{ "run_info": { "created_at": "2026-07-15T11:32:15+00:00", "total_time": 1289.498055625998, "experiment_name": "prefixtuning/llama-3.2-3B-lr_0.005_zero_init", "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, "use_gc": false, "query_template": "Question: {query} Think step by step.\nAnswer:", "seed": 0, "grad_norm_clip": 1.0, "optimizer_type": "AdamW", "optimizer_kwargs": { "lr": 0.005 }, "lr_scheduler": "cosine", "use_amp": false, "autocast_adapter_dtype": true, "generation_kwargs": { "max_length": 800, "max_new_tokens": 300 }, "attn_implementation": null, "init_kv_cache_prefix": null }, "peft_config": { "task_type": "CAUSAL_LM", "peft_type": "PREFIX_TUNING", "auto_mapping": null, "peft_version": "0.19.2.dev0@UNKNOWN", "base_model_name_or_path": "meta-llama/Llama-3.2-3B", "revision": null, "inference_mode": false, "num_virtual_tokens": 200, "token_dim": 1024, "num_transformer_submodules": 1, "num_attention_heads": 8, "num_layers": 28, "modules_to_save": null, "init_weights": "zero", "encoder_hidden_size": 3072, "prefix_projection": false }, "error_msg": "" }, "train_info": { "accelerator_memory_reserved_avg": 13973039821, "accelerator_memory_max": 21952987136, "accelerator_memory_reserved_99th": 19184746496, "train_time": 1013.0709846949758, "file_size": 45875328, "num_trainable_params": 11468800, "num_total_params": 3224218624, "status": "success", "metrics": [ { "step": 250, "valid accuracy": 0.2, "train loss": 1.267790999174118, "train samples": 1000, "train time": 27.60945090391033, "eval time": 16.052957416002755, "tokens / sec": 7668.352432536578, "mem allocated avg": 7055651895.296, "mem reserved avg": 14030399864.832, "elapsed time": 68.95821050899394 }, { "step": 500, "valid accuracy": 0.32, "train loss": 0.7903533728122711, "train samples": 2000, "train time": 27.258527728932677, "eval time": 12.650989898000262, "tokens / sec": 7630.456129852915, "mem allocated avg": 7047714314.24, "mem reserved avg": 13710366081.024, "elapsed time": 111.4505693300016 }, { "step": 750, "valid accuracy": 0.24, "train loss": 0.7466984703540802, "train samples": 3000, "train time": 27.920779504042002, "eval time": 12.175382749002893, "tokens / sec": 7678.9045223096955, "mem allocated avg": 7058404001.792, "mem reserved avg": 13860606050.304, "elapsed time": 154.0307989619905 }, { "step": 1000, "valid accuracy": 0.3, "train loss": 0.7248673977851867, "train samples": 4000, "train time": 27.492948249011533, "eval time": 15.898315195008763, "tokens / sec": 7577.797699724343, "mem allocated avg": 7051448203.264, "mem reserved avg": 13920156778.496, "elapsed time": 200.048776636002 }, { "step": 1250, "valid accuracy": 0.42, "train loss": 0.7227549761533737, "train samples": 5000, "train time": 27.528990936058108, "eval time": 15.895380795001984, "tokens / sec": 7575.214089189593, "mem allocated avg": 7049940750.336, "mem reserved avg": 13951748276.224, "elapsed time": 245.96436756699404 }, { "step": 1500, "valid accuracy": 0.36, "train loss": 0.7167108998298645, "train samples": 6000, "train time": 27.818393783119973, "eval time": 13.1919198139949, "tokens / sec": 7524.913250995129, "mem allocated avg": 7052220014.592, "mem reserved avg": 13883515338.752, "elapsed time": 289.5983049130009 }, { "step": 1750, "valid accuracy": 0.42, "train loss": 0.7072831872701645, "train samples": 7000, "train time": 28.521202208925388, "eval time": 15.277781643992057, "tokens / sec": 7340.328730409713, "mem allocated avg": 7052470607.872, "mem reserved avg": 14259249479.68, "elapsed time": 335.9976941280038 }, { "step": 2000, "valid accuracy": 0.24, "train loss": 0.7117487057447434, "train samples": 8000, "train time": 28.099234301873366, "eval time": 15.90539890700893, "tokens / sec": 7391.518137779042, "mem allocated avg": 7049094109.184, "mem reserved avg": 13900166725.632, "elapsed time": 382.6564563080028 }, { "step": 2250, "valid accuracy": 0.42, "train loss": 0.7034323381185532, "train samples": 9000, "train time": 28.748507894968498, "eval time": 15.899852960996213, "tokens / sec": 7476.840216727204, "mem allocated avg": 7059705278.464, "mem reserved avg": 14218715725.824, "elapsed time": 429.91266906999226 }, { "step": 2500, "valid accuracy": 0.38, "train loss": 0.7010540157556534, "train samples": 10000, "train time": 27.407462475908687, "eval time": 12.587275673009572, "tokens / sec": 7514.997062608264, "mem allocated avg": 7045442709.504, "mem reserved avg": 13782650716.16, "elapsed time": 472.4968782359938 }, { "step": 2750, "valid accuracy": 0.42, "train loss": 0.6944627047777175, "train samples": 11000, "train time": 28.060264764819294, "eval time": 12.707225862002815, "tokens / sec": 7550.926613694926, "mem allocated avg": 7056413902.848, "mem reserved avg": 14152219230.208, "elapsed time": 515.7938446079934 }, { "step": 3000, "valid accuracy": 0.38, "train loss": 0.686576269030571, "train samples": 12000, "train time": 28.080755243980093, "eval time": 10.618470949993934, "tokens / sec": 7433.240245372226, "mem allocated avg": 7051565406.208, "mem reserved avg": 14111886802.944, "elapsed time": 557.1905497909902 }, { "step": 3250, "valid accuracy": 0.28, "train loss": 0.6949699774980546, "train samples": 13000, "train time": 28.329060007308726, "eval time": 13.425027397999656, "tokens / sec": 7444.687538011809, "mem allocated avg": 7052971309.056, "mem reserved avg": 13899864735.744, "elapsed time": 601.468493524997 }, { "step": 3500, "valid accuracy": 0.38, "train loss": 0.68068468105793, "train samples": 14000, "train time": 27.756195911890245, "eval time": 11.647340005991282, "tokens / sec": 7556.871289777391, "mem allocated avg": 7052541753.344, "mem reserved avg": 13944114642.944, "elapsed time": 643.4692882009986 }, { "step": 3750, "valid accuracy": 0.36, "train loss": 0.677130599141121, "train samples": 15000, "train time": 28.431524478903157, "eval time": 14.646411609006464, "tokens / sec": 7621.926856605899, "mem allocated avg": 7062554853.376, "mem reserved avg": 14302173986.816, "elapsed time": 689.1648219609924 }, { "step": 4000, "valid accuracy": 0.4, "train loss": 0.6929642289876938, "train samples": 16000, "train time": 27.266019412854803, "eval time": 9.270167384005617, "tokens / sec": 7495.520226309476, "mem allocated avg": 7044563021.824, "mem reserved avg": 13981729161.216, "elapsed time": 728.311128166999 }, { "step": 4250, "valid accuracy": 0.36, "train loss": 0.672608556509018, "train samples": 17000, "train time": 27.917999273864552, "eval time": 11.928998513991246, "tokens / sec": 7571.781843188595, "mem allocated avg": 7055033632.768, "mem reserved avg": 13958316556.288, "elapsed time": 770.647137781998 }, { "step": 4500, "valid accuracy": 0.4, "train loss": 0.6801882609128952, "train samples": 18000, "train time": 28.206865660875337, "eval time": 10.22083333900082, "tokens / sec": 7367.638875532931, "mem allocated avg": 7049804318.72, "mem reserved avg": 13896987443.2, "elapsed time": 811.7163661529921 }, { "step": 4750, "valid accuracy": 0.36, "train loss": 0.6726717475652695, "train samples": 19000, "train time": 27.823305535057443, "eval time": 15.901436312007718, "tokens / sec": 7545.4370342688535, "mem allocated avg": 7052277383.168, "mem reserved avg": 14017733066.752, "elapsed time": 857.9259271740011 }, { "step": 5000, "valid accuracy": 0.38, "train loss": 0.6788799942731857, "train samples": 20000, "train time": 27.60435515793506, "eval time": 12.944955084007233, "tokens / sec": 7545.186214579206, "mem allocated avg": 7049017004.032, "mem reserved avg": 13678195769.344, "elapsed time": 901.0064869999915 }, { "step": 5000, "test accuracy": 0.35784685367702807, "train loss": 0.6788799942731857, "train samples": 20000, "train total tokens": 4198051, "forgetting": 0.3955150842666626 } ] }, "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": "740312add88f781978c0658806c59bc2815b9866", "created_at": "2022-04-12T10:22:10+00:00" } }, "package_info": { "transformers-version": "5.13.1", "transformers-commit-hash": null, "peft-version": "0.19.2.dev0", "peft-commit-hash": "d787f51bbaa196862733ed53c1b7def75b49ab29", "datasets-version": "4.2.0", "datasets-commit-hash": null, "bitsandbytes-version": "0.49.2", "bitsandbytes-commit-hash": null, "torch-version": "2.13.0+cu130", "torch-commit-hash": null }, "system_info": { "system": "Linux", "release": "6.17.0-1009-aws", "version": "#9~24.04.2-Ubuntu SMP Fri Mar 6 23:50:29 UTC 2026", "machine": "x86_64", "processor": "x86_64", "accelerator": "NVIDIA L40S" }, "pytorch_info": "PyTorch built with:\n - GCC 13.3\n - C++ Version: 202002\n - Intel(R) oneAPI Math Kernel Library Version 2024.2-Product Build 20240605 for Intel(R) 64 architecture applications\n - Intel(R) MKL-DNN v3.12.0 (Git Hash 80afa71049cd69a3df32adcccb623b12cd7baa22)\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 13.0\n - NVCC architecture flags: -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 (built against CUDA 12.8)\n - Built with CuDNN 92.0\n - Magma 2.6.1\n - Build settings: BLAS_INFO=mkl, BUILD_TYPE=Release, COMMIT_SHA=cf30153c4c131c8164ee7798e5022d810682e2cb, CUDA_FLAGS= -DLIBCUDACXX_ENABLE_SIMPLIFIED_COMPLEX_OPERATIONS -Xfatbin -compress-all -DONNX_NAMESPACE=onnx_torch -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 -Xcudafe --diag_suppress=cc_clobber_ignored,--diag_suppress=field_without_dll_interface,--diag_suppress=base_class_has_different_dll_interface,--diag_suppress=dll_interface_conflict_none_assumed,--diag_suppress=dll_interface_conflict_dllexport_assumed,--diag_suppress=bad_friend_decl --expt-relaxed-constexpr --expt-extended-lambda -Xfatbin -compress-all --threads 2 -compress-mode=size -Wno-deprecated-gpu-targets --expt-extended-lambda -DCUB_WRAPPED_NAMESPACE=at_cuda_detail -DDISABLE_CUSPARSE_DEPRECATED -DCUDA_HAS_FP16=1 -D__CUDA_NO_HALF_OPERATORS__ -D__CUDA_NO_HALF_CONVERSIONS__ -D__CUDA_NO_HALF2_OPERATORS__ -D__CUDA_NO_BFLOAT16_CONVERSIONS__ -DC10_NODEPRECATED, CUDA_VERSION=13.0, CUDNN_VERSION=9.20.0, CXX_COMPILER=/opt/rh/gcc-toolset-13/root/usr/bin/c++, CXX_FLAGS= -fvisibility-inlines-hidden -DUSE_PTHREADPOOL -DNDEBUG -DUSE_KINETO -DHAS_CUPTI -DUSE_FBGEMM -DUSE_MSLK -DUSE_PYTORCH_QNNPACK -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=2.13.0, USE_CUDA=1, 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=ON, USE_NNPACK=ON, USE_OPENMP=ON, USE_ROCM=OFF, USE_ROCM_KERNEL_ASSERT=OFF, USE_XCCL=OFF, USE_XPU=OFF, \n" } }