{ "run_info": { "created_at": "2026-01-10T08:32:50+00:00", "total_time": 1712.8715220090235, "experiment_name": "randlora/llama-3.2-3B-default", "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": "RANDLORA", "auto_mapping": null, "peft_version": null, "base_model_name_or_path": "meta-llama/Llama-3.2-3B", "revision": null, "inference_mode": false, "r": 32, "target_modules": [ "v_proj", "q_proj" ], "projection_prng_key": 0, "save_projection": true, "sparse": false, "very_sparse": false, "randlora_dropout": 0.0, "fan_in_fan_out": false, "randlora_alpha": 640, "bias": "none", "modules_to_save": null, "init_weights": true, "layers_to_transform": null, "layers_pattern": null }, "error_msg": "" }, "train_info": { "accelerator_memory_reserved_avg": 15193692176, "accelerator_memory_max": 22791847936, "accelerator_memory_reserved_99th": 20508049408, "train_time": 1465.4105980930617, "file_size": 2211281240, "num_trainable_params": 9289728, "num_total_params": 3222039552, "status": "success", "metrics": [ { "step": 250, "valid accuracy": 0.38, "train loss": 0.9159073839187623, "train samples": 1000, "train time": 51.456300139427185, "eval time": 8.755500879022293, "tokens / sec": 4114.539899416034, "mem allocated avg": 6994605942.784, "mem reserved avg": 15311281586.176, "elapsed time": 84.91182701400248 }, { "step": 500, "valid accuracy": 0.34, "train loss": 0.7010563344955444, "train samples": 2000, "train time": 52.27174317778554, "eval time": 14.115188423951622, "tokens / sec": 3979.1096939807776, "mem allocated avg": 6986201692.16, "mem reserved avg": 14981995167.744, "elapsed time": 154.25366032699822 }, { "step": 750, "valid accuracy": 0.36, "train loss": 0.6810104308128357, "train samples": 3000, "train time": 52.0417874493869, "eval time": 8.199473799031693, "tokens / sec": 4119.785474480774, "mem allocated avg": 6997675347.968, "mem reserved avg": 15224560156.672, "elapsed time": 217.4881883119815 }, { "step": 1000, "valid accuracy": 0.28, "train loss": 0.6662307423353195, "train samples": 4000, "train time": 51.765757714165375, "eval time": 7.892937373020686, "tokens / sec": 4024.5909496846825, "mem allocated avg": 6988936497.152, "mem reserved avg": 15255941939.2, "elapsed time": 280.21508718596306 }, { "step": 1250, "valid accuracy": 0.38, "train loss": 0.6655746358633041, "train samples": 5000, "train time": 52.88386500120396, "eval time": 14.099892683967482, "tokens / sec": 3943.3199520355106, "mem allocated avg": 6988741001.216, "mem reserved avg": 15194075955.2, "elapsed time": 350.0570412139641 }, { "step": 1500, "valid accuracy": 0.38, "train loss": 0.6581832259893418, "train samples": 6000, "train time": 52.76923948794138, "eval time": 14.133205129997805, "tokens / sec": 3966.9133387422708, "mem allocated avg": 6991358814.208, "mem reserved avg": 15142125305.856, "elapsed time": 419.9533305739751 }, { "step": 1750, "valid accuracy": 0.36, "train loss": 0.6510355799198151, "train samples": 7000, "train time": 53.25456840818515, "eval time": 13.085501310008112, "tokens / sec": 3931.212030399676, "mem allocated avg": 6991980468.224, "mem reserved avg": 15527758004.224, "elapsed time": 489.19832555996254 }, { "step": 2000, "valid accuracy": 0.46, "train loss": 0.6502977329492569, "train samples": 8000, "train time": 53.048176583775785, "eval time": 14.269489981001243, "tokens / sec": 3915.2335362931512, "mem allocated avg": 6987123507.2, "mem reserved avg": 15160764792.832, "elapsed time": 559.521614014986 }, { "step": 2250, "valid accuracy": 0.46, "train loss": 0.6383942588567734, "train samples": 9000, "train time": 54.19255740486551, "eval time": 14.067514252034016, "tokens / sec": 3966.374910011196, "mem allocated avg": 6999139655.68, "mem reserved avg": 15462352027.648, "elapsed time": 630.752589983982 }, { "step": 2500, "valid accuracy": 0.42, "train loss": 0.6331259000301361, "train samples": 10000, "train time": 51.30145365791395, "eval time": 14.185831039969344, "tokens / sec": 4014.837500968684, "mem allocated avg": 6985442590.72, "mem reserved avg": 15002463371.264, "elapsed time": 699.3249540069955 }, { "step": 2750, "valid accuracy": 0.5, "train loss": 0.6209538959264755, "train samples": 11000, "train time": 52.28545083821518, "eval time": 8.485180826974101, "tokens / sec": 4052.3892708817802, "mem allocated avg": 6994917660.672, "mem reserved avg": 15347461652.48, "elapsed time": 763.0090306919883 }, { "step": 3000, "valid accuracy": 0.42, "train loss": 0.608346626996994, "train samples": 12000, "train time": 50.35002116265241, "eval time": 7.221828919020481, "tokens / sec": 4145.599051998574, "mem allocated avg": 6990200410.112, "mem reserved avg": 15265429454.848, "elapsed time": 823.5788191119791 }, { "step": 3250, "valid accuracy": 0.5, "train loss": 0.6135392972230911, "train samples": 13000, "train time": 50.94754268962424, "eval time": 14.057202987954952, "tokens / sec": 4139.571584145337, "mem allocated avg": 6993134067.712, "mem reserved avg": 15059187138.56, "elapsed time": 891.3881165689672 }, { "step": 3500, "valid accuracy": 0.52, "train loss": 0.5913569860458374, "train samples": 14000, "train time": 51.22462305892259, "eval time": 10.398285192961339, "tokens / sec": 4094.710462949216, "mem allocated avg": 6989725995.008, "mem reserved avg": 15092934508.544, "elapsed time": 955.9796246159822 }, { "step": 3750, "valid accuracy": 0.46, "train loss": 0.586094471335411, "train samples": 15000, "train time": 54.32701862248359, "eval time": 7.748301387007814, "tokens / sec": 3988.8623652599267, "mem allocated avg": 7002481041.408, "mem reserved avg": 15310149124.096, "elapsed time": 1020.95719456597 }, { "step": 4000, "valid accuracy": 0.48, "train loss": 0.5912544637918472, "train samples": 16000, "train time": 51.06255231570685, "eval time": 14.150051894015633, "tokens / sec": 4002.4047120953414, "mem allocated avg": 6982720993.28, "mem reserved avg": 15162954219.52, "elapsed time": 1089.1380827649846 }, { "step": 4250, "valid accuracy": 0.5, "train loss": 0.5755240997076034, "train samples": 17000, "train time": 52.308833500603214, "eval time": 14.18892294500256, "tokens / sec": 4041.1721281752975, "mem allocated avg": 6993589919.744, "mem reserved avg": 15129659834.368, "elapsed time": 1158.4711514469818 }, { "step": 4500, "valid accuracy": 0.56, "train loss": 0.5807081238031387, "train samples": 18000, "train time": 50.61249336565379, "eval time": 9.862378416000865, "tokens / sec": 4106.0612939695175, "mem allocated avg": 6988681547.776, "mem reserved avg": 15200375799.808, "elapsed time": 1221.9246637789765 }, { "step": 4750, "valid accuracy": 0.52, "train loss": 0.5690834984779358, "train samples": 19000, "train time": 51.20076460047858, "eval time": 8.214846904971637, "tokens / sec": 4100.309861349955, "mem allocated avg": 6991610628.096, "mem reserved avg": 15140388864.0, "elapsed time": 1284.1992750130012 }, { "step": 5000, "valid accuracy": 0.5, "train loss": 0.5778189089298248, "train samples": 20000, "train time": 51.87798086879775, "eval time": 8.667530598992016, "tokens / sec": 4014.8054436958814, "mem allocated avg": 6988335155.2, "mem reserved avg": 14901984624.64, "elapsed time": 1347.629596157989 }, { "step": 5000, "test accuracy": 0.49962092494313876, "train loss": 0.5778189089298248, "train samples": 20000, "train total tokens": 4198051, "forgetting": 0.46999263763427734 } ] }, "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": "cc7b047b6e5bb11b4f1af84efc572db110a51b3c", "created_at": "2022-04-12T10:22:10+00:00" } }, "package_info": { "transformers-version": "4.57.1", "transformers-commit-hash": null, "peft-version": "0.18.1.dev0", "peft-commit-hash": "8be1a16f5e06ca5e197d2af74bdfc5b3c8072d26", "datasets-version": "4.2.0", "datasets-commit-hash": null, "bitsandbytes-version": "0.46.0", "bitsandbytes-commit-hash": null, "torch-version": "2.9.0+cu128", "torch-commit-hash": null }, "system_info": { "system": "Linux", "release": "6.14.0-1016-aws", "version": "#16~24.04.1-Ubuntu SMP Tue Oct 14 02:15:09 UTC 2025", "machine": "x86_64", "processor": "x86_64", "accelerator": "NVIDIA L40S" }, "pytorch_info": "PyTorch built with:\n - GCC 13.3\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 12.8\n - NVCC architecture flags: -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;-gencode;arch=compute_100,code=sm_100;-gencode;arch=compute_120,code=sm_120\n - CuDNN 90.7.1\n - Built with CuDNN 90.8\n - Magma 2.6.1\n - Build settings: BLAS_INFO=mkl, BUILD_TYPE=Release, COMMIT_SHA=0fabc3ba44823f257e70ce397d989c8de5e362c1, CUDA_VERSION=12.8, CUDNN_VERSION=9.8.0, CXX_COMPILER=/opt/rh/gcc-toolset-13/root/usr/bin/c++, CXX_FLAGS= -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 -DC10_NODEPRECATED -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 -faligned-new -Wno-maybe-uninitialized -fno-math-errno -fno-trapping-math -Werror=format -Wno-dangling-reference -Wno-error=dangling-reference -Wno-stringop-overflow, LAPACK_INFO=mkl, PERF_WITH_AVX=1, PERF_WITH_AVX2=1, TORCH_VERSION=2.9.0, 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, USE_XCCL=OFF, USE_XPU=OFF, \n" } }