| device cuda |
| dtype fp32 |
| [transformers] The fast path is not available because one of the required library is not installed. Falling back to torch implementation. To install follow https://github.com/fla-org/flash-linear-attention#installation and https://github.com/Dao-AILab/causal-conv1d |
|
Loading weights: 0%| | 0/426 [00:00<?, ?it/s]
Loading weights: 21%|██ | 89/426 [00:00<00:00, 862.68it/s]
Loading weights: 41%|████▏ | 176/426 [00:00<00:00, 789.32it/s]
Loading weights: 60%|██████ | 256/426 [00:00<00:00, 735.83it/s]
Loading weights: 78%|███████▊ | 332/426 [00:00<00:00, 734.89it/s]
Loading weights: 95%|█████████▌| 406/426 [00:00<00:00, 722.33it/s]
Loading weights: 100%|██████████| 426/426 [00:00<00:00, 737.27it/s] |
| trainable 33776128 |
| train encoded 3000 skipped_over_window 0 |
| dev encoded 500 skipped_over_window 0 |
| /root/venv/lib/python3.12/site-packages/triton/backends/nvidia/driver.c:9:10: fatal error: Python.h: No such file or directory |
| 9 | #include <Python.h> |
| | ^~~~~~~~~~ |
| compilation terminated. |
| EVAL {"epoch": 0, "dev_n": 500, "dev_pos": 85, "auc": 0.8983416017009214, "eval_seconds": 33.6, "eval_batched": true, "tpr@0.5": 0.7764705882352941, "fpr@0.5": 0.10120481927710843, "tpr@0.9": 0.2823529411764706, "fpr@0.9": 0.021686746987951807} |
| /root/work/ft_action.py:173: UserWarning: Converting a tensor with requires_grad=True to a scalar may lead to unexpected behavior. |
| Consider using tensor.detach() first. (Triggered internally at /__w/pytorch/pytorch/torch/csrc/autograd/generated/python_variable_methods.cpp:820.) |
| (loss / nq).backward(); tot += float(loss) |
| STEP {"step": 10, "of": 94, "loss": 0.39333093479581294, "lr": 2.9790281926131377e-05, "min": 2.3020944277445476} |
| STEP {"step": 20, "of": 94, "loss": 0.00793515532859601, "lr": 2.815142971168314e-05, "min": 4.684007918834686} |
| STEP {"step": 30, "of": 94, "loss": 0.008664748337650963, "lr": 2.507581554661477e-05, "min": 7.149123617013296} |
| STEP {"step": 40, "of": 94, "loss": 0.14514939102809876, "lr": 2.0942699789226136e-05, "min": 9.776532284418742} |
| STEP {"step": 50, "of": 94, "loss": 0.0899001021462027, "lr": 1.6261745511317765e-05, "min": 12.395656911532084} |
| STEP {"step": 60, "of": 94, "loss": 0.18478131373376527, "lr": 1.1610170888299289e-05, "min": 14.709938764572144} |
| STEP {"step": 70, "of": 94, "loss": 0.12211553269298747, "lr": 7.561571229433195e-06, "min": 16.99185705979665} |
| STEP {"step": 80, "of": 94, "loss": 0.18828649175839018, "lr": 4.615187751213268e-06, "min": 19.12233568827311} |
| STEP {"step": 90, "of": 94, "loss": 0.0011363789153620019, "lr": 3.134345108054867e-06, "min": 21.606734609603883} |
| Warning: You are sending unauthenticated requests to the HF Hub. Please set a HF_TOKEN to enable higher rate limits and faster downloads. |
| SAVED /root/work/ft/secjev_spot2/epoch1 |
| EVAL {"epoch": 1, "dev_n": 500, "dev_pos": 85, "auc": 0.9913536498936925, "eval_seconds": 29.2, "eval_batched": true, "tpr@0.5": 0.9294117647058824, "fpr@0.5": 0.004819277108433735, "tpr@0.9": 0.8470588235294118, "fpr@0.9": 0.0} |
| TRAIN_DONE |
|
|
