{ "run_info": { "created_at": "2025-06-19T16:12:05+00:00", "total_time": 2590.9341236870005, "experiment_name": "loha/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": "LOHA", "auto_mapping": null, "base_model_name_or_path": "meta-llama/Llama-3.2-3B", "revision": null, "inference_mode": false, "rank_pattern": {}, "alpha_pattern": {}, "r": 32, "alpha": 64, "rank_dropout": 0.0, "module_dropout": 0.0, "use_effective_conv2d": false, "target_modules": [ "q_proj", "v_proj" ], "exclude_modules": null, "init_weights": true, "layers_to_transform": null, "layers_pattern": null, "modules_to_save": null }, "error_msg": "" }, "train_info": { "accelerator_memory_reserved_avg": 13446820344, "accelerator_memory_max": 23886561280, "accelerator_memory_reserved_99th": 19247870771, "train_time": 2340.7451966560056, "file_size": 73429560, "num_trainable_params": 18350080, "num_total_params": 3231099904, "status": "success", "metrics": [ { "step": 250, "valid accuracy": 0.0, "train loss": 1.2914833688735963, "train samples": 1000, "train time": 47.4107696449737, "eval time": 14.298813604000316, "tokens / sec": 4465.630943885038, "mem allocated avg": 7073903032.32, "mem reserved avg": 13501707845.632, "elapsed time": 120.40146815400112 }, { "step": 500, "valid accuracy": 0.36, "train loss": 0.9051123185157776, "train samples": 2000, "train time": 47.1910586300055, "eval time": 14.155041256999539, "tokens / sec": 4407.508668766131, "mem allocated avg": 7065529796.608, "mem reserved avg": 13391154380.8, "elapsed time": 234.63223427900084 }, { "step": 750, "valid accuracy": 0.34, "train loss": 0.7515897085666656, "train samples": 3000, "train time": 48.203471163995346, "eval time": 14.26281827999992, "tokens / sec": 4447.8332124791605, "mem allocated avg": 7076336949.248, "mem reserved avg": 13550454046.72, "elapsed time": 350.3563782780002 }, { "step": 1000, "valid accuracy": 0.4, "train loss": 0.7082941273450851, "train samples": 4000, "train time": 47.1758063940124, "eval time": 14.148222272999192, "tokens / sec": 4416.161925457669, "mem allocated avg": 7067704416.256, "mem reserved avg": 13415607173.12, "elapsed time": 464.4216197680016 }, { "step": 1250, "valid accuracy": 0.32, "train loss": 0.6994056793451309, "train samples": 5000, "train time": 47.32811543400385, "eval time": 14.265782994998517, "tokens / sec": 4406.218123998481, "mem allocated avg": 7067674988.544, "mem reserved avg": 13411052158.976, "elapsed time": 578.864341718001 }, { "step": 1500, "valid accuracy": 0.38, "train loss": 0.6889224811792374, "train samples": 6000, "train time": 47.48961014100678, "eval time": 9.485757000999001, "tokens / sec": 4407.932585221307, "mem allocated avg": 7068496666.624, "mem reserved avg": 13434196328.448, "elapsed time": 688.8339757740014 }, { "step": 1750, "valid accuracy": 0.36, "train loss": 0.6795901688337326, "train samples": 7000, "train time": 47.5112849769921, "eval time": 8.524607335999463, "tokens / sec": 4406.426812101222, "mem allocated avg": 7070726457.344, "mem reserved avg": 13451493638.144, "elapsed time": 797.8088079910012 }, { "step": 2000, "valid accuracy": 0.42, "train loss": 0.680127969622612, "train samples": 8000, "train time": 47.15311444799954, "eval time": 14.09636382700046, "tokens / sec": 4404.714353047605, "mem allocated avg": 7067623004.16, "mem reserved avg": 13389174669.312, "elapsed time": 911.8939753530012 }, { "step": 2250, "valid accuracy": 0.42, "train loss": 0.6731046036481857, "train samples": 9000, "train time": 48.44002798400652, "eval time": 14.30888277199847, "tokens / sec": 4437.404538060332, "mem allocated avg": 7078766321.664, "mem reserved avg": 13582146207.744, "elapsed time": 1028.5093922580018 }, { "step": 2500, "valid accuracy": 0.42, "train loss": 0.6711453741788864, "train samples": 10000, "train time": 46.86391301901131, "eval time": 8.751619284999833, "tokens / sec": 4395.002182520381, "mem allocated avg": 7063469082.624, "mem reserved avg": 13336376770.56, "elapsed time": 1137.0856770440005 }, { "step": 2750, "valid accuracy": 0.44, "train loss": 0.6645345565080643, "train samples": 11000, "train time": 47.92562343400823, "eval time": 7.835686906000774, "tokens / sec": 4421.037950434847, "mem allocated avg": 7074535438.336, "mem reserved avg": 13512352989.184, "elapsed time": 1246.1237790820014 }, { "step": 3000, "valid accuracy": 0.4, "train loss": 0.6566170369386672, "train samples": 12000, "train time": 47.50991778100797, "eval time": 14.152554526999666, "tokens / sec": 4393.419516365485, "mem allocated avg": 7068629661.696, "mem reserved avg": 13428215250.944, "elapsed time": 1360.8028930970013 }, { "step": 3250, "valid accuracy": 0.42, "train loss": 0.6667062133550644, "train samples": 13000, "train time": 47.62723316902702, "eval time": 14.332656014001259, "tokens / sec": 4428.1598146069355, "mem allocated avg": 7071043653.632, "mem reserved avg": 13457114005.504, "elapsed time": 1476.0946507730005 }, { "step": 3500, "valid accuracy": 0.42, "train loss": 0.6537795497179031, "train samples": 14000, "train time": 47.07006615899445, "eval time": 14.135684340000807, "tokens / sec": 4456.12290604184, "mem allocated avg": 7069669969.92, "mem reserved avg": 13439749586.944, "elapsed time": 1590.4238928290015 }, { "step": 3750, "valid accuracy": 0.46, "train loss": 0.6509792991876602, "train samples": 15000, "train time": 48.58318820000204, "eval time": 14.298812560000442, "tokens / sec": 4460.452432802484, "mem allocated avg": 7081669246.976, "mem reserved avg": 13624240242.688, "elapsed time": 1707.3096692510007 }, { "step": 4000, "valid accuracy": 0.46, "train loss": 0.6675102390050888, "train samples": 16000, "train time": 46.83876558602242, "eval time": 14.188353157000165, "tokens / sec": 4363.330191199334, "mem allocated avg": 7062227976.192, "mem reserved avg": 13316957143.04, "elapsed time": 1821.413719397 }, { "step": 4250, "valid accuracy": 0.46, "train loss": 0.6494157313108444, "train samples": 17000, "train time": 46.9989987980116, "eval time": 8.258924301999286, "tokens / sec": 4497.7341093688, "mem allocated avg": 7072862310.4, "mem reserved avg": 13470619664.384, "elapsed time": 1930.0706906220003 }, { "step": 4500, "valid accuracy": 0.44, "train loss": 0.6580193819999695, "train samples": 18000, "train time": 47.171681194990015, "eval time": 9.717189478000364, "tokens / sec": 4405.566957449713, "mem allocated avg": 7068038127.616, "mem reserved avg": 13393654185.984, "elapsed time": 2040.1967968460012 }, { "step": 4750, "valid accuracy": 0.48, "train loss": 0.6511869616508484, "train samples": 19000, "train time": 47.517527918005726, "eval time": 14.28858694399969, "tokens / sec": 4418.138089217562, "mem allocated avg": 7069871403.008, "mem reserved avg": 13443927113.728, "elapsed time": 2155.4679406510004 }, { "step": 5000, "valid accuracy": 0.46, "train loss": 0.6569721374511719, "train samples": 20000, "train time": 46.99870921700858, "eval time": 9.378413720998651, "tokens / sec": 4431.6110691104805, "mem allocated avg": 7066192863.232, "mem reserved avg": 13386213490.688, "elapsed time": 2265.1425104650007 }, { "step": 5000, "test accuracy": 0.4184988627748294, "train loss": 0.6569721374511719, "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": "0.15.2.dev0", "peft-commit-hash": "5fe7f8f8abe914d313fc3751f2ea92de7718fbaf", "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": "6.8.0-1029-aws", "version": "#31-Ubuntu SMP Wed Apr 23 18:42:41 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" } }