gitextract_yhm008o_/ ├── .gitignore ├── README.md ├── deepspeed/ │ └── zero1.json ├── fast_lambda_setup.sh ├── fast_runpod_setup.sh ├── p2l/ │ ├── auto_eval_utils.py │ ├── auto_evals.py │ ├── dataset.py │ ├── endpoint.py │ ├── eval.py │ ├── model.py │ └── train.py ├── probe_barrier.py ├── route/ │ ├── chat.py │ ├── cost_optimizers.py │ ├── datatypes.py │ ├── example_config.yaml │ ├── openai_server.py │ ├── requirements.txt │ ├── routers.py │ └── utils.py ├── serve_requirements.txt ├── train_requirements.txt └── training_configs/ ├── Llama3.1-8B-full-train.yaml ├── Qwen2.5-1.5B-bag-chrono-eps-0.016-04302025.yaml ├── Qwen2.5-1.5B-bag-chrono-eps-0.032-04302025.yaml ├── Qwen2.5-1.5B-bag-chrono-eps-0.06-04302025.yaml ├── Qwen2.5-1.5B-bag-chrono-eps-0.112-04302025.yaml ├── Qwen2.5-1.5B-bag-chrono-eps-0.2-04302025.yaml ├── Qwen2.5-1.5B-bag-full-train-02222025.yaml ├── Qwen2.5-1.5B-full-train.yaml ├── Qwen2.5-1.5B-rk-full-train-half-batch.yaml ├── Qwen2.5-1.5B-rk-full-train.yaml ├── Qwen2.5-3B-bag-full-train-02222025.yaml ├── Qwen2.5-3B-bag-full-train-02242025.yaml ├── Qwen2.5-3B-freeze-test-part-2.yaml ├── Qwen2.5-3B-freeze-test.yaml ├── Qwen2.5-3B-full-train-double-batch.yaml ├── Qwen2.5-3B-full-train.yaml ├── Qwen2.5-3B-rk-full-train-half-batch.yaml ├── Qwen2.5-3B-rk-full-train.yaml ├── Qwen2.5-3B-training-bt_data_11092024 copy.yaml ├── Qwen2.5-7B-bag-full-train-02222025.yaml ├── Qwen2.5-7B-bag-full-train-02242025.yaml ├── Qwen2.5-7B-bag-full-train-03132025.yaml ├── Qwen2.5-7B-bag-full-train-chrono.yaml ├── Qwen2.5-7B-bt-full-train-02222025.yaml ├── Qwen2.5-7B-full-train.yaml ├── Qwen2.5-7B-rk-full-train-abs.yaml ├── Qwen2.5-7B-rk-full-train-half-batch.yaml ├── Qwen2.5-7B-rk-full-train.yaml ├── debug.yaml ├── init_debug_qwen_1.5b_he.yaml ├── init_debug_qwen_1.5b_reset_params.yaml ├── init_debug_qwen_1.5b_xavier.yaml ├── init_debug_qwen_3b_he.yaml ├── init_debug_qwen_3b_reset_params.yaml ├── init_debug_qwen_3b_xavier.yaml └── qwen_1.5B_geom_test.yaml