{ "run_info": { "created_at": "2025-07-31T11:51:03+00:00", "total_time": 1813.5205606600002, "experiment_name": "trainable_tokens/llama-3.2-3B-sos+eos", "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": 0.2 }, "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": "CAUSAL_LM", "peft_type": "TRAINABLE_TOKENS", "auto_mapping": null, "base_model_name_or_path": "meta-llama/Llama-3.2-3B", "revision": null, "inference_mode": false, "token_indices": [ 128000, 128001 ], "target_modules": "model.embed_tokens", "init_weights": true }, "error_msg": "" }, "train_info": { "accelerator_memory_reserved_avg": 12730137942, "accelerator_memory_max": 20956839936, "accelerator_memory_reserved_99th": 16957675929, "train_time": 1571.9034050299992, "file_size": 49424, "num_trainable_params": 6144, "num_total_params": 3212755968, "status": "success", "metrics": [ { "step": 250, "valid accuracy": 0.2, "train loss": 0.9600473783016205, "train samples": 1000, "train time": 28.046887916000458, "eval time": 11.940005464000023, "tokens / sec": 7548.751955443032, "mem allocated avg": 6772620728.32, "mem reserved avg": 12816987717.632, "elapsed time": 79.72835956 }, { "step": 500, "valid accuracy": 0.1, "train loss": 0.9827968046665192, "train samples": 2000, "train time": 27.568676805000052, "eval time": 11.946740159, "tokens / sec": 7544.613093736749, "mem allocated avg": 6765130303.488, "mem reserved avg": 12640885669.888, "elapsed time": 153.151956972 }, { "step": 750, "valid accuracy": 0.26, "train loss": 1.078141197681427, "train samples": 3000, "train time": 28.346493393999594, "eval time": 11.913883551000026, "tokens / sec": 7563.581040517151, "mem allocated avg": 6775356669.952, "mem reserved avg": 12829595795.456, "elapsed time": 227.42617082600003 }, { "step": 1000, "valid accuracy": 0.18, "train loss": 0.9337297527790069, "train samples": 4000, "train time": 27.80469767599925, "eval time": 11.935028867000028, "tokens / sec": 7492.834571613906, "mem allocated avg": 6767140114.432, "mem reserved avg": 12714495705.088, "elapsed time": 301.119469732 }, { "step": 1250, "valid accuracy": 0.06, "train loss": 1.1530333435535431, "train samples": 5000, "train time": 27.832769999000448, "eval time": 11.899406999999997, "tokens / sec": 7492.534879118722, "mem allocated avg": 6766889218.048, "mem reserved avg": 12706635579.392, "elapsed time": 374.949399381 }, { "step": 1500, "valid accuracy": 0.16, "train loss": 0.9257006254196167, "train samples": 6000, "train time": 27.848633210000685, "eval time": 11.963534283000058, "tokens / sec": 7516.7423270463205, "mem allocated avg": 6768338049.024, "mem reserved avg": 12713531015.168, "elapsed time": 448.64371043399996 }, { "step": 1750, "valid accuracy": 0.04, "train loss": 0.9480950233936309, "train samples": 7000, "train time": 28.10500999800047, "eval time": 11.907140037999966, "tokens / sec": 7449.027771735166, "mem allocated avg": 6769886017.536, "mem reserved avg": 12724176158.72, "elapsed time": 523.0294692360001 }, { "step": 2000, "valid accuracy": 0.22, "train loss": 1.099897492647171, "train samples": 8000, "train time": 27.50086192600054, "eval time": 11.908490246000042, "tokens / sec": 7552.345106814087, "mem allocated avg": 6767001008.128, "mem reserved avg": 12671663472.64, "elapsed time": 596.5166793630001 }, { "step": 2250, "valid accuracy": 0.2, "train loss": 1.0596771531105043, "train samples": 9000, "train time": 28.679387931000633, "eval time": 11.905819370000017, "tokens / sec": 7494.860089662325, "mem allocated avg": 6777668958.208, "mem reserved avg": 12765641048.064, "elapsed time": 671.4013165209999 }, { "step": 2500, "valid accuracy": 0.28, "train loss": 0.8838931715488434, "train samples": 10000, "train time": 27.525735084001553, "eval time": 11.918955123999922, "tokens / sec": 7482.70661515273, "mem allocated avg": 6762954442.752, "mem reserved avg": 12636850749.44, "elapsed time": 744.8136404429999 }, { "step": 2750, "valid accuracy": 0.28, "train loss": 0.8453443908691406, "train samples": 11000, "train time": 28.56140573299922, "eval time": 11.910845225000003, "tokens / sec": 7418.437382975074, "mem allocated avg": 6773394008.064, "mem reserved avg": 12781445185.536, "elapsed time": 819.4217107059999 }, { "step": 3000, "valid accuracy": 0.28, "train loss": 0.8956673395633697, "train samples": 12000, "train time": 27.520784680998986, "eval time": 11.922119511000119, "tokens / sec": 7584.485777548084, "mem allocated avg": 6768260833.28, "mem reserved avg": 12729855246.336, "elapsed time": 892.9293128969998 }, { "step": 3250, "valid accuracy": 0.28, "train loss": 0.8384639675617218, "train samples": 13000, "train time": 27.425537425002176, "eval time": 11.90852308500007, "tokens / sec": 7689.9495799026545, "mem allocated avg": 6769824622.592, "mem reserved avg": 12732145336.32, "elapsed time": 966.4361498369999 }, { "step": 3500, "valid accuracy": 0.28, "train loss": 0.8318528242111206, "train samples": 14000, "train time": 27.599476771002628, "eval time": 11.926854094000191, "tokens / sec": 7599.781754572018, "mem allocated avg": 6769281437.696, "mem reserved avg": 12748083691.52, "elapsed time": 1040.349463558 }, { "step": 3750, "valid accuracy": 0.3, "train loss": 0.825278183221817, "train samples": 15000, "train time": 28.668002616999956, "eval time": 11.920088525999972, "tokens / sec": 7559.054702733158, "mem allocated avg": 6780301985.792, "mem reserved avg": 12867151593.472, "elapsed time": 1115.464965257 }, { "step": 4000, "valid accuracy": 0.34, "train loss": 0.8279166626930237, "train samples": 16000, "train time": 27.677960625997912, "eval time": 11.956045786999994, "tokens / sec": 7383.961656771504, "mem allocated avg": 6761672556.544, "mem reserved avg": 12607280906.24, "elapsed time": 1189.311786065 }, { "step": 4250, "valid accuracy": 0.26, "train loss": 0.7829471210241318, "train samples": 17000, "train time": 27.69402594300027, "eval time": 11.900985279999986, "tokens / sec": 7633.018053607661, "mem allocated avg": 6771771314.176, "mem reserved avg": 12787006832.64, "elapsed time": 1263.137991501 }, { "step": 4500, "valid accuracy": 0.36, "train loss": 0.788821615934372, "train samples": 18000, "train time": 27.622178668999823, "eval time": 11.941970926000067, "tokens / sec": 7523.591911062131, "mem allocated avg": 6766746431.488, "mem reserved avg": 12679666204.672, "elapsed time": 1336.699760115 }, { "step": 4750, "valid accuracy": 0.3, "train loss": 0.7817396788597107, "train samples": 19000, "train time": 28.053782109000622, "eval time": 11.908781481000005, "tokens / sec": 7483.447300770342, "mem allocated avg": 6769035872.256, "mem reserved avg": 12704303546.368, "elapsed time": 1411.011856929 }, { "step": 5000, "valid accuracy": 0.32, "train loss": 0.786205664396286, "train samples": 20000, "train time": 27.865934093999613, "eval time": 11.930896392000022, "tokens / sec": 7474.359169063313, "mem allocated avg": 6766191118.336, "mem reserved avg": 12745357393.92, "elapsed time": 1484.8831304389998 }, { "step": 5000, "test accuracy": 0.2880970432145565, "train loss": 0.786205664396286, "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" } }