{ "run_info": { "created_at": "2025-07-31T14:11:12+00:00", "total_time": 2493.9155955019996, "experiment_name": "oft/llama-3.2-3B-rank32", "peft_branch": "main", "train_config": { "model_id": "meta-llama/Llama-3.2-3B", "dtype": "bfloat16", "max_seq_length": 768, "batch_size": 4, "batch_size_eval": 50, "max_steps": 5000, "eval_steps": 250, "compile": false, "query_template": "Question: {query} Think step by step.\nAnswer:", "seed": 0, "grad_norm_clip": 1.0, "optimizer_type": "AdamW", "optimizer_kwargs": { "lr": 0.0001, "weight_decay": 0.1 }, "lr_scheduler": "cosine", "use_amp": false, "autocast_adapter_dtype": true, "generation_kwargs": { "max_length": 800, "max_new_tokens": 300 }, "attn_implementation": null }, "peft_config": { "task_type": null, "peft_type": "OFT", "auto_mapping": null, "base_model_name_or_path": "meta-llama/Llama-3.2-3B", "revision": null, "inference_mode": false, "r": 32, "oft_block_size": 0, "module_dropout": 0.0, "target_modules": [ "v_proj", "q_proj" ], "fan_in_fan_out": false, "bias": "none", "exclude_modules": null, "init_weights": true, "layers_to_transform": null, "layers_pattern": null, "modules_to_save": null, "coft": false, "eps": 6e-05, "block_share": false, "use_cayley_neumann": true, "num_cayley_neumann_terms": 5 }, "error_msg": "" }, "train_info": { "accelerator_memory_reserved_avg": 12057354384, "accelerator_memory_max": 22294822912, "accelerator_memory_reserved_99th": 17939310837, "train_time": 2214.446992367006, "file_size": 32693568, "num_trainable_params": 8171520, "num_total_params": 3220921344, "status": "success", "metrics": [ { "step": 250, "valid accuracy": 0.36, "train loss": 0.9631057088375091, "train samples": 1000, "train time": 43.418166981995455, "eval time": 16.96007740999994, "tokens / sec": 4876.276791873667, "mem allocated avg": 6903823460.352, "mem reserved avg": 12108561383.424, "elapsed time": 113.91408998500083 }, { "step": 500, "valid accuracy": 0.36, "train loss": 0.7144306401014328, "train samples": 2000, "train time": 42.455775934988196, "eval time": 16.150497423999695, "tokens / sec": 4899.097835792689, "mem allocated avg": 6896105342.976, "mem reserved avg": 11994249822.208, "elapsed time": 220.49977440600014 }, { "step": 750, "valid accuracy": 0.52, "train loss": 0.6711842056512832, "train samples": 3000, "train time": 43.15603912099323, "eval time": 10.51256339000065, "tokens / sec": 4968.041654585135, "mem allocated avg": 6906686986.24, "mem reserved avg": 12155101380.608, "elapsed time": 322.5515955810006 }, { "step": 1000, "valid accuracy": 0.48, "train loss": 0.6508683092594146, "train samples": 4000, "train time": 42.42713372799517, "eval time": 16.934662378998837, "tokens / sec": 4910.442485595753, "mem allocated avg": 6897939019.776, "mem reserved avg": 12025262505.984, "elapsed time": 429.7382754350001 }, { "step": 1250, "valid accuracy": 0.4, "train loss": 0.6453732433319092, "train samples": 5000, "train time": 42.549762738994104, "eval time": 16.92903551499876, "tokens / sec": 4901.03790423462, "mem allocated avg": 6897900118.016, "mem reserved avg": 12017234608.128, "elapsed time": 537.135011331 }, { "step": 1500, "valid accuracy": 0.5, "train loss": 0.636857116818428, "train samples": 6000, "train time": 42.7670685170051, "eval time": 16.97714005600028, "tokens / sec": 4894.677312679627, "mem allocated avg": 6899436058.624, "mem reserved avg": 12045822984.192, "elapsed time": 644.8122739440005 }, { "step": 1750, "valid accuracy": 0.48, "train loss": 0.6280697054862976, "train samples": 7000, "train time": 42.93359049599712, "eval time": 11.770931148001182, "tokens / sec": 4876.251848060996, "mem allocated avg": 6900382935.04, "mem reserved avg": 12059630632.96, "elapsed time": 747.525349122001 }, { "step": 2000, "valid accuracy": 0.4, "train loss": 0.6299525223970414, "train samples": 8000, "train time": 42.82682755300084, "eval time": 11.5680384089992, "tokens / sec": 4849.670448808364, "mem allocated avg": 6896952041.472, "mem reserved avg": 12003611508.736, "elapsed time": 849.5279627600012 }, { "step": 2250, "valid accuracy": 0.42, "train loss": 0.6208749743700027, "train samples": 9000, "train time": 43.43083962600576, "eval time": 16.986704689999897, "tokens / sec": 4949.20203825146, "mem allocated avg": 6908628027.392, "mem reserved avg": 12188169273.344, "elapsed time": 958.0240945160003 }, { "step": 2500, "valid accuracy": 0.42, "train loss": 0.6179436918497085, "train samples": 10000, "train time": 42.63891591101674, "eval time": 17.232789900999705, "tokens / sec": 4830.493355643306, "mem allocated avg": 6893492830.208, "mem reserved avg": 11953867063.296, "elapsed time": 1065.2266578140006 }, { "step": 2750, "valid accuracy": 0.42, "train loss": 0.6097300077676773, "train samples": 11000, "train time": 43.157022238001446, "eval time": 17.135427543998958, "tokens / sec": 4909.537058222485, "mem allocated avg": 6904392247.296, "mem reserved avg": 12124977889.28, "elapsed time": 1173.5244531360004 }, { "step": 3000, "valid accuracy": 0.42, "train loss": 0.600518134355545, "train samples": 12000, "train time": 42.90499155000907, "eval time": 17.038416949999373, "tokens / sec": 4864.958422301702, "mem allocated avg": 6898886381.568, "mem reserved avg": 12038994657.28, "elapsed time": 1281.100714346001 }, { "step": 3250, "valid accuracy": 0.54, "train loss": 0.6095727566480637, "train samples": 13000, "train time": 42.991201876006016, "eval time": 17.145920277998812, "tokens / sec": 4905.678157318666, "mem allocated avg": 6900920473.6, "mem reserved avg": 12070426771.456, "elapsed time": 1389.080374264 }, { "step": 3500, "valid accuracy": 0.54, "train loss": 0.59402192902565, "train samples": 14000, "train time": 43.139979139998104, "eval time": 10.18719298600081, "tokens / sec": 4862.079309758545, "mem allocated avg": 6899826102.272, "mem reserved avg": 12054404530.176, "elapsed time": 1490.7450829120007 }, { "step": 3750, "valid accuracy": 0.58, "train loss": 0.5927710949182511, "train samples": 15000, "train time": 43.49427866901169, "eval time": 10.884315328999946, "tokens / sec": 4982.333461582249, "mem allocated avg": 6910839183.36, "mem reserved avg": 12223619530.752, "elapsed time": 1593.6702795590008 }, { "step": 4000, "valid accuracy": 0.52, "train loss": 0.6036465883255004, "train samples": 16000, "train time": 42.54699739801072, "eval time": 10.508950370000093, "tokens / sec": 4803.464697829781, "mem allocated avg": 6892073494.528, "mem reserved avg": 11931788247.04, "elapsed time": 1694.1543825910012 }, { "step": 4250, "valid accuracy": 0.5, "train loss": 0.5904108211994171, "train samples": 17000, "train time": 42.904117188016244, "eval time": 10.362485865000053, "tokens / sec": 4927.009663749569, "mem allocated avg": 6902539771.904, "mem reserved avg": 12087044603.904, "elapsed time": 1795.3652429800004 }, { "step": 4500, "valid accuracy": 0.56, "train loss": 0.5975252593755722, "train samples": 18000, "train time": 42.7045542899923, "eval time": 9.970661539999128, "tokens / sec": 4866.413043179837, "mem allocated avg": 6897064284.16, "mem reserved avg": 12006883065.856, "elapsed time": 1895.7771126360003 }, { "step": 4750, "valid accuracy": 0.54, "train loss": 0.588557964682579, "train samples": 19000, "train time": 42.698231221012975, "eval time": 10.72399718899942, "tokens / sec": 4916.8078863342525, "mem allocated avg": 6900484192.256, "mem reserved avg": 12052575813.632, "elapsed time": 1997.1282366079995 }, { "step": 5000, "valid accuracy": 0.56, "train loss": 0.5946548076868057, "train samples": 20000, "train time": 42.98944765599845, "eval time": 10.321189939999385, "tokens / sec": 4844.909887343902, "mem allocated avg": 6896923324.416, "mem reserved avg": 12004861411.328, "elapsed time": 2098.129397994 }, { "step": 5000, "test accuracy": 0.5056861258529188, "train loss": 0.5946548076868057, "train samples": 20000, "train total tokens": 4198051 } ] }, "meta_info": { "model_info": { "sha": "13afe5124825b4f3751f836b40dafda64c1ed062", "created_at": "2024-09-18T15:23:48+00:00" }, "dataset_info": { "metamath": { "sha": "aa4f34d3d2d3231299b5b03d9b3e5a20da45aa18", "created_at": "2023-09-21T17:22:46+00:00" }, "gsm8k": { "sha": "e53f048856ff4f594e959d75785d2c2d37b678ee", "created_at": "2022-04-12T10:22:10+00:00" } }, "package_info": { "transformers-version": "4.52.4", "transformers-commit-hash": null, "peft-version": "0.16.1.dev0", "peft-commit-hash": "25e5c6b25c4589eb2683484ede1ba3d985d8a760", "datasets-version": "3.6.0", "datasets-commit-hash": null, "bitsandbytes-version": "0.46.0", "bitsandbytes-commit-hash": null, "torch-version": "2.7.1+cu126", "torch-commit-hash": null }, "system_info": { "system": "Linux", "release": "6.8.0-1031-aws", "version": "#33-Ubuntu SMP Fri Jun 20 18:11:07 UTC 2025", "machine": "x86_64", "processor": "x86_64", "accelerator": "NVIDIA L40S" }, "pytorch_info": "PyTorch built with:\n - GCC 11.2\n - C++ Version: 201703\n - Intel(R) oneAPI Math Kernel Library Version 2024.2-Product Build 20240605 for Intel(R) 64 architecture applications\n - Intel(R) MKL-DNN v3.7.1 (Git Hash 8d263e693366ef8db40acc569cc7d8edf644556d)\n - OpenMP 201511 (a.k.a. OpenMP 4.5)\n - LAPACK is enabled (usually provided by MKL)\n - NNPACK is enabled\n - CPU capability usage: AVX2\n - CUDA Runtime 12.6\n - NVCC architecture flags: -gencode;arch=compute_50,code=sm_50;-gencode;arch=compute_60,code=sm_60;-gencode;arch=compute_70,code=sm_70;-gencode;arch=compute_75,code=sm_75;-gencode;arch=compute_80,code=sm_80;-gencode;arch=compute_86,code=sm_86;-gencode;arch=compute_90,code=sm_90\n - CuDNN 90.7.1 (built against CUDA 12.8)\n - Built with CuDNN 90.5.1\n - Magma 2.6.1\n - Build settings: BLAS_INFO=mkl, BUILD_TYPE=Release, COMMIT_SHA=e2d141dbde55c2a4370fac5165b0561b6af4798b, CUDA_VERSION=12.6, CUDNN_VERSION=9.5.1, CXX_COMPILER=/opt/rh/gcc-toolset-11/root/usr/bin/c++, CXX_FLAGS= -D_GLIBCXX_USE_CXX11_ABI=1 -fvisibility-inlines-hidden -DUSE_PTHREADPOOL -DNDEBUG -DUSE_KINETO -DLIBKINETO_NOROCTRACER -DLIBKINETO_NOXPUPTI=ON -DUSE_FBGEMM -DUSE_PYTORCH_QNNPACK -DUSE_XNNPACK -DSYMBOLICATE_MOBILE_DEBUG_HANDLE -O2 -fPIC -Wall -Wextra -Werror=return-type -Werror=non-virtual-dtor -Werror=range-loop-construct -Werror=bool-operation -Wnarrowing -Wno-missing-field-initializers -Wno-unknown-pragmas -Wno-unused-parameter -Wno-strict-overflow -Wno-strict-aliasing -Wno-stringop-overflow -Wsuggest-override -Wno-psabi -Wno-error=old-style-cast -fdiagnostics-color=always -faligned-new -Wno-maybe-uninitialized -fno-math-errno -fno-trapping-math -Werror=format -Wno-stringop-overflow, LAPACK_INFO=mkl, PERF_WITH_AVX=1, PERF_WITH_AVX2=1, TORCH_VERSION=2.7.1, USE_CUDA=ON, USE_CUDNN=ON, USE_CUSPARSELT=1, USE_GFLAGS=OFF, USE_GLOG=OFF, USE_GLOO=ON, USE_MKL=ON, USE_MKLDNN=ON, USE_MPI=OFF, USE_NCCL=1, USE_NNPACK=ON, USE_OPENMP=ON, USE_ROCM=OFF, USE_ROCM_KERNEL_ASSERT=OFF, \n" } }