{ "run_info": { "created_at": "2026-01-10T08:13:23+00:00", "total_time": 1162.917062154971, "experiment_name": "ptuning/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": "CAUSAL_LM", "peft_type": "P_TUNING", "auto_mapping": null, "peft_version": "0.18.1.dev0@UNKNOWN", "base_model_name_or_path": "meta-llama/Llama-3.2-3B", "revision": null, "inference_mode": false, "num_virtual_tokens": 20, "token_dim": 3072, "num_transformer_submodules": 1, "num_attention_heads": 24, "num_layers": 28, "modules_to_save": null, "encoder_reparameterization_type": "MLP", "encoder_hidden_size": 3072, "encoder_num_layers": 2, "encoder_dropout": 0.0 }, "error_msg": "" }, "train_info": { "accelerator_memory_reserved_avg": 14104160475, "accelerator_memory_max": 20950548480, "accelerator_memory_reserved_99th": 19218300928, "train_time": 959.7272375969915, "file_size": 245880, "num_trainable_params": 28382208, "num_total_params": 3241132032, "status": "success", "metrics": [ { "step": 250, "valid accuracy": 0.12, "train loss": 0.9454967255592346, "train samples": 1000, "train time": 29.694366472947877, "eval time": 11.680081729020458, "tokens / sec": 7129.938272731966, "mem allocated avg": 7266438576.128, "mem reserved avg": 14210964652.032, "elapsed time": 62.83816469996236 }, { "step": 500, "valid accuracy": 0.36, "train loss": 0.7915761218070984, "train samples": 2000, "train time": 30.064934794034343, "eval time": 11.6559070070507, "tokens / sec": 6918.192286958545, "mem allocated avg": 7258846287.872, "mem reserved avg": 13882106052.608, "elapsed time": 107.20484153198777 }, { "step": 750, "valid accuracy": 0.32, "train loss": 0.7556082547903061, "train samples": 3000, "train time": 30.023478755319957, "eval time": 11.630144803028088, "tokens / sec": 7141.111186591247, "mem allocated avg": 7269281730.56, "mem reserved avg": 14078433034.24, "elapsed time": 151.32677704497473 }, { "step": 1000, "valid accuracy": 0.28, "train loss": 0.7297289679050446, "train samples": 4000, "train time": 29.457952680124436, "eval time": 11.804060328984633, "tokens / sec": 7072.317695063931, "mem allocated avg": 7260512626.688, "mem reserved avg": 14043024719.872, "elapsed time": 195.22547230595956 }, { "step": 1250, "valid accuracy": 0.38, "train loss": 0.7234208507537842, "train samples": 5000, "train time": 29.54452143079834, "eval time": 11.626074030995369, "tokens / sec": 7058.43215258894, "mem allocated avg": 7259926929.408, "mem reserved avg": 14128538189.824, "elapsed time": 238.97390070999973 }, { "step": 1500, "valid accuracy": 0.36, "train loss": 0.7123123975992203, "train samples": 6000, "train time": 30.126697142142802, "eval time": 8.240376622998156, "tokens / sec": 6948.355440768741, "mem allocated avg": 7261799639.04, "mem reserved avg": 14042739507.2, "elapsed time": 279.96362110297196 }, { "step": 1750, "valid accuracy": 0.42, "train loss": 0.7047000522613526, "train samples": 7000, "train time": 29.664779813203495, "eval time": 11.633714700001292, "tokens / sec": 7057.358973108514, "mem allocated avg": 7263532308.48, "mem reserved avg": 14314966614.016, "elapsed time": 323.87410058197565 }, { "step": 2000, "valid accuracy": 0.44, "train loss": 0.7051602005958557, "train samples": 8000, "train time": 29.419329382071737, "eval time": 11.648401496990118, "tokens / sec": 7059.848214166664, "mem allocated avg": 7259382927.36, "mem reserved avg": 14145265074.176, "elapsed time": 367.623096796975 }, { "step": 2250, "valid accuracy": 0.44, "train loss": 0.6980359296798706, "train samples": 9000, "train time": 30.83813531260239, "eval time": 11.641796185984276, "tokens / sec": 6970.20094830957, "mem allocated avg": 7271977609.216, "mem reserved avg": 14313431498.752, "elapsed time": 412.71304436796345 }, { "step": 2500, "valid accuracy": 0.5, "train loss": 0.6962304570674896, "train samples": 10000, "train time": 29.281850074534304, "eval time": 7.721943242009729, "tokens / sec": 7033.947632261268, "mem allocated avg": 7256155267.072, "mem reserved avg": 13909410971.648, "elapsed time": 452.32350523298373 }, { "step": 2750, "valid accuracy": 0.4, "train loss": 0.6884069440364837, "train samples": 11000, "train time": 29.80504708213266, "eval time": 11.643363915965892, "tokens / sec": 7108.8966716317345, "mem allocated avg": 7266777749.504, "mem reserved avg": 14297912573.952, "elapsed time": 496.34979761595605 }, { "step": 3000, "valid accuracy": 0.34, "train loss": 0.6823657184839249, "train samples": 12000, "train time": 29.4866174882045, "eval time": 11.663472635962535, "tokens / sec": 7078.838394519087, "mem allocated avg": 7261770512.384, "mem reserved avg": 14176470695.936, "elapsed time": 540.1913515449851 }, { "step": 3250, "valid accuracy": 0.32, "train loss": 0.6892943847179412, "train samples": 13000, "train time": 30.05981874052668, "eval time": 9.359989358054008, "tokens / sec": 7016.043636872069, "mem allocated avg": 7263956414.464, "mem reserved avg": 14069599830.016, "elapsed time": 582.1699139429838 }, { "step": 3500, "valid accuracy": 0.34, "train loss": 0.6757683880329132, "train samples": 14000, "train time": 29.95969556894852, "eval time": 11.67419690900715, "tokens / sec": 7001.072474761515, "mem allocated avg": 7262049505.28, "mem reserved avg": 14046656987.136, "elapsed time": 626.462889155955 }, { "step": 3750, "valid accuracy": 0.32, "train loss": 0.6726086636781693, "train samples": 15000, "train time": 30.04007897916017, "eval time": 11.621128266968299, "tokens / sec": 7213.795947418589, "mem allocated avg": 7273184010.24, "mem reserved avg": 14348948865.024, "elapsed time": 670.6827715289546 }, { "step": 4000, "valid accuracy": 0.36, "train loss": 0.687798623919487, "train samples": 16000, "train time": 28.8626030132873, "eval time": 6.513009499001782, "tokens / sec": 7080.892873934969, "mem allocated avg": 7255382544.384, "mem reserved avg": 14072502288.384, "elapsed time": 708.7366754019749 }, { "step": 4250, "valid accuracy": 0.34, "train loss": 0.6694162834882736, "train samples": 17000, "train time": 29.575119761226233, "eval time": 6.9559804140008055, "tokens / sec": 7147.528115072473, "mem allocated avg": 7265205653.504, "mem reserved avg": 14095831007.232, "elapsed time": 747.7560498069506 }, { "step": 4500, "valid accuracy": 0.42, "train loss": 0.6779570702314377, "train samples": 18000, "train time": 29.56798030581558, "eval time": 6.977328688022681, "tokens / sec": 7028.4814130211425, "mem allocated avg": 7259650541.568, "mem reserved avg": 14029410009.088, "elapsed time": 786.97692555096 }, { "step": 4750, "valid accuracy": 0.32, "train loss": 0.6683919057846069, "train samples": 19000, "train time": 30.30186787992716, "eval time": 6.962075466988608, "tokens / sec": 6928.252767515686, "mem allocated avg": 7262556618.752, "mem reserved avg": 14063425814.528, "elapsed time": 826.7863181569846 }, { "step": 5000, "valid accuracy": 0.38, "train loss": 0.6739522596597671, "train samples": 20000, "train time": 29.33842807688052, "eval time": 7.4907851399620995, "tokens / sec": 7099.221521146537, "mem allocated avg": 7259827462.144, "mem reserved avg": 13813571125.248, "elapsed time": 866.1968886919785 }, { "step": 5000, "test accuracy": 0.3821076573161486, "train loss": 0.6739522596597671, "train samples": 20000, "train total tokens": 4198051, "forgetting": 3.1548709869384766 } ] }, "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" } }