15/876 | Loss: 1.1770 | CE: 1.1770 | LR: 1.74e-05 | Epoch: 1/2 |
| 2026-06-25 12:19:32,923 [INFO] Step 20/876 | Loss: 1.1770 | CE: 1.1770 | LR: 2.33e-05 | Epoch: 1/2 |
| 2026-06-25 12:20:06,351 [INFO] Step 25/876 | Loss: 1.0719 | CE: 1.0719 | LR: 2.91e-05 | Epoch: 1/2 |
| 2026-06-25 12:20:31,972 [INFO] Step 30/876 | Loss: 1.0259 | CE: 1.0259 | LR: 3.49e-05 | Epoch: 1/2 |
| 2026-06-25 12:21:00,550 [INFO] Step 35/876 | Loss: 0.9809 | CE: 0.9809 | LR: 4.07e-05 | Epoch: 1/2 |
| 2026-06-25 12:21:31,736 [INFO] Step 40/876 | Loss: 0.9678 | CE: 0.9678 | LR: 4.65e-05 | Epoch: 1/2 |
| 2026-06-25 12:22:08,473 [INFO] Step 45/876 | Loss: 0.9542 | CE: 0.9542 | LR: 5.00e-05 | Epoch: 1/2 |
| 2026-06-25 12:22:38,313 [INFO] Step 50/876 | Loss: 0.9198 | CE: 0.9198 | LR: 5.00e-05 | Epoch: 1/2 |
| 2026-06-25 12:23:07,570 [INFO] Step 55/876 | Loss: 0.8811 | CE: 0.8811 | LR: 5.00e-05 | Epoch: 1/2 |
| 2026-06-25 12:23:40,618 [INFO] Step 60/876 | Loss: 0.8912 | CE: 0.8912 | LR: 4.99e-05 | Epoch: 1/2 |
| 2026-06-25 12:24:13,328 [INFO] Step 65/876 | Loss: 0.7963 | CE: 0.7963 | LR: 4.99e-05 | Epoch: 1/2 |
| 2026-06-25 12:24:45,377 [INFO] Step 70/876 | Loss: 0.8171 | CE: 0.8171 | LR: 4.99e-05 | Epoch: 1/2 |
| 2026-06-25 12:25:18,076 [INFO] Step 75/876 | Loss: 0.7703 | CE: 0.7703 | LR: 4.98e-05 | Epoch: 1/2 |
| 2026-06-25 12:25:56,914 [INFO] Step 80/876 | Loss: 0.8153 | CE: 0.8153 | LR: 4.98e-05 | Epoch: 1/2 |
| 2026-06-25 12:26:26,129 [INFO] Step 85/876 | Loss: 0.7535 | CE: 0.7535 | LR: 4.97e-05 | Epoch: 1/2 |
| 2026-06-25 12:26:54,762 [INFO] Step 90/876 | Loss: 0.7338 | CE: 0.7338 | LR: 4.96e-05 | Epoch: 1/2 |
| 2026-06-25 12:27:24,258 [INFO] Step 95/876 | Loss: 0.7791 | CE: 0.7791 | LR: 4.95e-05 | Epoch: 1/2 |
| 2026-06-25 12:27:57,283 [INFO] Step 100/876 | Loss: 0.7514 | CE: 0.7514 | LR: 4.94e-05 | Epoch: 1/2 |
| 2026-06-25 12:28:24,751 [INFO] Step 105/876 | Loss: 0.8108 | CE: 0.8108 | LR: 4.93e-05 | Epoch: 1/2 |
| 2026-06-25 12:28:57,393 [INFO] Step 110/876 | Loss: 0.7927 | CE: 0.7927 | LR: 4.92e-05 | Epoch: 1/2 |
| 2026-06-25 12:29:39,280 [INFO] Step 115/876 | Loss: 0.7540 | CE: 0.7540 | LR: 4.91e-05 | Epoch: 1/2 |
| 2026-06-25 12:30:11,220 [INFO] Step 120/876 | Loss: 0.7562 | CE: 0.7562 | LR: 4.90e-05 | Epoch: 1/2 |
| 2026-06-25 12:30:49,770 [INFO] Step 125/876 | Loss: 0.7578 | CE: 0.7578 | LR: 4.88e-05 | Epoch: 1/2 |
| 2026-06-25 12:31:24,990 [INFO] Step 130/876 | Loss: 0.7958 | CE: 0.7958 | LR: 4.87e-05 | Epoch: 1/2 |
| 2026-06-25 12:31:59,385 [INFO] Step 135/876 | Loss: 0.8040 | CE: 0.8040 | LR: 4.85e-05 | Epoch: 1/2 |
| 2026-06-25 12:32:36,359 [INFO] Step 140/876 | Loss: 0.7411 | CE: 0.7411 | LR: 4.83e-05 | Epoch: 1/2 |
| 2026-06-25 12:33:09,419 [INFO] Step 145/876 | Loss: 0.7738 | CE: 0.7738 | LR: 4.82e-05 | Epoch: 1/2 |
| 2026-06-25 12:33:38,113 [INFO] Step 150/876 | Loss: 0.6960 | CE: 0.6960 | LR: 4.80e-05 | Epoch: 1/2 |
| 2026-06-25 12:34:06,053 [INFO] Step 155/876 | Loss: 0.8285 | CE: 0.8285 | LR: 4.78e-05 | Epoch: 1/2 |
| 2026-06-25 12:34:39,905 [INFO] Step 160/876 | Loss: 0.7504 | CE: 0.7504 | LR: 4.76e-05 | Epoch: 1/2 |
| 2026-06-25 12:35:14,812 [INFO] Step 165/876 | Loss: 0.7828 | CE: 0.7828 | LR: 4.74e-05 | Epoch: 1/2 |
| 2026-06-25 12:35:52,587 [INFO] Step 170/876 | Loss: 0.6965 | CE: 0.6965 | LR: 4.72e-05 | Epoch: 1/2 |
| 2026-06-25 12:36:29,090 [INFO] Step 175/876 | Loss: 0.8126 | CE: 0.8126 | LR: 4.70e-05 | Epoch: 1/2 |
| 2026-06-25 12:37:00,549 [INFO] Step 180/876 | Loss: 0.7264 | CE: 0.7264 | LR: 4.67e-05 | Epoch: 1/2 |
| 2026-06-25 12:37:34,419 [INFO] Step 185/876 | Loss: 0.7558 | CE: 0.7558 | LR: 4.65e-05 | Epoch: 1/2 |
| 2026-06-25 12:38:08,999 [INFO] Step 190/876 | Loss: 0.7921 | CE: 0.7921 | LR: 4.63e-05 | Epoch: 1/2 |
| 2026-06-25 12:38:41,317 [INFO] Step 195/876 | Loss: 0.7703 | CE: 0.7703 | LR: 4.60e-05 | Epoch: 1/2 |
| 2026-06-25 12:39:12,543 [INFO] Step 200/876 | Loss: 0.7483 | CE: 0.7483 | LR: 4.57e-05 | Epoch: 1/2 |
| 2026-06-25 12:39:46,399 [INFO] Step 205/876 | Loss: 0.7828 | CE: 0.7828 | LR: 4.55e-05 | Epoch: 1/2 |
| 2026-06-25 12:40:16,025 [INFO] Step 210/876 | Loss: 0.7614 | CE: 0.7614 | LR: 4.52e-05 | Epoch: 1/2 |
| 2026-06-25 12:40:50,049 [INFO] Step 215/876 | Loss: 0.7305 | CE: 0.7305 | LR: 4.49e-05 | Epoch: 1/2 |
| 2026-06-25 12:41:15,206 [INFO] Step 220/876 | Loss: 0.6887 | CE: 0.6887 | LR: 4.46e-05 | Epoch: 1/2 |
| 2026-06-25 12:41:46,769 [INFO] Step 225/876 | Loss: 0.7394 | CE: 0.7394 | LR: 4.43e-05 | Epoch: 1/2 |
| 2026-06-25 12:42:17,711 [INFO] Step 230/876 | Loss: 0.7862 | CE: 0.7862 | LR: 4.40e-05 | Epoch: 1/2 |
| 2026-06-25 12:42:46,810 [INFO] Step 235/876 | Loss: 0.7118 | CE: 0.7118 | LR: 4.37e-05 | Epoch: 1/2 |
| 2026-06-25 12:43:16,181 [INFO] Step 240/876 | Loss: 0.7983 | CE: 0.7983 | LR: 4.34e-05 | Epoch: 1/2 |
| 2026-06-25 12:43:53,697 [INFO] Step 245/876 | Loss: 0.8313 | CE: 0.8313 | LR: 4.31e-05 | Epoch: 1/2 |
| 2026-06-25 12:44:26,205 [INFO] Step 250/876 | Loss: 0.7233 | CE: 0.7233 | LR: 4.28e-05 | Epoch: 1/2 |
| 2026-06-25 12:44:53,255 [INFO] Step 255/876 | Loss: 0.6789 | CE: 0.6789 | LR: 4.24e-05 | Epoch: 1/2 |
| 2026-06-25 12:45:29,725 [INFO] Step 260/876 | Loss: 0.7869 | CE: 0.7869 | LR: 4.21e-05 | Epoch: 1/2 |
| 2026-06-25 12:45:59,860 [INFO] Step 265/876 | Loss: 0.8162 | CE: 0.8162 | LR: 4.17e-05 | Epoch: 1/2 |
| 2026-06-25 12:46:29,949 [INFO] Step 270/876 | Loss: 0.7815 | CE: 0.7815 | LR: 4.14e-05 | Epoch: 1/2 |
| 2026-06-25 12:47:06,307 [INFO] Step 275/876 | Loss: 0.7788 | CE: 0.7788 | LR: 4.10e-05 | Epoch: 1/2 |
| 2026-06-25 12:47:34,352 [INFO] Step 280/876 | Loss: 0.7321 | CE: 0.7321 | LR: 4.07e-05 | Epoch: 1/2 |
| 2026-06-25 12:48:12,829 [INFO] Step 285/876 | Loss: 0.8614 | CE: 0.8614 | LR: 4.03e-05 | Epoch: 1/2 |
| 2026-06-25 12:48:46,212 [INFO] Step 290/876 | Loss: 0.8278 | CE: 0.8278 | LR: 3.99e-05 | Epoch: 1/2 |
| 2026-06-25 12:49:22,516 [INFO] Step 295/876 | Loss: 0.8367 | CE: 0.8367 | LR: 3.95e-05 | Epoch: 1/2 |
| 2026-06-25 12:49:52,143 [INFO] Step 300/876 | Loss: 0.7896 | CE: 0.7896 | LR: 3.91e-05 | Epoch: 1/2 |
| 2026-06-25 12:50:27,869 [INFO] Step 305/876 | Loss: 0.8530 | CE: 0.8530 | LR: 3.88e-05 | Epoch: 1/2 |
| 2026-06-25 12:51:03,928 [INFO] Step 310/876 | Loss: 0.7734 | CE: 0.7734 | LR: 3.84e-05 | Epoch: 1/2 |
| 2026-06-25 12:51:36,109 [INFO] Step 315/876 | Loss: 0.7914 | CE: 0.7914 | LR: 3.80e-05 | Epoch: 1/2 |
| 2026-06-25 12:52:16,544 [INFO] Step 320/876 | Loss: 0.8215 | CE: 0.8215 | LR: 3.76e-05 | Epoch: 1/2 |
| 2026-06-25 12:52:47,744 [INFO] Step 325/876 | Loss: 0.8091 | CE: 0.8091 | LR: 3.71e-05 | Epoch: 1/2 |
| 2026-06-25 12:53:18,227 [INFO] Step 330/876 | Loss: 0.8102 | CE: 0.8102 | LR: 3.67e-05 | Epoch: 1/2 |
| 2026-06-25 12:53:51,346 [INFO] Step 335/876 | Loss: 0.7677 | CE: 0.7677 | LR: 3.63e-05 | Epoch: 1/2 |
| 2026-06-25 12:54:20,052 [INFO] Step 340/876 | Loss: 0.7925 | CE: 0.7925 | LR: 3.59e-05 | Epoch: 1/2 |
| 2026-06-25 12:54:50,084 [INFO] Step 345/876 | Loss: 0.7576 | CE: 0.7576 | LR: 3.55e-05 | Epoch: 1/2 |
| 2026-06-25 12:55:20,164 [INFO] Step 350/876 | Loss: 0.7477 | CE: 0.7477 | LR: 3.50e-05 | Epoch: 1/2 |
| 2026-06-25 12:55:51,472 [INFO] Step 355/876 | Loss: 0.8388 | CE: 0.8388 | LR: 3.46e-05 | Epoch: 1/2 |
| 2026-06-25 12:56:33,494 [INFO] Step 360/876 | Loss: 0.7876 | CE: 0.7876 | LR: 3.42e-05 | Epoch: 1/2 |
| 2026-06-25 12:57:13,108 [INFO] Step 365/876 | Loss: 0.8618 | CE: 0.8618 | LR: 3.37e-05 | Epoch: 1/2 |
| 2026-06-25 12:57:43,781 [INFO] Step 370/876 | Loss: 0.7917 | CE: 0.7917 | LR: 3.33e-05 | Epoch: 1/2 |
| 2026-06-25 12:58:16,681 [INFO] Step 375/876 | Loss: 0.8142 | CE: 0.8142 | LR: 3.28e-05 | Epoch: 1/2 |
| 2026-06-25 12:58:51,010 [INFO] Step 380/876 | Loss: 0.8301 | CE: 0.8301 | LR: 3.24e-05 | Epoch: 1/2 |
| 2026-06-25 12:59:22,799 [INFO] Step 385/876 | Loss: 0.7648 | CE: 0.7648 | LR: 3.19e-05 | Epoch: 1/2 |
| 2026-06-25 12:59:51,137 [INFO] Step 390/876 | Loss: 0.8340 | CE: 0.8340 | LR: 3.15e-05 | Epoch: 1/2 |
| 2026-06-25 13:00:27,074 [INFO] Step 395/876 | Loss: 0.8094 | CE: 0.8094 | LR: 3.10e-05 | Epoch: 1/2 |
| 2026-06-25 13:01:03,965 [INFO] Step 400/876 | Loss: 0.7836 | CE: 0.7836 | LR: 3.06e-05 | Epoch: 1/2 |
| 2026-06-25 13:01:36,378 [INFO] Step 405/876 | Loss: 0.7796 | CE: 0.7796 | LR: 3.01e-05 | Epoch: 1/2 |
| 2026-06-25 13:02:06,293 [INFO] Step 410/876 | Loss: 0.7862 | CE: 0.7862 | LR: 2.96e-05 | Epoch: 1/2 |
| 2026-06-25 13:02:45,209 [INFO] Step 415/876 | Loss: 0.8292 | CE: 0.8292 | LR: 2.92e-05 | Epoch: 1/2 |
| 2026-06-25 13:03:20,551 [INFO] Step 420/876 | Loss: 0.8806 | CE: 0.8806 | LR: 2.87e-05 | Epoch: 1/2 |
| 2026-06-25 13:03:51,816 [INFO] Step 425/876 | Loss: 0.8715 | CE: 0.8715 | LR: 2.82e-05 | Epoch: 1/2 |
| 2026-06-25 13:04:25,024 [INFO] Step 430/876 | Loss: 1.0088 | CE: 1.0088 | LR: 2.78e-05 | Epoch: 1/2 |
| 2026-06-25 13:05:03,563 [INFO] Step 435/876 | Loss: 1.1248 | CE: 1.1248 | LR: 2.73e-05 | Epoch: 1/2 |
| 2026-06-25 13:06:34,425 [INFO] Step 438 | Eval Loss: 2.8371 |
| 2026-06-25 13:06:36,398 [INFO] Saved trainable-only checkpoint to runs/phase2_lora_encoder_qwen_wikilingua/best.pt (575 model tensors, stores_base_encoder_weights=False) |
| 2026-06-25 13:06:37,932 [INFO] New best eval loss: 2.8371 |
| 2026-06-25 13:06:39,909 [INFO] Saved trainable-only checkpoint to runs/phase2_lora_encoder_qwen_wikilingua/epoch_1.pt (575 model tensors, stores_base_encoder_weights=False) |
| 2026-06-25 13:06:41,439 [INFO] Saved epoch 1 checkpoint: runs/phase2_lora_encoder_qwen_wikilingua/epoch_1.pt |
| 2026-06-25 13:06:42,877 [INFO] Uploading to HF: runs/phase2_lora_encoder_qwen_wikilingua/epoch_1.pt -> BienKieu/llm2seq-wikilingua/checkpoints/phase2_lora_encoder_qwen_wikilingua/epochs/epoch_001_step_438.pt |
| 2026-06-25 13:06:43,144 [INFO] HTTP Request: POST https://huggingface.co/api/models/BienKieu/llm2seq-wikilingua/preupload/main "HTTP/1.1 200 OK" |
| 2026-06-25 13:07:02,346 [INFO] HTTP Request: POST https://huggingface.co/api/models/BienKieu/llm2seq-wikilingua/commit/main "HTTP/1.1 200 OK" |
| 2026-06-25 13:07:02,347 [INFO] Uploading to HF: runs/phase2_lora_encoder_qwen_wikilingua/config.yaml -> BienKieu/llm2seq-wikilingua/checkpoints/phase2_lora_encoder_qwen_wikilingua/config.yaml |
| 2026-06-25 13:07:02,520 [INFO] HTTP Request: POST https://huggingface.co/api/models/BienKieu/llm2seq-wikilingua/preupload/main "HTTP/1.1 200 OK" |
| 2026-06-25 13:07:02,982 [INFO] HTTP Request: POST https://huggingface.co/api/models/BienKieu/llm2seq-wikilingua/commit/main "HTTP/1.1 200 OK" |
| 2026-06-25 13:07:02,982 [INFO] Uploading to HF: runs/phase2_lora_encoder_qwen_wikilingua/train.log -> BienKieu/llm2seq-wikilingua/checkpoints/phase2_lora_encoder_qwen_wikilingua/train.log |
| 2026-06-25 13:07:03,157 [INFO] HTTP Request: POST https://huggingface.co/api/models/BienKieu/llm2seq-wikilingua/preupload/main "HTTP/1.1 200 OK" |
| 2026-06-25 13:07:03,769 [INFO] HTTP Request: POST https://huggingface.co/api/models/BienKieu/llm2seq-wikilingua/commit/main "HTTP/1.1 200 OK" |
| 2026-06-25 13:07:03,770 [INFO] Uploading to HF: runs/phase2_lora_encoder_qwen_wikilingua/checkpoint_manifest.json -> BienKieu/llm2seq-wikilingua/checkpoints/phase2_lora_encoder_qwen_wikilingua/checkpoint_manifest.json |
| 2026-06-25 13:07:03,954 [INFO] HTTP Request: POST https://huggingface.co/api/models/BienKieu/llm2seq-wikilingua/preupload/main "HTTP/1.1 200 OK" |
| 2026-06-25 13:07:03,955 [WARNING] No files have been modified since last commit. Skipping to prevent empty commit. |
| 2026-06-25 13:07:04,103 [INFO] HTTP Request: GET https://huggingface.co/api/models/BienKieu/llm2seq-wikilingua/revision/main "HTTP/1.1 200 OK" |
| 2026-06-25 13:07:05,543 [INFO] Uploading to HF: runs/phase2_lora_encoder_qwen_wikilingua/best.pt -> BienKieu/llm2seq-wikilingua/checkpoints/phase2_lora_encoder_qwen_wikilingua/best.pt |
| 2026-06-25 13:07:05,709 [INFO] HTTP Request: POST https://huggingface.co/api/models/BienKieu/llm2seq-wikilingua/preupload/main "HTTP/1.1 200 OK" |
| 2026-06-25 13:07:16,176 [INFO] HTTP Request: POST https://huggingface.co/api/models/BienKieu/llm2seq-wikilingua/commit/main "HTTP/1.1 200 OK" |
| 2026-06-25 13:07:28,293 [INFO] Step 440/876 | Loss: 1.0481 | CE: 1.0481 | LR: 2.68e-05 | Epoch: 2/2 |
| 2026-06-25 13:08:05,851 [INFO] Step 445/876 | Loss: 0.9572 | CE: 0.9572 | LR: 2.64e-05 | Epoch: 2/2 |
| 2026-06-25 13:08:41,978 [INFO] Step 450/876 | Loss: 0.9171 | CE: 0.9171 | LR: 2.59e-05 | Epoch: 2/2 |
| 2026-06-25 13:09:16,574 [INFO] Step 455/876 | Loss: 0.9959 | CE: 0.9959 | LR: 2.54e-05 | Epoch: 2/2 |
| 2026-06-25 13:09:49,317 [INFO] Step 460/876 | Loss: 0.9336 | CE: 0.9336 | LR: 2.50e-05 | Epoch: 2/2 |
| 2026-06-25 13:10:15,509 [INFO] Step 465/876 | Loss: 0.9317 | CE: 0.9317 | LR: 2.45e-05 | Epoch: 2/2 |
| 2026-06-25 13:10:43,946 [INFO] Step 470/876 | Loss: 0.8930 | CE: 0.8930 | LR: 2.40e-05 | Epoch: 2/2 |
| 2026-06-25 13:11:18,404 [INFO] Step 475/876 | Loss: 0.9817 | CE: 0.9817 | LR: 2.35e-05 | Epoch: 2/2 |
| 2026-06-25 13:11:55,274 [INFO] Step 480/876 | Loss: 0.9801 | CE: 0.9801 | LR: 2.31e-05 | Epoch: 2/2 |
| 2026-06-25 13:12:28,462 [INFO] Step 485/876 | Loss: 0.9966 | CE: 0.9966 | LR: 2.26e-05 | Epoch: 2/2 |
| 2026-06-25 13:13:15,152 [INFO] Step 490/876 | Loss: 0.9548 | CE: 0.9548 | LR: 2.21e-05 | Epoch: 2/2 |
| 2026-06-25 13:13:51,930 [INFO] Step 495/876 | Loss: 1.0046 | CE: 1.0046 | LR: 2.17e-05 | Epoch: 2/2 |
| 2026-06-25 13:14:31,774 [INFO] Step 500/876 | Loss: 0.9345 | CE: 0.9345 | LR: 2.12e-05 | Epoch: 2/2 |
| 2026-06-25 13:15:05,380 [INFO] Step 505/876 | Loss: 0.8942 | CE: 0.8942 | LR: 2.07e-05 | Epoch: 2/2 |
| 2026-06-25 13:15:38,655 [INFO] Step 510/876 | Loss: 0.9612 | CE: 0.9612 | LR: 2.03e-05 | Epoch: 2/2 |
| 2026-06-25 13:16:07,666 [INFO] Step 515/876 | Loss: 0.9557 | CE: 0.9557 | LR: 1.98e-05 | Epoch: 2/2 |
| 2026-06-25 13:16:36,465 [INFO] Step 520/876 | Loss: 0.9236 | CE: 0.9236 | LR: 1.93e-05 | Epoch: 2/2 |
| 2026-06-25 13:17:15,819 [INFO] Step 525/876 | Loss: 0.9467 | CE: 0.9467 | LR: 1.89e-05 | Epoch: 2/2 |
| 2026-06-25 13:17:55,365 [INFO] Step 530/876 | Loss: 0.8808 | CE: 0.8808 | LR: 1.84e-05 | Epoch: 2/2 |
| 2026-06-25 13:18:29,330 [INFO] Step 535/876 | Loss: 0.9588 | CE: 0.9588 | LR: 1.80e-05 | Epoch: 2/2 |
| 2026-06-25 13:19:05,654 [INFO] Step 540/876 | Loss: 0.9688 | CE: 0.9688 | LR: 1.75e-05 | Epoch: 2/2 |
| 2026-06-25 13:19:34,753 [INFO] Step 545/876 | Loss: 0.9391 | CE: 0.9391 | LR: 1.71e-05 | Epoch: 2/2 |
| 2026-06-25 13:20:05,248 [INFO] Step 550/876 | Loss: 0.9413 | CE: 0.9413 | LR: 1.66e-05 | Epoch: 2/2 |
| 2026-06-25 13:20:35,625 [INFO] Step 555/876 | Loss: 0.9270 | CE: 0.9270 | LR: 1.62e-05 | Epoch: 2/2 |
| 2026-06-25 13:21:05,307 [INFO] Step 560/876 | Loss: 0.9118 | CE: 0.9118 | LR: 1.57e-05 | Epoch: 2/2 |
| 2026-06-25 13:21:39,707 [INFO] Step 565/876 | Loss: 0.9413 | CE: 0.9413 | LR: 1.53e-05 | Epoch: 2/2 |
| 2026-06-25 13:22:14,154 [INFO] Step 570/876 | Loss: 0.9483 | CE: 0.9483 | LR: 1.49e-05 | Epoch: 2/2 |
| 2026-06-25 13:22:43,019 [INFO] Step 575/876 | Loss: 0.9603 | CE: 0.9603 | LR: 1.45e-05 | Epoch: 2/2 |
| 2026-06-25 13:23:14,980 [INFO] Step 580/876 | Loss: 0.8935 | CE: 0.8935 | LR: 1.40e-05 | Epoch: 2/2 |
| 2026-06-25 13:23:48,661 [INFO] Step 585/876 | Loss: 0.9275 | CE: 0.9275 | LR: 1.36e-05 | Epoch: 2/2 |
| 2026-06-25 13:24:23,922 [INFO] Step 590/876 | Loss: 0.8906 | CE: 0.8906 | LR: 1.32e-05 | Epoch: 2/2 |
| 2026-06-25 13:24:54,557 [INFO] Step 595/876 | Loss: 0.9981 | CE: 0.9981 | LR: 1.28e-05 | Epoch: 2/2 |
| 2026-06-25 13:25:22,405 [INFO] Step 600/876 | Loss: 0.9478 | CE: 0.9478 | LR: 1.24e-05 | Epoch: 2/2 |
| 2026-06-25 13:25:53,768 [INFO] Step 605/876 | Loss: 0.9366 | CE: 0.9366 | LR: 1.20e-05 | Epoch: 2/2 |
| 2026-06-25 13:26:29,758 [INFO] Step 610/876 | Loss: 0.9987 | CE: 0.9987 | LR: 1.16e-05 | Epoch: 2/2 |
| 2026-06-25 13:27:04,278 [INFO] Step 615/876 | Loss: 0.9159 | CE: 0.9159 | LR: 1.12e-05 | Epoch: 2/2 |
| 2026-06-25 13:27:36,777 [INFO] Step 620/876 | Loss: 0.8865 | CE: 0.8865 | LR: 1.08e-05 | Epoch: 2/2 |
| 2026-06-25 13:28:16,297 [INFO] Step 625/876 | Loss: 0.9068 | CE: 0.9068 | LR: 1.04e-05 | Epoch: 2/2 |
| 2026-06-25 13:28:54,511 [INFO] Step 630/876 | Loss: 0.9141 | CE: 0.9141 | LR: 1.00e-05 | Epoch: 2/2 |
| 2026-06-25 13:29:26,147 [INFO] Step 635/876 | Loss: 0.9231 | CE: 0.9231 | LR: 9.63e-06 | Epoch: 2/2 |
| 2026-06-25 13:29:58,795 [INFO] Step 640/876 | Loss: 0.9408 | CE: 0.9408 | LR: 9.27e-06 | Epoch: 2/2 |
| 2026-06-25 13:30:29,671 [INFO] Step 645/876 | Loss: 0.9363 | CE: 0.9363 | LR: 8.90e-06 | Epoch: 2/2 |
| 2026-06-25 13:30:55,321 [INFO] Step 650/876 | Loss: 0.9086 | CE: 0.9086 | LR: 8.54e-06 | Epoch: 2/2 |
| 2026-06-25 13:31:34,755 [INFO] Step 655/876 | Loss: 0.9958 | CE: 0.9958 | LR: 8.19e-06 | Epoch: 2/2 |
| 2026-06-25 13:32:06,848 [INFO] Step 660/876 | Loss: 0.9772 | CE: 0.9772 | LR: 7.85e-06 | Epoch: 2/2 |
| 2026-06-25 13:32:46,518 [INFO] Step 665/876 | Loss: 0.9745 | CE: 0.9745 | LR: 7.51e-06 | Epoch: 2/2 |
| 2026-06-25 13:33:22,136 [INFO] Step 670/876 | Loss: 0.9166 | CE: 0.9166 | LR: 7.17e-06 | Epoch: 2/2 |
| 2026-06-25 13:33:54,939 [INFO] Step 675/876 | Loss: 0.9516 | CE: 0.9516 | LR: 6.85e-06 | Epoch: 2/2 |
| 2026-06-25 13:34:24,951 [INFO] Step 680/876 | Loss: 0.9407 | CE: 0.9407 | LR: 6.52e-06 | Epoch: 2/2 |
| 2026-06-25 13:34:59,912 [INFO] Step 685/876 | Loss: 0.9694 | CE: 0.9694 | LR: 6.21e-06 | Epoch: 2/2 |
| 2026-06-25 13:35:38,016 [INFO] Step 690/876 | Loss: 0.9175 | CE: 0.9175 | LR: 5.90e-06 | Epoch: 2/2 |
| 2026-06-25 13:36:05,414 [INFO] Step 695/876 | Loss: 0.9295 | CE: 0.9295 | LR: 5.60e-06 | Epoch: 2/2 |
| 2026-06-25 13:36:37,082 [INFO] Step 700/876 | Loss: 0.9362 | CE: 0.9362 | LR: 5.31e-06 | Epoch: 2/2 |
| 2026-06-25 13:37:06,985 [INFO] Step 705/876 | Loss: 0.9315 | CE: 0.9315 | LR: 5.02e-06 | Epoch: 2/2 |
| 2026-06-25 13:37:42,710 [INFO] Step 710/876 | Loss: 0.9054 | CE: 0.9054 | LR: 5.00e-06 | Epoch: 2/2 |
| 2026-06-25 13:38:13,166 [INFO] Step 715/876 | Loss: 0.9611 | CE: 0.9611 | LR: 5.00e-06 | Epoch: 2/2 |
| 2026-06-25 13:38:43,823 [INFO] Step 720/876 | Loss: 0.9577 | CE: 0.9577 | LR: 5.00e-06 | Epoch: 2/2 |
| 2026-06-25 13:39:13,899 [INFO] Step 725/876 | Loss: 0.9207 | CE: 0.9207 | LR: 5.00e-06 | Epoch: 2/2 |
| 2026-06-25 13:39:50,473 [INFO] Step 730/876 | Loss: 0.9440 | CE: 0.9440 | LR: 5.00e-06 | Epoch: 2/2 |
| 2026-06-25 13:40:22,365 [INFO] Step 735/876 | Loss: 0.9464 | CE: 0.9464 | LR: 5.00e-06 | Epoch: 2/2 |
| 2026-06-25 13:40:58,019 [INFO] Step 740/876 | Loss: 0.9538 | CE: 0.9538 | LR: 5.00e-06 | Epoch: 2/2 |
| 2026-06-25 13:41:35,471 [INFO] Step 745/876 | Loss: 1.0175 | CE: 1.0175 | LR: 5.00e-06 | Epoch: 2/2 |
| 2026-06-25 13:42:08,446 [INFO] Step 750/876 | Loss: 0.9310 | CE: 0.9310 | LR: 5.00e-06 | Epoch: 2/2 |
| 2026-06-25 13:42:36,051 [INFO] Step 755/876 | Loss: 0.9782 | CE: 0.9782 | LR: 5.00e-06 | Epoch: 2/2 |
| 2026-06-25 13:43:06,056 [INFO] Step 760/876 | Loss: 0.9575 | CE: 0.9575 | LR: 5.00e-06 | Epoch: 2/2 |
| 2026-06-25 13:43:38,410 [INFO] Step 765/876 | Loss: 0.9689 | CE: 0.9689 | LR: 5.00e-06 | Epoch: 2/2 |
| 2026-06-25 13:44:21,023 [INFO] Step 770/876 | Loss: 0.9786 | CE: 0.9786 | LR: 5.00e-06 | Epoch: 2/2 |
| 2026-06-25 13:44:52,402 [INFO] Step 775/876 | Loss: 0.9339 | CE: 0.9339 | LR: 5.00e-06 | Epoch: 2/2 |
| 2026-06-25 13:45:26,505 [INFO] Step 780/876 | Loss: 0.8858 | CE: 0.8858 | LR: 5.00e-06 | Epoch: 2/2 |
| 2026-06-25 13:45:58,064 [INFO] Step 785/876 | Loss: 0.9162 | CE: 0.9162 | LR: 5.00e-06 | Epoch: 2/2 |
| 2026-06-25 13:46:26,010 [INFO] Step 790/876 | Loss: 0.9180 | CE: 0.9180 | LR: 5.00e-06 | Epoch: 2/2 |
| 2026-06-25 13:47:01,220 [INFO] Step 795/876 | Loss: 0.8986 | CE: 0.8986 | LR: 5.00e-06 | Epoch: 2/2 |
| 2026-06-25 13:47:29,298 [INFO] Step 800/876 | Loss: 0.8913 | CE: 0.8913 | LR: 5.00e-06 | Epoch: 2/2 |
| 2026-06-25 13:48:07,516 [INFO] Step 805/876 | Loss: 0.9999 | CE: 0.9999 | LR: 5.00e-06 | Epoch: 2/2 |
| 2026-06-25 13:48:42,037 [INFO] Step 810/876 | Loss: 0.9329 | CE: 0.9329 | LR: 5.00e-06 | Epoch: 2/2 |
| 2026-06-25 13:49:15,146 [INFO] Step 815/876 | Loss: 0.9532 | CE: 0.9532 | LR: 5.00e-06 | Epoch: 2/2 |
| 2026-06-25 13:49:48,260 [INFO] Step 820/876 | Loss: 0.9271 | CE: 0.9271 | LR: 5.00e-06 | Epoch: 2/2 |
| 2026-06-25 13:50:21,847 [INFO] Step 825/876 | Loss: 0.9426 | CE: 0.9426 | LR: 5.00e-06 | Epoch: 2/2 |
| 2026-06-25 13:50:55,369 [INFO] Step 830/876 | Loss: 0.9478 | CE: 0.9478 | LR: 5.00e-06 | Epoch: 2/2 |
| 2026-06-25 13:51:27,589 [INFO] Step 835/876 | Loss: 0.8968 | CE: 0.8968 | LR: 5.00e-06 | Epoch: 2/2 |
| 2026-06-25 13:51:55,937 [INFO] Step 840/876 | Loss: 0.9110 | CE: 0.9110 | LR: 5.00e-06 | Epoch: 2/2 |
| 2026-06-25 13:52:23,577 [INFO] Step 845/876 | Loss: 0.9074 | CE: 0.9074 | LR: 5.00e-06 | Epoch: 2/2 |
| 2026-06-25 13:52:55,898 [INFO] Step 850/876 | Loss: 0.9235 | CE: 0.9235 | LR: 5.00e-06 | Epoch: 2/2 |
| 2026-06-25 13:53:31,703 [INFO] Step 855/876 | Loss: 0.9684 | CE: 0.9684 | LR: 5.00e-06 | Epoch: 2/2 |
| 2026-06-25 13:54:01,724 [INFO] Step 860/876 | Loss: 0.9564 | CE: 0.9564 | LR: 5.00e-06 | Epoch: 2/2 |
| 2026-06-25 13:54:37,428 [INFO] Step 865/876 | Loss: 0.9307 | CE: 0.9307 | LR: 5.00e-06 | Epoch: 2/2 |
| 2026-06-25 13:55:22,224 [INFO] Step 870/876 | Loss: 1.0528 | CE: 1.0528 | LR: 5.00e-06 | Epoch: 2/2 |
| 2026-06-25 13:55:52,714 [INFO] Step 875/876 | Loss: 0.9604 | CE: 0.9604 | LR: 5.00e-06 | Epoch: 3/2 |
| 2026-06-25 13:57:12,780 [INFO] Step 876 | Eval Loss: 2.8109 |
| 2026-06-25 13:57:15,260 [INFO] Saved trainable-only checkpoint to runs/phase2_lora_encoder_qwen_wikilingua/best.pt (575 model tensors, stores_base_encoder_weights=False) |
| 2026-06-25 13:57:16,668 [INFO] New best eval loss: 2.8109 |
| 2026-06-25 13:57:18,563 [INFO] Saved trainable-only checkpoint to runs/phase2_lora_encoder_qwen_wikilingua/epoch_2.pt (575 model tensors, stores_base_encoder_weights=False) |
| 2026-06-25 13:57:19,944 [INFO] Saved epoch 2 checkpoint: runs/phase2_lora_encoder_qwen_wikilingua/epoch_2.pt |
| 2026-06-25 13:57:21,216 [INFO] Uploading to HF: runs/phase2_lora_encoder_qwen_wikilingua/epoch_2.pt -> BienKieu/llm2seq-wikilingua/checkpoints/phase2_lora_encoder_qwen_wikilingua/epochs/epoch_002_step_876.pt |
| 2026-06-25 13:57:21,472 [INFO] HTTP Request: POST https://huggingface.co/api/models/BienKieu/llm2seq-wikilingua/preupload/main "HTTP/1.1 200 OK" |
| 2026-06-25 13:57:39,806 [INFO] HTTP Request: POST https://huggingface.co/api/models/BienKieu/llm2seq-wikilingua/commit/main "HTTP/1.1 200 OK" |
| 2026-06-25 13:57:39,806 [INFO] Uploading to HF: runs/phase2_lora_encoder_qwen_wikilingua/config.yaml -> BienKieu/llm2seq-wikilingua/checkpoints/phase2_lora_encoder_qwen_wikilingua/config.yaml |
| 2026-06-25 13:57:39,969 [INFO] HTTP Request: POST https://huggingface.co/api/models/BienKieu/llm2seq-wikilingua/preupload/main "HTTP/1.1 200 OK" |
| 2026-06-25 13:57:39,970 [WARNING] No files have been modified since last commit. Skipping to prevent empty commit. |
| 2026-06-25 13:57:40,109 [INFO] HTTP Request: GET https://huggingface.co/api/models/BienKieu/llm2seq-wikilingua/revision/main "HTTP/1.1 200 OK" |
| 2026-06-25 13:57:40,110 [INFO] Uploading to HF: runs/phase2_lora_encoder_qwen_wikilingua/train.log -> BienKieu/llm2seq-wikilingua/checkpoints/phase2_lora_encoder_qwen_wikilingua/train.log |
| 2026-06-25 13:57:40,260 [INFO] HTTP Request: POST https://huggingface.co/api/models/BienKieu/llm2seq-wikilingua/preupload/main "HTTP/1.1 200 OK" |
| 2026-06-25 13:57:40,722 [INFO] HTTP Request: POST https://huggingface.co/api/models/BienKieu/llm2seq-wikilingua/commit/main "HTTP/1.1 200 OK" |
| 2026-06-25 13:57:40,723 [INFO] Uploading to HF: runs/phase2_lora_encoder_qwen_wikilingua/checkpoint_manifest.json -> BienKieu/llm2seq-wikilingua/checkpoints/phase2_lora_encoder_qwen_wikilingua/checkpoint_manifest.json |
| 2026-06-25 13:57:40,889 [INFO] HTTP Request: POST https://huggingface.co/api/models/BienKieu/llm2seq-wikilingua/preupload/main "HTTP/1.1 200 OK" |
| 2026-06-25 13:57:40,890 [WARNING] No files have been modified since last commit. Skipping to prevent empty commit. |
| 2026-06-25 13:57:41,027 [INFO] HTTP Request: GET https://huggingface.co/api/models/BienKieu/llm2seq-wikilingua/revision/main "HTTP/1.1 200 OK" |
| 2026-06-25 13:57:42,304 [INFO] Uploading to HF: runs/phase2_lora_encoder_qwen_wikilingua/best.pt -> BienKieu/llm2seq-wikilingua/checkpoints/phase2_lora_encoder_qwen_wikilingua/best.pt |
| 2026-06-25 13:57:42,461 [INFO] HTTP Request: POST https://huggingface.co/api/models/BienKieu/llm2seq-wikilingua/preupload/main "HTTP/1.1 200 OK" |
| 2026-06-25 13:57:52,357 [INFO] HTTP Request: POST https://huggingface.co/api/models/BienKieu/llm2seq-wikilingua/commit/main "HTTP/1.1 200 OK" |
| 2026-06-25 13:57:52,386 [INFO] Deleted old local epoch checkpoint: runs/phase2_lora_encoder_qwen_wikilingua/epoch_1.pt |
| 2026-06-25 13:57:52,386 [INFO] Running ROUGE evaluation at epoch 2 (max 500 samples)... |
| 2026-06-25 14:32:18,978 [INFO] Seed: 42 |
| 2026-06-25 14:32:18,979 [INFO] Using CUDA: NVIDIA L40 |
| 2026-06-25 14:32:18,980 [INFO] Enabled TF32 matmul/cudnn for CUDA |
| 2026-06-25 14:32:19,147 [INFO] HTTP Request: HEAD https://huggingface.co/Qwen/Qwen3-Embedding-0.6B/resolve/main/config.json "HTTP/1.1 307 Temporary Redirect" |
| 2026-06-25 14:32:19,154 [INFO] HTTP Request: HEAD https://huggingface.co/api/resolve-cache/models/Qwen/Qwen3-Embedding-0.6B/97b0c614be4d77ee51c0cef4e5f07c00f9eb65b3/config.json "HTTP/1.1 200 OK" |
| 2026-06-25 14:32:19,294 [INFO] HTTP Request: HEAD https://huggingface.co/Qwen/Qwen3-Embedding-0.6B/resolve/main/tokenizer_config.json "HTTP/1.1 307 Temporary Redirect" |
| 2026-06-25 14:32:19,300 [INFO] HTTP Request: HEAD https://huggingface.co/api/resolve-cache/models/Qwen/Qwen3-Embedding-0.6B/97b0c614be4d77ee51c0cef4e5f07c00f9eb65b3/tokenizer_config.json "HTTP/1.1 200 OK" |
| 2026-06-25 14:32:19,439 [INFO] HTTP Request: GET https://huggingface.co/api/models/Qwen/Qwen3-Embedding-0.6B/tree/main/additional_chat_templates?recursive=false&expand=false "HTTP/1.1 404 Not Found" |
| 2026-06-25 14:32:19,576 [INFO] HTTP Request: GET https://huggingface.co/api/models/Qwen/Qwen3-Embedding-0.6B/tree/main?recursive=true&expand=false "HTTP/1.1 200 OK" |
| 2026-06-25 14:32:20,543 [INFO] HTTP Request: GET https://huggingface.co/api/models/Qwen/Qwen3-Embedding-0.6B "HTTP/1.1 200 OK" |
| 2026-06-25 14:32:20,716 [INFO] HTTP Request: HEAD https://huggingface.co/Qwen/Qwen3-Embedding-0.6B/resolve/main/config.json "HTTP/1.1 307 Temporary Redirect" |
| 2026-06-25 14:32:20,726 [INFO] HTTP Request: HEAD https://huggingface.co/api/resolve-cache/models/Qwen/Qwen3-Embedding-0.6B/97b0c614be4d77ee51c0cef4e5f07c00f9eb65b3/config.json "HTTP/1.1 200 OK" |
| 2026-06-25 14:32:20,989 [INFO] HTTP Request: HEAD https://huggingface.co/Qwen/Qwen3-Embedding-0.6B/resolve/main/config.json "HTTP/1.1 307 Temporary Redirect" |
| 2026-06-25 14:32:20,996 [INFO] HTTP Request: HEAD https://huggingface.co/api/resolve-cache/models/Qwen/Qwen3-Embedding-0.6B/97b0c614be4d77ee51c0cef4e5f07c00f9eb65b3/config.json "HTTP/1.1 200 OK" |
| 2026-06-25 14:32:28,958 [INFO] LLM2Seq Model Summary |
| ================================================== |
| Encoder: Qwen/Qwen3-Embedding-0.6B |
| Trainable: True |
| LoRA: True |
| Adaptor: gated_residual |
| Layer Fusion: True |
| Fusion Type: tokenwise |
| Salience Gate: True |
| EncStack: True |
| Global Tokens: 16 |
| Decoder: |
| Layers: 8 |
| Hidden Size: 1024 |
| Heads: 16 |
| FFN Size: 4096 |
| ================================================== |
| Total params: 1,000,232,449 |
| Trainable params: 404,455,937 |
| Frozen params: 595,776,512 |
| 2026-06-25 14:32:29,912 [INFO] Enabled gradient checkpointing for encoder |
| 2026-06-25 14:32:30,362 [INFO] Resuming from checkpoint: runs/phase2_lora_encoder_qwen_wikilingua/best.pt |
| 2026-06-25 14:32:31,797 [INFO] Checkpoint load allowed 310 missing keys for stage=phase2_lora_encoder_qwen_wikilingua_tiep context=resume, prefixes=('encoder.',) |
| 2026-06-25 14:32:31,797 [INFO] Reset global step because training.reset_global_step_on_resume=true |
| 2026-06-25 14:32:31,797 [INFO] Reset best eval loss because training.reset_best_eval_loss_on_resume=true |
| 2026-06-25 14:32:31,797 [INFO] Resumed at step 0 |
| 2026-06-25 14:32:31,961 [INFO] HTTP Request: POST https://huggingface.co/api/repos/create "HTTP/1.1 409 Conflict" |
| 2026-06-25 14:32:31,962 [INFO] Hugging Face uploads enabled: repo=BienKieu/llm2seq-wikilingua path=checkpoints/phase2_lora_encoder_qwen_wikilingua |
| 2026-06-25 14:32:31,962 [INFO] Train file: llm2seq/data/processed/train.jsonl (13,999 examples) |
| 2026-06-25 14:32:31,962 [INFO] Eval file: llm2seq/data/processed/validation.jsonl (1,680 examples) |
| 2026-06-25 14:32:31,962 [INFO] Starting training for 3 epochs (1,314 optimizer steps)... |
| 2026-06-25 14:32:31,963 [INFO] Batch size: 8 |
| 2026-06-25 14:32:31,963 [INFO] Gradient accumulation: 4 |
| 2026-06-25 14:32:31,963 [INFO] Effective batch size: 32 |
| 2026-06-25 14:32:31,963 [INFO] Steps per epoch: 438 |
| 2026-06-25 14:32:31,963 [INFO] Warmup steps: 65 |
| 2026-06-25 14:32:31,963 [INFO] Eval every: 438 optimizer steps |
| 2026-06-25 14:32:31,963 [INFO] Save every: 438 optimizer steps |
| 2026-06-25 14:32:31,963 [INFO] Mixed precision: fp16=False, bf16=True |
| 2026-06-25 14:32:31,963 [INFO] Initial optimizer LRs: adaptor=0.00e+00, decoder=0.00e+00, encoder=0.00e+00 |
| 2026-06-25 14:33:08,219 [INFO] Step 5/1314 | Loss: 0.9635 | CE: 0.9635 | LR: 3.85e-06 | Epoch: 1/3 |
| 2026-06-25 14:33:43,418 [INFO] Step 10/1314 | Loss: 0.9186 | CE: 0.9186 | LR: 7.69e-06 | Epoch: 1/3 |
| 2026-06-25 14:34:10,671 [INFO] Step 15/1314 | Loss: 0.9296 | CE: 0.9296 | LR: 1.15e-05 | Epoch: 1/3 |
| 2026-06-25 14:34:49,234 [INFO] Step 20/1314 | Loss: 0.9152 | CE: 0.9152 | LR: 1.54e-05 | Epoch: 1/3 |
| 2026-06-25 14:35:23,251 [INFO] Step 25/1314 | Loss: 0.7991 | CE: 0.7991 | LR: 1.92e-05 | Epoch: 1/3 |
| 2026-06-25 14:35:49,159 [INFO] Step 30/1314 | Loss: 0.7311 | CE: 0.7311 | LR: 2.31e-05 | Epoch: 1/3 |
| 2026-06-25 14:36:17,943 [INFO] Step 35/1314 | Loss: 0.6724 | CE: 0.6724 | LR: 2.69e-05 | Epoch: 1/3 |
| 2026-06-25 14:36:49,292 [INFO] Step 40/1314 | Loss: 0.6478 | CE: 0.6478 | LR: 3.08e-05 | Epoch: 1/3 |
| 2026-06-25 14:37:26,376 [INFO] Step 45/1314 | Loss: 0.6369 | CE: 0.6369 | LR: 3.46e-05 | Epoch: 1/3 |
| 2026-06-25 14:37:56,576 [INFO] Step 50/1314 | Loss: 0.5983 | CE: 0.5983 | LR: 3.85e-05 | Epoch: 1/3 |
| 2026-06-25 14:38:26,124 [INFO] Step 55/1314 | Loss: 0.5871 | CE: 0.5871 | LR: 4.23e-05 | Epoch: 1/3 |
| 2026-06-25 14:38:59,480 [INFO] Step 60/1314 | Loss: 0.5744 | CE: 0.5744 | LR: 4.62e-05 | Epoch: 1/3 |
| 2026-06-25 14:39:32,483 [INFO] Step 65/1314 | Loss: 0.5212 | CE: 0.5212 | LR: 5.00e-05 | Epoch: 1/3 |
| 2026-06-25 14:40:04,852 [INFO] Step 70/1314 | Loss: 0.5320 | CE: 0.5320 | LR: 5.00e-05 | Epoch: 1/3 |
| 2026-06-25 14:40:37,811 [INFO] Step 75/1314 | Loss: 0.4820 | CE: 0.4820 | LR: 5.00e-05 | Epoch: 1/3 |
| 2026-06-25 14:41:16,535 [INFO] Step 80/1314 | Loss: 0.5322 | CE: 0.5322 | LR: 5.00e-05 | Epoch: 1/3 |
| 2026-06-25 14:41:45,191 [INFO] Step 85/1314 | Loss: 0.4807 | CE: 0.4807 | LR: 5.00e-05 | Epoch: 1/3 |
| 2026-06-25 14:42:13,498 [INFO] Step 90/1314 | Loss: 0.4724 | CE: 0.4724 | LR: 5.00e-05 | Epoch: 1/3 |
| 2026-06-25 14:42:43,224 [INFO] Step 95/1314 | Loss: 0.5068 | CE: 0.5068 | LR: 4.99e-05 | Epoch: 1/3 |
| 2026-06-25 14:43:16,534 [INFO] Step 100/1314 | Loss: 0.4933 | CE: 0.4933 | LR: 4.99e-05 | Epoch: 1/3 |
| 2026-06-25 14:43:44,315 [INFO] Step 105/1314 | Loss: 0.5311 | CE: 0.5311 | LR: 4.99e-05 | Epoch: 1/3 |
| 2026-06-25 14:44:17,099 [INFO] Step 110/1314 | Loss: 0.5274 | CE: 0.5274 | LR: 4.98e-05 | Epoch: 1/3 |
| 2026-06-25 14:44:59,120 [INFO] Step 115/1314 | Loss: 0.4937 | CE: 0.4937 | LR: 4.98e-05 | Epoch: 1/3 |
| 2026-06-25 14:45:31,190 [INFO] Step 120/1314 | Loss: 0.4940 | CE: 0.4940 | LR: 4.98e-05 | Epoch: 1/3 |
| 2026-06-25 14:46:10,001 [INFO] Step 125/1314 | Loss: 0.5020 | CE: 0.5020 | LR: 4.97e-05 | Epoch: 1/3 |
| 2026-06-25 14:46:45,472 [INFO] Step 130/1314 | Loss: 0.5317 | CE: 0.5317 | LR: 4.97e-05 | Epoch: 1/3 |
| 2026-06-25 14:47:20,008 [INFO] Step 135/1314 | Loss: 0.5492 | CE: 0.5492 | LR: 4.96e-05 | Epoch: 1/3 |
| 2026-06-25 14:47:57,176 [INFO] Step 140/1314 | Loss: 0.4953 | CE: 0.4953 | LR: 4.96e-05 | Epoch: 1/3 |
| 2026-06-25 14:48:30,426 [INFO] Step 145/1314 | Loss: 0.5194 | CE: 0.5194 | LR: 4.95e-05 | Epoch: 1/3 |
| 2026-06-25 14:48:59,341 [INFO] Step 150/1314 | Loss: 0.4449 | CE: 0.4449 | LR: 4.94e-05 | Epoch: 1/3 |
| 2026-06-25 14:49:27,469 [INFO] Step 155/1314 | Loss: 0.5620 | CE: 0.5620 | LR: 4.94e-05 | Epoch: 1/3 |
| 2026-06-25 14:50:01,529 [INFO] Step 160/1314 | Loss: 0.5062 | CE: 0.5062 | LR: 4.93e-05 | Epoch: 1/3 |
| 2026-06-25 14:50:36,621 [INFO] Step 165/1314 | Loss: 0.5348 | CE: 0.5348 | LR: 4.92e-05 | Epoch: 1/3 |
| 2026-06-25 14:51:14,606 [INFO] Step 170/1314 | Loss: 0.4596 | CE: 0.4596 | LR: 4.91e-05 | Epoch: 1/3 |
| 2026-06-25 14:51:51,300 [INFO] Step 175/1314 | Loss: 0.5517 | CE: 0.5517 | LR: 4.90e-05 | Epoch: 1/3 |
| 2026-06-25 14:52:22,948 [INFO] Step 180/1314 | Loss: 0.4755 | CE: 0.4755 | LR: 4.90e-05 | Epoch: 1/3 |
| 2026-06-25 14:52:57,041 [INFO] Step 185/1314 | Loss: 0.5214 | CE: 0.5214 | LR: 4.89e-05 | Epoch: 1/3 |
| 2026-06-25 14:53:31,738 [INFO] Step 190/1314 | Loss: 0.5569 | CE: 0.5569 | LR: 4.88e-05 | Epoch: 1/3 |
| 2026-06-25 14:54:04,175 [INFO] Step 195/1314 | Loss: 0.5246 | CE: 0.5246 | LR: 4.87e-05 | Epoch: 1/3 |
| 2026-06-25 14:54:35,545 [INFO] Step 200/1314 | Loss: 0.5166 | CE: 0.5166 | LR: 4.86e-05 | Epoch: 1/3 |
| 2026-06-25 14:55:09,529 [INFO] Step 205/1314 | Loss: 0.5498 | CE: 0.5498 | LR: 4.85e-05 | Epoch: 1/3 |
| 2026-06-25 14:55:39,340 [INFO] Step 210/1314 | Loss: 0.5347 | CE: 0.5347 | LR: 4.84e-05 | Epoch: 1/3 |
| 2026-06-25 14:56:13,508 [INFO] Step 215/1314 | Loss: 0.4961 | CE: 0.4961 | LR: 4.82e-05 | Epoch: 1/3 |
| 2026-06-25 14:56:38,839 [INFO] Step 220/1314 | Loss: 0.4688 | CE: 0.4688 | LR: 4.81e-05 | Epoch: 1/3 |
| 2026-06-25 14:57:10,588 [INFO] Step 225/1314 | Loss: 0.5135 | CE: 0.5135 | LR: 4.80e-05 | Epoch: 1/3 |
| 2026-06-25 14:57:41,643 [INFO] Step 230/1314 | Loss: 0.5414 | CE: 0.5414 | LR: 4.79e-05 | Epoch: 1/3 |
| 2026-06-25 14:58:10,885 [INFO] Step 235/1314 | Loss: 0.4985 | CE: 0.4985 | LR: 4.77e-05 | Epoch: 1/3 |
| 2026-06-25 14:58:40,401 [INFO] Step 240/1314 | Loss: 0.5581 | CE: 0.5581 | LR: 4.76e-05 | Epoch: 1/3 |
| 2026-06-25 14:59:18,073 [INFO] Step 245/1314 | Loss: 0.5880 | CE: 0.5880 | LR: 4.75e-05 | Epoch: 1/3 |
| 2026-06-25 14:59:50,707 [INFO] Step 250/1314 | Loss: 0.4943 | CE: 0.4943 | LR: 4.73e-05 | Epoch: 1/3 |
| 2026-06-25 15:00:17,986 [INFO] Step 255/1314 | Loss: 0.4603 | CE: 0.4603 | LR: 4.72e-05 | Epoch: 1/3 |
| 2026-06-25 15:00:54,656 [INFO] Step 260/1314 | Loss: 0.5511 | CE: 0.5511 | LR: 4.71e-05 | Epoch: 1/3 |
| 2026-06-25 15:01:24,984 [INFO] Step 265/1314 | Loss: 0.5765 | CE: 0.5765 | LR: 4.69e-05 | Epoch: 1/3 |
| 2026-06-25 15:01:55,212 [INFO] Step 270/1314 | Loss: 0.5482 | CE: 0.5482 | LR: 4.67e-05 | Epoch: 1/3 |
| 2026-06-25 15:02:31,740 [INFO] Step 275/1314 | Loss: 0.5468 | CE: 0.5468 | LR: 4.66e-05 | Epoch: 1/3 |
| 2026-06-25 15:02:59,875 [INFO] Step 280/1314 | Loss: 0.5146 | CE: 0.5146 | LR: 4.64e-05 | Epoch: 1/3 |
| 2026-06-25 15:03:38,440 [INFO] Step 285/1314 | Loss: 0.6161 | CE: 0.6161 | LR: 4.63e-05 | Epoch: 1/3 |
| 2026-06-25 15:04:11,911 [INFO] Step 290/1314 | Loss: 0.6046 | CE: 0.6046 | LR: 4.61e-05 | Epoch: 1/3 |
| 2026-06-25 15:04:48,277 [INFO] Step 295/1314 | Loss: 0.6142 | CE: 0.6142 | LR: 4.59e-05 | Epoch: 1/3 |
| 2026-06-25 15:05:17,999 [INFO] Step 300/1314 | Loss: 0.5711 | CE: 0.5711 | LR: 4.58e-05 | Epoch: 1/3 |
| 2026-06-25 15:05:53,876 [INFO] Step 305/1314 | Loss: 0.6431 | CE: 0.6431 | LR: 4.56e-05 | Epoch: 1/3 |
| 2026-06-25 15:06:30,145 [INFO] Step 310/1314 | Loss: 0.5616 | CE: 0.5616 | LR: 4.54e-05 | Epoch: 1/3 |
| 2026-06-25 15:07:02,501 [INFO] Step 315/1314 | Loss: 0.5821 | CE: 0.5821 | LR: 4.52e-05 | Epoch: 1/3 |
| 2026-06-25 15:07:43,143 [INFO] Step 320/1314 | Loss: 0.6123 | CE: 0.6123 | LR: 4.50e-05 | Epoch: 1/3 |
| 2026-06-25 15:08:14,523 [INFO] Step 325/1314 | Loss: 0.5998 | CE: 0.5998 | LR: 4.48e-05 | Epoch: 1/3 |
| 2026-06-25 15:08:45,238 [INFO] Step 330/1314 | Loss: 0.5951 | CE: 0.5951 | LR: 4.46e-05 | Epoch: 1/3 |
| 2026-06-25 15:09:18,565 [INFO] Step 335/1314 | Loss: 0.5544 | CE: 0.5544 | LR: 4.45e-05 | Epoch: 1/3 |
| 2026-06-25 15:09:47,432 [INFO] Step 340/1314 | Loss: 0.5872 | CE: 0.5872 | LR: 4.43e-05 | Epoch: 1/3 |
| 2026-06-25 15:10:17,604 [INFO] Step 345/1314 | Loss: 0.5449 | CE: 0.5449 | LR: 4.41e-05 | Epoch: 1/3 |
| 2026-06-25 15:10:47,883 [INFO] Step 350/1314 | Loss: 0.5716 | CE: 0.5716 | LR: 4.38e-05 | Epoch: 1/3 |
| 2026-06-25 15:11:19,378 [INFO] Step 355/1314 | Loss: 0.6207 | CE: 0.6207 | LR: 4.36e-05 | Epoch: 1/3 |
| 2026-06-25 15:12:01,697 [INFO] Step 360/1314 | Loss: 0.5928 | CE: 0.5928 | LR: 4.34e-05 | Epoch: 1/3 |
| 2026-06-25 15:12:41,584 [INFO] Step 365/1314 | Loss: 0.6507 | CE: 0.6507 | LR: 4.32e-05 | Epoch: 1/3 |
| 2026-06-25 15:13:12,484 [INFO] Step 370/1314 | Loss: 0.5941 | CE: 0.5941 | LR: 4.30e-05 | Epoch: 1/3 |
| 2026-06-25 15:13:45,553 [INFO] Step 375/1314 | Loss: 0.6137 | CE: 0.6137 | LR: 4.28e-05 | Epoch: 1/3 |
| 2026-06-25 15:14:20,152 [INFO] Step 380/1314 | Loss: 0.6264 | CE: 0.6264 | LR: 4.26e-05 | Epoch: 1/3 |
| 2026-06-25 15:14:52,196 [INFO] Step 385/1314 | Loss: 0.5747 | CE: 0.5747 | LR: 4.23e-05 | Epoch: 1/3 |
| 2026-06-25 15:15:20,725 [INFO] Step 390/1314 | Loss: 0.6272 | CE: 0.6272 | LR: 4.21e-05 | Epoch: 1/3 |
| 2026-06-25 15:15:57,164 [INFO] Step 395/1314 | Loss: 0.6054 | CE: 0.6054 | LR: 4.19e-05 | Epoch: 1/3 |
| 2026-06-25 15:16:34,347 [INFO] Step 400/1314 | Loss: 0.5988 | CE: 0.5988 | LR: 4.16e-05 | Epoch: 1/3 |
| 2026-06-25 15:17:06,935 [INFO] Step 405/1314 | Loss: 0.5886 | CE: 0.5886 | LR: 4.14e-05 | Epoch: 1/3 |
| 2026-06-25 15:17:37,017 [INFO] Step 410/1314 | Loss: 0.5973 | CE: 0.5973 | LR: 4.12e-05 | Epoch: 1/3 |
| 2026-06-25 15:18:16,191 [INFO] Step 415/1314 | Loss: 0.6419 | CE: 0.6419 | LR: 4.09e-05 | Epoch: 1/3 |
| 2026-06-25 15:18:51,682 [INFO] Step 420/1314 | Loss: 0.6875 | CE: 0.6875 | LR: 4.07e-05 | Epoch: 1/3 |
| 2026-06-25 15:19:23,192 [INFO] Step 425/1314 | Loss: 0.6807 | CE: 0.6807 | LR: 4.04e-05 | Epoch: 1/3 |
| 2026-06-25 15:19:56,519 [INFO] Step 430/1314 | Loss: 0.7905 | CE: 0.7905 | LR: 4.02e-05 | Epoch: 1/3 |
| 2026-06-25 15:20:35,268 [INFO] Step 435/1314 | Loss: 0.8923 | CE: 0.8923 | LR: 3.99e-05 | Epoch: 1/3 |
| 2026-06-25 15:22:05,888 [INFO] Step 438 | Eval Loss: 3.0199 |
| 2026-06-25 15:22:08,143 [INFO] Saved trainable-only checkpoint to runs/phase2_lora_encoder_qwen_wikilingua/best.pt (575 model tensors, stores_base_encoder_weights=False) |
| 2026-06-25 15:22:09,723 [INFO] New best eval loss: 3.0199 |
| 2026-06-25 15:22:11,740 [INFO] Saved trainable-only checkpoint to runs/phase2_lora_encoder_qwen_wikilingua/epoch_1.pt (575 model tensors, stores_base_encoder_weights=False) |
| 2026-06-25 15:22:13,304 [INFO] Saved epoch 1 checkpoint: runs/phase2_lora_encoder_qwen_wikilingua/epoch_1.pt |
| 2026-06-25 15:22:14,859 [INFO] Uploading to HF: runs/phase2_lora_encoder_qwen_wikilingua/epoch_1.pt -> BienKieu/llm2seq-wikilingua/checkpoints/phase2_lora_encoder_qwen_wikilingua/epochs/epoch_001_step_438.pt |
| 2026-06-25 15:22:15,033 [INFO] HTTP Request: POST https://huggingface.co/api/models/BienKieu/llm2seq-wikilingua/preupload/main "HTTP/1.1 200 OK" |
| 2026-06-25 15:22:34,708 [INFO] HTTP Request: POST https://huggingface.co/api/models/BienKieu/llm2seq-wikilingua/commit/main "HTTP/1.1 200 OK" |
| 2026-06-25 15:22:34,709 [INFO] Uploading to HF: runs/phase2_lora_encoder_qwen_wikilingua/config.yaml -> BienKieu/llm2seq-wikilingua/checkpoints/phase2_lora_encoder_qwen_wikilingua/config.yaml |
| 2026-06-25 15:22:34,869 [INFO] HTTP Request: POST https://huggingface.co/api/models/BienKieu/llm2seq-wikilingua/preupload/main "HTTP/1.1 200 OK" |
| 2026-06-25 15:22:35,238 [INFO] HTTP Request: POST https://huggingface.co/api/models/BienKieu/llm2seq-wikilingua/commit/main "HTTP/1.1 200 OK" |
| 2026-06-25 15:22:35,238 [INFO] Uploading to HF: runs/phase2_lora_encoder_qwen_wikilingua/train.log -> BienKieu/llm2seq-wikilingua/checkpoints/phase2_lora_encoder_qwen_wikilingua/train.log |
| 2026-06-25 15:22:35,393 [INFO] HTTP Request: POST https://huggingface.co/api/models/BienKieu/llm2seq-wikilingua/preupload/main "HTTP/1.1 200 OK" |
| 2026-06-25 15:22:35,875 [INFO] HTTP Request: POST https://huggingface.co/api/models/BienKieu/llm2seq-wikilingua/commit/main "HTTP/1.1 200 OK" |
| 2026-06-25 15:22:35,876 [INFO] Uploading to HF: runs/phase2_lora_encoder_qwen_wikilingua/checkpoint_manifest.json -> BienKieu/llm2seq-wikilingua/checkpoints/phase2_lora_encoder_qwen_wikilingua/checkpoint_manifest.json |
| 2026-06-25 15:22:36,043 [INFO] HTTP Request: POST https://huggingface.co/api/models/BienKieu/llm2seq-wikilingua/preupload/main "HTTP/1.1 200 OK" |
| 2026-06-25 15:22:36,044 [WARNING] No files have been modified since last commit. Skipping to prevent empty commit. |
| 2026-06-25 15:22:36,185 [INFO] HTTP Request: GET https://huggingface.co/api/models/BienKieu/llm2seq-wikilingua/revision/main "HTTP/1.1 200 OK" |
| 2026-06-25 15:22:37,651 [INFO] Uploading to HF: runs/phase2_lora_encoder_qwen_wikilingua/best.pt -> BienKieu/llm2seq-wikilingua/checkpoints/phase2_lora_encoder_qwen_wikilingua/best.pt |
| 2026-06-25 15:22:37,803 [INFO] HTTP Request: POST https://huggingface.co/api/models/BienKieu/llm2seq-wikilingua/preupload/main "HTTP/1.1 200 OK" |
| 2026-06-25 15:22:48,540 [INFO] HTTP Request: POST https://huggingface.co/api/models/BienKieu/llm2seq-wikilingua/commit/main "HTTP/1.1 200 OK" |
| 2026-06-25 15:22:48,560 [INFO] Deleted old local epoch checkpoint: runs/phase2_lora_encoder_qwen_wikilingua/epoch_1.pt |
| 2026-06-25 15:23:00,661 [INFO] Step 440/1314 | Loss: 0.7817 | CE: 0.7817 | LR: 3.97e-05 | Epoch: 2/3 |
| 2026-06-25 15:23:37,972 [INFO] Step 445/1314 | Loss: 0.6797 | CE: 0.6797 | LR: 3.94e-05 | Epoch: 2/3 |
| 2026-06-25 15:24:13,546 [INFO] Step 450/1314 | Loss: 0.6367 | CE: 0.6367 | LR: 3.92e-05 | Epoch: 2/3 |
| 2026-06-25 15:24:47,896 [INFO] Step 455/1314 | Loss: 0.7072 | CE: 0.7072 | LR: 3.89e-05 | Epoch: 2/3 |
| 2026-06-25 15:25:20,671 [INFO] Step 460/1314 | Loss: 0.6627 | CE: 0.6627 | LR: 3.86e-05 | Epoch: 2/3 |
| 2026-06-25 15:25:46,967 [INFO] Step 465/1314 | Loss: 0.6533 | CE: 0.6533 | LR: 3.84e-05 | Epoch: 2/3 |
| 2026-06-25 15:26:15,446 [INFO] Step 470/1314 | Loss: 0.6286 | CE: 0.6286 | LR: 3.81e-05 | Epoch: 2/3 |
| 2026-06-25 15:26:50,056 [INFO] Step 475/1314 | Loss: 0.7186 | CE: 0.7186 | LR: 3.78e-05 | Epoch: 2/3 |
| 2026-06-25 15:27:27,593 [INFO] Step 480/1314 | Loss: 0.7121 | CE: 0.7121 | LR: 3.76e-05 | Epoch: 2/3 |
| 2026-06-25 15:28:01,552 [INFO] Step 485/1314 | Loss: 0.7310 | CE: 0.7310 | LR: 3.73e-05 | Epoch: 2/3 |
| 2026-06-25 15:28:48,558 [INFO] Step 490/1314 | Loss: 0.7053 | CE: 0.7053 | LR: 3.70e-05 | Epoch: 2/3 |
| 2026-06-25 15:29:25,370 [INFO] Step 495/1314 | Loss: 0.7334 | CE: 0.7334 | LR: 3.67e-05 | Epoch: 2/3 |
| 2026-06-25 15:30:05,289 [INFO] Step 500/1314 | Loss: 0.7028 | CE: 0.7028 | LR: 3.65e-05 | Epoch: 2/3 |
| 2026-06-25 15:30:39,054 [INFO] Step 505/1314 | Loss: 0.6529 | CE: 0.6529 | LR: 3.62e-05 | Epoch: 2/3 |
| 2026-06-25 15:31:12,405 [INFO] Step 510/1314 | Loss: 0.7157 | CE: 0.7157 | LR: 3.59e-05 | Epoch: 2/3 |
| 2026-06-25 15:31:41,424 [INFO] Step 515/1314 | Loss: 0.7230 | CE: 0.7230 | LR: 3.56e-05 | Epoch: 2/3 |
| 2026-06-25 15:32:10,215 [INFO] Step 520/1314 | Loss: 0.6909 | CE: 0.6909 | LR: 3.53e-05 | Epoch: 2/3 |
| 2026-06-25 15:32:49,593 [INFO] Step 525/1314 | Loss: 0.7147 | CE: 0.7147 | LR: 3.51e-05 | Epoch: 2/3 |
| 2026-06-25 15:33:29,221 [INFO] Step 530/1314 | Loss: 0.6619 | CE: 0.6619 | LR: 3.48e-05 | Epoch: 2/3 |
| 2026-06-25 15:34:03,332 [INFO] Step 535/1314 | Loss: 0.7287 | CE: 0.7287 | LR: 3.45e-05 | Epoch: 2/3 |
| 2026-06-25 15:34:39,827 [INFO] Step 540/1314 | Loss: 0.7410 | CE: 0.7410 | LR: 3.42e-05 | Epoch: 2/3 |
| 2026-06-25 15:35:09,155 [INFO] Step 545/1314 | Loss: 0.7227 | CE: 0.7227 | LR: 3.39e-05 | Epoch: 2/3 |
| 2026-06-25 15:35:39,770 [INFO] Step 550/1314 | Loss: 0.7222 | CE: 0.7222 | LR: 3.36e-05 | Epoch: 2/3 |
| 2026-06-25 15:36:10,286 [INFO] Step 555/1314 | Loss: 0.7214 | CE: 0.7214 | LR: 3.33e-05 | Epoch: 2/3 |
| 2026-06-25 15:36:40,146 [INFO] Step 560/1314 | Loss: 0.7026 | CE: 0.7026 | LR: 3.30e-05 | Epoch: 2/3 |
| 2026-06-25 15:37:14,738 [INFO] Step 565/1314 | Loss: 0.7336 | CE: 0.7336 | LR: 3.27e-05 | Epoch: 2/3 |
| 2026-06-25 15:37:49,348 [INFO] Step 570/1314 | Loss: 0.7467 | CE: 0.7467 | LR: 3.24e-05 | Epoch: 2/3 |
| 2026-06-25 15:38:18,296 [INFO] Step 575/1314 | Loss: 0.7615 | CE: 0.7615 | LR: 3.21e-05 | Epoch: 2/3 |
| 2026-06-25 15:38:50,353 [INFO] Step 580/1314 | Loss: 0.7075 | CE: 0.7075 | LR: 3.18e-05 | Epoch: 2/3 |
| 2026-06-25 15:39:24,259 [INFO] Step 585/1314 | Loss: 0.7346 | CE: 0.7346 | LR: 3.15e-05 | Epoch: 2/3 |
| 2026-06-25 15:39:59,691 [INFO] Step 590/1314 | Loss: 0.7085 | CE: 0.7085 | LR: 3.12e-05 | Epoch: 2/3 |
| 2026-06-25 15:40:30,495 [INFO] Step 595/1314 | Loss: 0.8065 | CE: 0.8065 | LR: 3.09e-05 | Epoch: 2/3 |
| 2026-06-25 15:40:58,525 [INFO] Step 600/1314 | Loss: 0.7649 | CE: 0.7649 | LR: 3.06e-05 | Epoch: 2/3 |
| 2026-06-25 15:41:30,103 [INFO] Step 605/1314 | Loss: 0.7643 | CE: 0.7643 | LR: 3.03e-05 | Epoch: 2/3 |
| 2026-06-25 15:42:06,455 [INFO] Step 610/1314 | Loss: 0.8194 | CE: 0.8194 | LR: 3.00e-05 | Epoch: 2/3 |
| 2026-06-25 15:42:41,174 [INFO] Step 615/1314 | Loss: 0.7452 | CE: 0.7452 | LR: 2.97e-05 | Epoch: 2/3 |
| 2026-06-25 15:43:13,848 [INFO] Step 620/1314 | Loss: 0.7231 | CE: 0.7231 | LR: 2.93e-05 | Epoch: 2/3 |
| 2026-06-25 15:43:53,542 [INFO] Step 625/1314 | Loss: 0.7428 | CE: 0.7428 | LR: 2.90e-05 | Epoch: 2/3 |
| 2026-06-25 15:44:31,938 [INFO] Step 630/1314 | Loss: 0.7511 | CE: 0.7511 | LR: 2.87e-05 | Epoch: 2/3 |
| 2026-06-25 15:45:03,782 [INFO] Step 635/1314 | Loss: 0.7611 | CE: 0.7611 | LR: 2.84e-05 | Epoch: 2/3 |
| 2026-06-25 15:45:36,604 [INFO] Step 640/1314 | Loss: 0.7880 | CE: 0.7880 | LR: 2.81e-05 | Epoch: 2/3 |
| 2026-06-25 15:46:07,643 [INFO] Step 645/1314 | Loss: 0.7733 | CE: 0.7733 | LR: 2.78e-05 | Epoch: 2/3 |
| 2026-06-25 15:46:33,439 [INFO] Step 650/1314 | Loss: 0.7492 | CE: 0.7492 | LR: 2.75e-05 | Epoch: 2/3 |
| 2026-06-25 15:47:13,001 [INFO] Step 655/1314 | Loss: 0.8454 | CE: 0.8454 | LR: 2.72e-05 | Epoch: 2/3 |
| 2026-06-25 15:47:45,316 [INFO] Step 660/1314 | Loss: 0.8146 | CE: 0.8146 | LR: 2.69e-05 | Epoch: 2/3 |
| 2026-06-25 15:48:25,194 [INFO] Step 665/1314 | Loss: 0.8335 | CE: 0.8335 | LR: 2.65e-05 | Epoch: 2/3 |
| 2026-06-25 15:49:00,949 [INFO] Step 670/1314 | Loss: 0.7761 | CE: 0.7761 | LR: 2.62e-05 | Epoch: 2/3 |
| 2026-06-25 15:49:33,883 [INFO] Step 675/1314 | Loss: 0.8086 | CE: 0.8086 | LR: 2.59e-05 | Epoch: 2/3 |
| 2026-06-25 15:50:04,085 [INFO] Step 680/1314 | Loss: 0.7982 | CE: 0.7982 | LR: 2.56e-05 | Epoch: 2/3 |
| 2026-06-25 15:50:39,195 [INFO] Step 685/1314 | Loss: 0.8333 | CE: 0.8333 | LR: 2.53e-05 | Epoch: 2/3 |
| 2026-06-25 15:51:17,424 [INFO] Step 690/1314 | Loss: 0.7773 | CE: 0.7773 | LR: 2.50e-05 | Epoch: 2/3 |
| 2026-06-25 15:51:45,007 [INFO] Step 695/1314 | Loss: 0.7960 | CE: 0.7960 | LR: 2.47e-05 | Epoch: 2/3 |
| 2026-06-25 15:52:16,807 [INFO] Step 700/1314 | Loss: 0.8015 | CE: 0.8015 | LR: 2.43e-05 | Epoch: 2/3 |
| 2026-06-25 15:52:46,891 [INFO] Step 705/1314 | Loss: 0.7949 | CE: 0.7949 | LR: 2.40e-05 | Epoch: 2/3 |
| 2026-06-25 15:53:22,763 [INFO] Step 710/1314 | Loss: 0.7764 | CE: 0.7764 | LR: 2.37e-05 | Epoch: 2/3 |
| 2026-06-25 15:53:53,403 [INFO] Step 715/1314 | Loss: 0.8223 | CE: 0.8223 | LR: 2.34e-05 | Epoch: 2/3 |
| 2026-06-25 15:54:24,247 [INFO] Step 720/1314 | Loss: 0.8237 | CE: 0.8237 | LR: 2.31e-05 | Epoch: 2/3 |
| 2026-06-25 15:54:54,498 [INFO] Step 725/1314 | Loss: 0.7872 | CE: 0.7872 | LR: 2.28e-05 | Epoch: 2/3 |
| 2026-06-25 15:55:31,275 [INFO] Step 730/1314 | Loss: 0.8131 | CE: 0.8131 | LR: 2.25e-05 | Epoch: 2/3 |
| 2026-06-25 15:56:03,379 [INFO] Step 735/1314 | Loss: 0.8157 | CE: 0.8157 | LR: 2.21e-05 | Epoch: 2/3 |
| 2026-06-25 15:56:39,205 [INFO] Step 740/1314 | Loss: 0.8238 | CE: 0.8238 | LR: 2.18e-05 | Epoch: 2/3 |
| 2026-06-25 15:57:16,710 [INFO] Step 745/1314 | Loss: 0.8754 | CE: 0.8754 | LR: 2.15e-05 | Epoch: 2/3 |
| 2026-06-25 15:57:49,748 [INFO] Step 750/1314 | Loss: 0.8011 | CE: 0.8011 | LR: 2.12e-05 | Epoch: 2/3 |
| 2026-06-25 15:58:17,452 [INFO] Step 755/1314 | Loss: 0.8424 | CE: 0.8424 | LR: 2.09e-05 | Epoch: 2/3 |
| 2026-06-25 15:58:47,517 [INFO] Step 760/1314 | Loss: 0.8341 | CE: 0.8341 | LR: 2.06e-05 | Epoch: 2/3 |
| 2026-06-25 15:59:19,977 [INFO] Step 765/1314 | Loss: 0.8354 | CE: 0.8354 | LR: 2.03e-05 | Epoch: 2/3 |
| 2026-06-25 16:00:02,761 [INFO] Step 770/1314 | Loss: 0.8428 | CE: 0.8428 | LR: 2.00e-05 | Epoch: 2/3 |
| 2026-06-25 16:00:34,295 [INFO] Step 775/1314 | Loss: 0.8037 | CE: 0.8037 | LR: 1.97e-05 | Epoch: 2/3 |
| 2026-06-25 16:01:08,516 [INFO] Step 780/1314 | Loss: 0.7577 | CE: 0.7577 | LR: 1.94e-05 | Epoch: 2/3 |
| 2026-06-25 16:01:40,221 [INFO] Step 785/1314 | Loss: 0.7911 | CE: 0.7911 | LR: 1.91e-05 | Epoch: 2/3 |
| 2026-06-25 16:02:08,346 [INFO] Step 790/1314 | Loss: 0.7954 | CE: 0.7954 | LR: 1.87e-05 | Epoch: 2/3 |
| 2026-06-25 16:02:43,685 [INFO] Step 795/1314 | Loss: 0.7744 | CE: 0.7744 | LR: 1.84e-05 | Epoch: 2/3 |
| 2026-06-25 16:03:11,946 [INFO] Step 800/1314 | Loss: 0.7665 | CE: 0.7665 | LR: 1.81e-05 | Epoch: 2/3 |
| 2026-06-25 16:03:50,289 [INFO] Step 805/1314 | Loss: 0.8592 | CE: 0.8592 | LR: 1.78e-05 | Epoch: 2/3 |
| 2026-06-25 16:04:24,964 [INFO] Step 810/1314 | Loss: 0.8053 | CE: 0.8053 | LR: 1.75e-05 | Epoch: 2/3 |
| 2026-06-25 16:04:58,188 [INFO] Step 815/1314 | Loss: 0.8237 | CE: 0.8237 | LR: 1.72e-05 | Epoch: 2/3 |
| 2026-06-25 16:05:31,323 [INFO] Step 820/1314 | Loss: 0.8049 | CE: 0.8049 | LR: 1.69e-05 | Epoch: 2/3 |
| 2026-06-25 16:06:04,967 [INFO] Step 825/1314 | Loss: 0.8193 | CE: 0.8193 | LR: 1.66e-05 | Epoch: 2/3 |
| 2026-06-25 16:06:38,455 [INFO] Step 830/1314 | Loss: 0.8251 | CE: 0.8251 | LR: 1.63e-05 | Epoch: 2/3 |
| 2026-06-25 16:07:10,621 [INFO] Step 835/1314 | Loss: 0.7689 | CE: 0.7689 | LR: 1.61e-05 | Epoch: 2/3 |
| 2026-06-25 16:07:38,977 [INFO] Step 840/1314 | Loss: 0.7804 | CE: 0.7804 | LR: 1.58e-05 | Epoch: 2/3 |
| 2026-06-25 16:08:06,590 [INFO] Step 845/1314 | Loss: 0.7755 | CE: 0.7755 | LR: 1.55e-05 | Epoch: 2/3 |
| 2026-06-25 16:08:38,883 [INFO] Step 850/1314 | Loss: 0.7959 | CE: 0.7959 | LR: 1.52e-05 | Epoch: 2/3 |
| 2026-06-25 16:09:14,682 [INFO] Step 855/1314 | Loss: 0.8318 | CE: 0.8318 | LR: 1.49e-05 | Epoch: 2/3 |
| 2026-06-25 16:09:44,705 [INFO] Step 860/1314 | Loss: 0.8250 | CE: 0.8250 | LR: 1.46e-05 | Epoch: 2/3 |
| 2026-06-25 16:10:20,462 [INFO] Step 865/1314 | Loss: 0.8052 | CE: 0.8052 | LR: 1.43e-05 | Epoch: 2/3 |
| 2026-06-25 16:11:05,334 [INFO] Step 870/1314 | Loss: 0.9280 | CE: 0.9280 | LR: 1.40e-05 | Epoch: 2/3 |
| 2026-06-25 16:11:35,755 [INFO] Step 875/1314 | Loss: 0.8484 | CE: 0.8484 | LR: 1.38e-05 | Epoch: 3/3 |
| 2026-06-25 16:12:56,059 [INFO] Step 876 | Eval Loss: 2.9621 |
| 2026-06-25 16:12:58,790 [INFO] Saved trainable-only checkpoint to runs/phase2_lora_encoder_qwen_wikilingua/best.pt (575 model tensors, stores_base_encoder_weights=False) |
| 2026-06-25 16:13:00,346 [INFO] New best eval loss: 2.9621 |
| 2026-06-25 16:13:02,656 [INFO] Saved trainable-only checkpoint to runs/phase2_lora_encoder_qwen_wikilingua/epoch_2.pt (575 model tensors, stores_base_encoder_weights=False) |
| 2026-06-25 16:13:04,108 [INFO] Saved epoch 2 checkpoint: runs/phase2_lora_encoder_qwen_wikilingua/epoch_2.pt |
| 2026-06-25 16:13:05,453 [INFO] Uploading to HF: runs/phase2_lora_encoder_qwen_wikilingua/epoch_2.pt -> BienKieu/llm2seq-wikilingua/checkpoints/phase2_lora_encoder_qwen_wikilingua/epochs/epoch_002_step_876.pt |
| 2026-06-25 16:13:05,625 [INFO] HTTP Request: POST https://huggingface.co/api/models/BienKieu/llm2seq-wikilingua/preupload/main "HTTP/1.1 200 OK" |
| 2026-06-25 16:13:24,350 [INFO] HTTP Request: POST https://huggingface.co/api/models/BienKieu/llm2seq-wikilingua/commit/main "HTTP/1.1 200 OK" |
| 2026-06-25 16:13:24,351 [INFO] Uploading to HF: runs/phase2_lora_encoder_qwen_wikilingua/config.yaml -> BienKieu/llm2seq-wikilingua/checkpoints/phase2_lora_encoder_qwen_wikilingua/config.yaml |
| 2026-06-25 16:13:24,528 [INFO] HTTP Request: POST https://huggingface.co/api/models/BienKieu/llm2seq-wikilingua/preupload/main "HTTP/1.1 200 OK" |
| 2026-06-25 16:13:24,528 [WARNING] No files have been modified since last commit. Skipping to prevent empty commit. |
| 2026-06-25 16:13:24,664 [INFO] HTTP Request: GET https://huggingface.co/api/models/BienKieu/llm2seq-wikilingua/revision/main "HTTP/1.1 200 OK" |
| 2026-06-25 16:13:24,666 [INFO] Uploading to HF: runs/phase2_lora_encoder_qwen_wikilingua/train.log -> BienKieu/llm2seq-wikilingua/checkpoints/phase2_lora_encoder_qwen_wikilingua/train.log |
| 2026-06-25 16:13:24,840 [INFO] HTTP Request: POST https://huggingface.co/api/models/BienKieu/llm2seq-wikilingua/preupload/main "HTTP/1.1 200 OK" |
|
|