{ "run_info": { "created_at": "2025-06-19T21:27:27+00:00", "total_time": 1921.5641919770023, "experiment_name": "ia3/llama-3.2-3B-lr_0.001", "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": 0.001 }, "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": "IA3", "auto_mapping": null, "base_model_name_or_path": "meta-llama/Llama-3.2-3B", "revision": null, "inference_mode": false, "target_modules": [ "k_proj", "down_proj", "v_proj" ], "exclude_modules": null, "feedforward_modules": [ "down_proj" ], "fan_in_fan_out": false, "modules_to_save": null, "init_ia3_weights": true }, "error_msg": "" }, "train_info": { "accelerator_memory_reserved_avg": 12023331867, "accelerator_memory_max": 23135780864, "accelerator_memory_reserved_99th": 18398356439, "train_time": 1746.0246657649877, "file_size": 1157064, "num_trainable_params": 286720, "num_total_params": 3213036544, "status": "success", "metrics": [ { "step": 250, "valid accuracy": 0.18, "train loss": 1.1670710837841034, "train samples": 1000, "train time": 30.829080988976784, "eval time": 10.962777282999014, "tokens / sec": 6867.509286952213, "mem allocated avg": 6781095491.584, "mem reserved avg": 12075594153.984, "elapsed time": 91.04478788100096 }, { "step": 500, "valid accuracy": 0.34, "train loss": 0.8285422480106354, "train samples": 2000, "train time": 30.237734625952726, "eval time": 10.93798775599862, "tokens / sec": 6878.656836331916, "mem allocated avg": 6773575256.064, "mem reserved avg": 11961039323.136, "elapsed time": 175.57074494799963 }, { "step": 750, "valid accuracy": 0.34, "train loss": 0.7387537934780121, "train samples": 3000, "train time": 30.784141963005823, "eval time": 10.918857135002327, "tokens / sec": 6964.657330961239, "mem allocated avg": 6784163356.672, "mem reserved avg": 12124793339.904, "elapsed time": 261.120397177001 }, { "step": 1000, "valid accuracy": 0.36, "train loss": 0.7030822492837906, "train samples": 4000, "train time": 30.625773959025537, "eval time": 6.545184372997028, "tokens / sec": 6802.636246147914, "mem allocated avg": 6775321157.632, "mem reserved avg": 11986549080.064, "elapsed time": 341.78445690100125 }, { "step": 1250, "valid accuracy": 0.34, "train loss": 0.6953592277765274, "train samples": 5000, "train time": 30.090904191973095, "eval time": 7.180137749001005, "tokens / sec": 6930.266989305977, "mem allocated avg": 6774968741.888, "mem reserved avg": 11983218802.688, "elapsed time": 422.45400445199994 }, { "step": 1500, "valid accuracy": 0.34, "train loss": 0.6861299908161164, "train samples": 6000, "train time": 30.086008766014857, "eval time": 10.923475695002708, "tokens / sec": 6957.75241003254, "mem allocated avg": 6776914077.696, "mem reserved avg": 12007201832.96, "elapsed time": 506.8615667560007 }, { "step": 1750, "valid accuracy": 0.34, "train loss": 0.6775313948392868, "train samples": 7000, "train time": 30.329398032976314, "eval time": 7.039293795001868, "tokens / sec": 6902.708710946855, "mem allocated avg": 6778176180.224, "mem reserved avg": 12032417988.608, "elapsed time": 587.730657346001 }, { "step": 2000, "valid accuracy": 0.36, "train loss": 0.6783386437892914, "train samples": 8000, "train time": 30.340071335995162, "eval time": 8.14293124300093, "tokens / sec": 6845.600252547578, "mem allocated avg": 6775202904.064, "mem reserved avg": 11967733432.32, "elapsed time": 669.6239550099999 }, { "step": 2250, "valid accuracy": 0.5, "train loss": 0.6720720986127854, "train samples": 9000, "train time": 31.104124111985584, "eval time": 7.4280358140022145, "tokens / sec": 6910.594853149151, "mem allocated avg": 6785809762.304, "mem reserved avg": 12167885619.2, "elapsed time": 752.2532132060005 }, { "step": 2500, "valid accuracy": 0.46, "train loss": 0.6705386472940444, "train samples": 10000, "train time": 30.09476044199255, "eval time": 7.5499184540021815, "tokens / sec": 6843.948812850663, "mem allocated avg": 6770963554.304, "mem reserved avg": 11912058241.024, "elapsed time": 833.2611769709983 }, { "step": 2750, "valid accuracy": 0.48, "train loss": 0.6631126835346222, "train samples": 11000, "train time": 30.640666239018174, "eval time": 10.92325482400338, "tokens / sec": 6915.025879241109, "mem allocated avg": 6781913962.496, "mem reserved avg": 12090299383.808, "elapsed time": 918.4276470139994 }, { "step": 3000, "valid accuracy": 0.38, "train loss": 0.6557366658449173, "train samples": 12000, "train time": 30.612569437977072, "eval time": 10.933225860997482, "tokens / sec": 6818.473712992361, "mem allocated avg": 6776591689.728, "mem reserved avg": 12003032694.784, "elapsed time": 1003.438990486 }, { "step": 3250, "valid accuracy": 0.44, "train loss": 0.6655691808462143, "train samples": 13000, "train time": 30.508301533980557, "eval time": 7.2082155700009025, "tokens / sec": 6912.905320707402, "mem allocated avg": 6778600480.768, "mem reserved avg": 12040143896.576, "elapsed time": 1084.7670880670012 }, { "step": 3500, "valid accuracy": 0.46, "train loss": 0.6528272937536239, "train samples": 14000, "train time": 30.571383574966603, "eval time": 7.452295711998886, "tokens / sec": 6860.991406740058, "mem allocated avg": 6777338779.648, "mem reserved avg": 12021227585.536, "elapsed time": 1166.2843480039992 }, { "step": 3750, "valid accuracy": 0.48, "train loss": 0.6513591132164002, "train samples": 15000, "train time": 31.176262214954477, "eval time": 6.50122426100279, "tokens / sec": 6950.897400909496, "mem allocated avg": 6788519866.368, "mem reserved avg": 12209543446.528, "elapsed time": 1248.1537826940003 }, { "step": 4000, "valid accuracy": 0.42, "train loss": 0.6660103598833084, "train samples": 16000, "train time": 30.1621740100818, "eval time": 10.007692241000768, "tokens / sec": 6775.804686084222, "mem allocated avg": 6769538811.904, "mem reserved avg": 11886321991.68, "elapsed time": 1331.4140659110017 }, { "step": 4250, "valid accuracy": 0.4, "train loss": 0.648773505806923, "train samples": 17000, "train time": 30.627343150990782, "eval time": 9.851157391000015, "tokens / sec": 6901.969882201866, "mem allocated avg": 6780366684.16, "mem reserved avg": 12050411552.768, "elapsed time": 1415.4855422520013 }, { "step": 4500, "valid accuracy": 0.42, "train loss": 0.6574939725399017, "train samples": 18000, "train time": 30.04905582394713, "eval time": 6.792122120001295, "tokens / sec": 6915.957733167199, "mem allocated avg": 6775072815.104, "mem reserved avg": 11969042055.168, "elapsed time": 1495.5897211369993 }, { "step": 4750, "valid accuracy": 0.4, "train loss": 0.6505398267507553, "train samples": 19000, "train time": 30.326544256924535, "eval time": 7.6139581239986, "tokens / sec": 6922.615324100572, "mem allocated avg": 6777039572.992, "mem reserved avg": 12019256262.656, "elapsed time": 1577.114852814997 }, { "step": 5000, "valid accuracy": 0.42, "train loss": 0.6568749620914459, "train samples": 20000, "train time": 30.342653310064634, "eval time": 6.5661308569979155, "tokens / sec": 6864.264567492972, "mem allocated avg": 6774530805.76, "mem reserved avg": 11958866673.664, "elapsed time": 1657.5746541439985 }, { "step": 5000, "test accuracy": 0.41243366186504926, "train loss": 0.6568749620914459, "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" } }