testing

MetaMathQA/ptuning--llama-3.2-3B-vt10-MLP.json

@@ -1,353 +0,0 @@
-{
-  "run_info": {
-    "created_at": "2026-07-01T16:24:03+00:00",
-    "total_time": 1246.6367577210185,
-    "experiment_name": "ptuning/llama-3.2-3B-vt10-MLP",
-    "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": 0.0001,
-        "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,
-      "init_kv_cache_prefix": null
-    },
-    "peft_config": {
-      "task_type": "CAUSAL_LM",
-      "peft_type": "P_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": 10,
-      "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": 13880178350,
-    "accelerator_memory_max": 20860370944,
-    "accelerator_memory_reserved_99th": 18765316096,
-    "train_time": 1028.5749839689524,
-    "file_size": 123000,
-    "num_trainable_params": 28351488,
-    "num_total_params": 3241101312,
-    "status": "success",
-    "metrics": [
-      {
-        "step": 250,
-        "valid accuracy": 0.18,
-        "train loss": 0.9363359813690185,
-        "train samples": 1000,
-        "train time": 31.64022434395156,
-        "eval time": 9.38078186599887,
-        "tokens / sec": 6691.450657822938,
-        "mem allocated avg": 7248517412.864,
-        "mem reserved avg": 13954751397.888,
-        "elapsed time": 136.20182907700655
-      },
-      {
-        "step": 500,
-        "valid accuracy": 0.28,
-        "train loss": 0.7981618006229401,
-        "train samples": 2000,
-        "train time": 31.24711563330493,
-        "eval time": 9.311849407007685,
-        "tokens / sec": 6656.454388971098,
-        "mem allocated avg": 7242140928.0,
-        "mem reserved avg": 13628501655.552,
-        "elapsed time": 179.63920924201375
-      },
-      {
-        "step": 750,
-        "valid accuracy": 0.44,
-        "train loss": 0.7558251311779022,
-        "train samples": 3000,
-        "train time": 32.0883745571773,
-        "eval time": 9.31352505000541,
-        "tokens / sec": 6681.578701282153,
-        "mem allocated avg": 7252390766.592,
-        "mem reserved avg": 13869690912.768,
-        "elapsed time": 223.71605581301264
-      },
-      {
-        "step": 1000,
-        "valid accuracy": 0.36,
-        "train loss": 0.7296249723434448,
-        "train samples": 4000,
-        "train time": 31.482282812765334,
-        "eval time": 9.314052008005092,
-        "tokens / sec": 6617.563320901387,
-        "mem allocated avg": 7242649475.072,
-        "mem reserved avg": 13836782403.584,
-        "elapsed time": 267.32254803701653
-      },
-      {
-        "step": 1250,
-        "valid accuracy": 0.36,
-        "train loss": 0.7244644690752029,
-        "train samples": 5000,
-        "train time": 31.3771616689628,
-        "eval time": 9.291028669016669,
-        "tokens / sec": 6646.1715753684175,
-        "mem allocated avg": 7242932428.8,
-        "mem reserved avg": 13901433405.44,
-        "elapsed time": 310.76940306302276
-      },
-      {
-        "step": 1500,
-        "valid accuracy": 0.34,
-        "train loss": 0.7148160248994827,
-        "train samples": 6000,
-        "train time": 31.504103975836188,
-        "eval time": 6.739314348989865,
-        "tokens / sec": 6644.562884904074,
-        "mem allocated avg": 7245105039.36,
-        "mem reserved avg": 13796483530.752,
-        "elapsed time": 351.9117020450067
-      },
-      {
-        "step": 1750,
-        "valid accuracy": 0.46,
-        "train loss": 0.7091476993560791,
-        "train samples": 7000,
-        "train time": 31.73713576295995,
-        "eval time": 9.282792913989397,
-        "tokens / sec": 6596.531002786201,
-        "mem allocated avg": 7245834381.312,
-        "mem reserved avg": 14123672797.184,
-        "elapsed time": 395.7068113429996
-      },
-      {
-        "step": 2000,
-        "valid accuracy": 0.36,
-        "train loss": 0.7094244191646576,
-        "train samples": 8000,
-        "train time": 31.38104235692299,
-        "eval time": 9.307759829011047,
-        "tokens / sec": 6618.518200820059,
-        "mem allocated avg": 7242345316.352,
-        "mem reserved avg": 13885562159.104,
-        "elapsed time": 439.24598261501524
-      },
-      {
-        "step": 2250,
-        "valid accuracy": 0.34,
-        "train loss": 0.7038368511199952,
-        "train samples": 9000,
-        "train time": 32.27969044083147,
-        "eval time": 9.300171014998341,
-        "tokens / sec": 6658.923833052201,
-        "mem allocated avg": 7254741288.96,
-        "mem reserved avg": 14144333938.688,
-        "elapsed time": 483.6396906840091
-      },
-      {
-        "step": 2500,
-        "valid accuracy": 0.3,
-        "train loss": 0.7028931931257248,
-        "train samples": 10000,
-        "train time": 31.062119146925397,
-        "eval time": 9.311563563009258,
-        "tokens / sec": 6630.809669674038,
-        "mem allocated avg": 7238613737.472,
-        "mem reserved avg": 13659019411.456,
-        "elapsed time": 526.8292486710125
-      },
-      {
-        "step": 2750,
-        "valid accuracy": 0.28,
-        "train loss": 0.6949501513242722,
-        "train samples": 11000,
-        "train time": 31.92746263978188,
-        "eval time": 9.288599150982918,
-        "tokens / sec": 6636.324420469121,
-        "mem allocated avg": 7249483808.768,
-        "mem reserved avg": 14054928154.624,
-        "elapsed time": 570.9226945360133
-      },
-      {
-        "step": 3000,
-        "valid accuracy": 0.32,
-        "train loss": 0.6887242776155472,
-        "train samples": 12000,
-        "train time": 31.612302809866378,
-        "eval time": 9.31309081602376,
-        "tokens / sec": 6602.840712219607,
-        "mem allocated avg": 7243983824.896,
-        "mem reserved avg": 13947251982.336,
-        "elapsed time": 614.6747809900262
-      },
-      {
-        "step": 3250,
-        "valid accuracy": 0.46,
-        "train loss": 0.6959465799331666,
-        "train samples": 13000,
-        "train time": 31.669745966763003,
-        "eval time": 9.295396289002383,
-        "tokens / sec": 6659.384013415767,
-        "mem allocated avg": 7247555878.912,
-        "mem reserved avg": 13834626531.328,
-        "elapsed time": 658.3630746390263
-      },
-      {
-        "step": 3500,
-        "valid accuracy": 0.36,
-        "train loss": 0.6824533659219741,
-        "train samples": 14000,
-        "train time": 31.705487885017646,
-        "eval time": 9.105592161999084,
-        "tokens / sec": 6615.573958699966,
-        "mem allocated avg": 7244949321.728,
-        "mem reserved avg": 13871158919.168,
-        "elapsed time": 701.9437575310003
-      },
-      {
-        "step": 3750,
-        "valid accuracy": 0.32,
-        "train loss": 0.6802876671552658,
-        "train samples": 15000,
-        "train time": 32.53438692694181,
-        "eval time": 9.274032826011535,
-        "tokens / sec": 6660.737160550203,
-        "mem allocated avg": 7255717593.088,
-        "mem reserved avg": 14150264684.544,
-        "elapsed time": 746.4688438280136
-      },
-      {
-        "step": 4000,
-        "valid accuracy": 0.34,
-        "train loss": 0.6954794878959656,
-        "train samples": 16000,
-        "train time": 31.074721232173033,
-        "eval time": 9.317057678010315,
-        "tokens / sec": 6576.824888404907,
-        "mem allocated avg": 7237630017.536,
-        "mem reserved avg": 13796911349.76,
-        "elapsed time": 789.6863839280268
-      },
-      {
-        "step": 4250,
-        "valid accuracy": 0.36,
-        "train loss": 0.6776138496398926,
-        "train samples": 17000,
-        "train time": 31.751483836880652,
-        "eval time": 7.7834975119913,
-        "tokens / sec": 6657.610116301494,
-        "mem allocated avg": 7248522713.088,
-        "mem reserved avg": 13859121266.688,
-        "elapsed time": 831.8446559070144
-      },
-      {
-        "step": 4500,
-        "valid accuracy": 0.3,
-        "train loss": 0.6849392641782761,
-        "train samples": 18000,
-        "train time": 31.408977424085606,
-        "eval time": 7.941874133015517,
-        "tokens / sec": 6616.515946827266,
-        "mem allocated avg": 7242914779.136,
-        "mem reserved avg": 13793170030.592,
-        "elapsed time": 874.0072624590248
-      },
-      {
-        "step": 4750,
-        "valid accuracy": 0.28,
-        "train loss": 0.6771339122056961,
-        "train samples": 19000,
-        "train time": 31.56458607915556,
-        "eval time": 9.30490853099036,
-        "tokens / sec": 6651.093078601728,
-        "mem allocated avg": 7245515804.672,
-        "mem reserved avg": 13896693841.92,
-        "elapsed time": 917.6395900130155
-      },
-      {
-        "step": 5000,
-        "valid accuracy": 0.38,
-        "train loss": 0.683932776093483,
-        "train samples": 20000,
-        "train time": 31.327182788023492,
-        "eval time": 9.292242444003932,
-        "tokens / sec": 6648.539110884439,
-        "mem allocated avg": 7242581661.696,
-        "mem reserved avg": 13599208636.416,
-        "elapsed time": 960.9633954950259
-      },
-      {
-        "step": 5000,
-        "test accuracy": 0.33131159969673996,
-        "train loss": 0.683932776093483,
-        "train samples": 20000,
-        "train total tokens": 4198051,
-        "forgetting": 1.3030481338500977
-      }
-    ]
-  },
-  "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": "5.12.1",
-      "transformers-commit-hash": null,
-      "peft-version": "0.19.2.dev0",
-      "peft-commit-hash": "36b70a3480a2f4eb274eb5486a050652a2b6cd35",
-      "datasets-version": "5.0.0",
-      "datasets-commit-hash": null,
-      "bitsandbytes-version": "0.49.2",
-      "bitsandbytes-commit-hash": null,
-      "torch-version": "2.12.0+cu130",
-      "torch-commit-hash": null
-    },
-    "system_info": {
-      "system": "Linux",
-      "release": "5.15.0-185-generic",
-      "version": "#195-Ubuntu SMP Fri Jun 19 17:11:50 UTC 2026",
-      "machine": "x86_64",
-      "processor": "x86_64",
-      "accelerator": "NVIDIA GeForce RTX 4090"
-    },
-    "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 v3.11.2 (Git Hash 03c022d3ffdcee958cfacbe720048e725fdf644c)\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 92.0  (built against CUDA 13.2)\n  - Magma 2.6.1\n  - Build settings: BLAS_INFO=mkl, BUILD_TYPE=Release, COMMIT_SHA=7661cd9c6b841b62b7f411aa52ec51f05457263b, 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 -DLIBKINETO_NOROCTRACER -DLIBKINETO_NOXPUPTI=ON -DUSE_FBGEMM -DUSE_MSLK -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=2.12.0, 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"
-  }
-}

MetaMathQA/ptuning--llama-3.2-3B-vt20-LSTM.json

@@ -1,353 +0,0 @@
-{
-  "run_info": {
-    "created_at": "2026-06-25T13:04:16+00:00",
-    "total_time": 1502.6669879511464,
-    "experiment_name": "ptuning/llama-3.2-3B-vt20-LSTM",
-    "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": 0.0001,
-        "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,
-      "init_kv_cache_prefix": null
-    },
-    "peft_config": {
-      "task_type": "CAUSAL_LM",
-      "peft_type": "P_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": 20,
-      "token_dim": 3072,
-      "num_transformer_submodules": 1,
-      "num_attention_heads": 24,
-      "num_layers": 28,
-      "modules_to_save": null,
-      "encoder_reparameterization_type": "LSTM",
-      "encoder_hidden_size": 3072,
-      "encoder_num_layers": 2,
-      "encoder_dropout": 0.0
-    },
-    "error_msg": ""
-  },
-  "train_info": {
-    "accelerator_memory_reserved_avg": 21550333952,
-    "accelerator_memory_max": 28781314048,
-    "accelerator_memory_reserved_99th": 26820749557,
-    "train_time": 1322.590588517487,
-    "file_size": 245880,
-    "num_trainable_params": 434279424,
-    "num_total_params": 3647029248,
-    "status": "success",
-    "metrics": [
-      {
-        "step": 250,
-        "valid accuracy": 0.22,
-        "train loss": 0.9354020810127258,
-        "train samples": 1000,
-        "train time": 50.49005669821054,
-        "eval time": 9.3927109639626,
-        "tokens / sec": 4193.2810902845295,
-        "mem allocated avg": 13755032733.696,
-        "mem reserved avg": 21591454908.416,
-        "elapsed time": 119.27638081205077
-      },
-      {
-        "step": 500,
-        "valid accuracy": 0.2,
-        "train loss": 0.7862321782112122,
-        "train samples": 2000,
-        "train time": 50.365367197664455,
-        "eval time": 9.254252284998074,
-        "tokens / sec": 4129.722695829867,
-        "mem allocated avg": 13747812104.192,
-        "mem reserved avg": 21267361038.336,
-        "elapsed time": 179.84149901219644
-      },
-      {
-        "step": 750,
-        "valid accuracy": 0.32,
-        "train loss": 0.7509800347089768,
-        "train samples": 3000,
-        "train time": 50.671867781318724,
-        "eval time": 9.250419595045969,
-        "tokens / sec": 4231.164340049126,
-        "mem allocated avg": 13758180833.28,
-        "mem reserved avg": 21481211822.08,
-        "elapsed time": 240.75460329209454
-      },
-      {
-        "step": 1000,
-        "valid accuracy": 0.28,
-        "train loss": 0.7235222470760345,
-        "train samples": 4000,
-        "train time": 50.121545881032944,
-        "eval time": 9.264669175958261,
-        "tokens / sec": 4156.615609871657,
-        "mem allocated avg": 13750246084.608,
-        "mem reserved avg": 21661600448.512,
-        "elapsed time": 301.1448435881175
-      },
-      {
-        "step": 1250,
-        "valid accuracy": 0.3,
-        "train loss": 0.7182610543966294,
-        "train samples": 5000,
-        "train time": 49.812087612459436,
-        "eval time": 9.25290141813457,
-        "tokens / sec": 4186.493881212853,
-        "mem allocated avg": 13749475418.112,
-        "mem reserved avg": 21545644720.128,
-        "elapsed time": 361.17410576017573
-      },
-      {
-        "step": 1500,
-        "valid accuracy": 0.36,
-        "train loss": 0.7089396673440933,
-        "train samples": 6000,
-        "train time": 50.24829094391316,
-        "eval time": 9.26507464889437,
-        "tokens / sec": 4165.932732590925,
-        "mem allocated avg": 13751175532.544,
-        "mem reserved avg": 21550510112.768,
-        "elapsed time": 421.5853925740812
-      },
-      {
-        "step": 1750,
-        "valid accuracy": 0.24,
-        "train loss": 0.7019602956771851,
-        "train samples": 7000,
-        "train time": 50.260878534056246,
-        "eval time": 9.258642738917843,
-        "tokens / sec": 4165.366903766779,
-        "mem allocated avg": 13752757673.984,
-        "mem reserved avg": 21844916699.136,
-        "elapsed time": 482.5431994642131
-      },
-      {
-        "step": 2000,
-        "valid accuracy": 0.32,
-        "train loss": 0.703207641363144,
-        "train samples": 8000,
-        "train time": 49.82990423007868,
-        "eval time": 6.006760481977835,
-        "tokens / sec": 4168.099521945881,
-        "mem allocated avg": 13748972744.704,
-        "mem reserved avg": 21610698375.168,
-        "elapsed time": 539.6426566240843
-      },
-      {
-        "step": 2250,
-        "valid accuracy": 0.26,
-        "train loss": 0.6957956521511078,
-        "train samples": 9000,
-        "train time": 50.82219616463408,
-        "eval time": 6.473255675053224,
-        "tokens / sec": 4229.411875545376,
-        "mem allocated avg": 13760203024.384,
-        "mem reserved avg": 21905373396.992,
-        "elapsed time": 598.0294065510388
-      },
-      {
-        "step": 2500,
-        "valid accuracy": 0.26,
-        "train loss": 0.6942941577434539,
-        "train samples": 10000,
-        "train time": 49.56321152066812,
-        "eval time": 6.418301625177264,
-        "tokens / sec": 4155.642737438648,
-        "mem allocated avg": 13745219473.408,
-        "mem reserved avg": 21274810122.24,
-        "elapsed time": 655.3750411360525
-      },
-      {
-        "step": 2750,
-        "valid accuracy": 0.3,
-        "train loss": 0.687157685995102,
-        "train samples": 11000,
-        "train time": 50.7322088687215,
-        "eval time": 9.255984412971884,
-        "tokens / sec": 4176.459190812671,
-        "mem allocated avg": 13755800027.136,
-        "mem reserved avg": 21811186106.368,
-        "elapsed time": 716.4680303991772
-      },
-      {
-        "step": 3000,
-        "valid accuracy": 0.28,
-        "train loss": 0.6796072626113892,
-        "train samples": 12000,
-        "train time": 49.92008067877032,
-        "eval time": 9.27642115810886,
-        "tokens / sec": 4181.303338493356,
-        "mem allocated avg": 13751262904.32,
-        "mem reserved avg": 21667245981.696,
-        "elapsed time": 776.6304151872173
-      },
-      {
-        "step": 3250,
-        "valid accuracy": 0.36,
-        "train loss": 0.688059378027916,
-        "train samples": 13000,
-        "train time": 50.41137535823509,
-        "eval time": 9.249134139157832,
-        "tokens / sec": 4183.599405913604,
-        "mem allocated avg": 13752641021.952,
-        "mem reserved avg": 21651131465.728,
-        "elapsed time": 837.2296860460192
-      },
-      {
-        "step": 3500,
-        "valid accuracy": 0.32,
-        "train loss": 0.6734865410327912,
-        "train samples": 14000,
-        "train time": 50.12707422138192,
-        "eval time": 9.280864350032061,
-        "tokens / sec": 4184.365500241589,
-        "mem allocated avg": 13751505661.952,
-        "mem reserved avg": 21513927393.28,
-        "elapsed time": 897.8190522671212
-      },
-      {
-        "step": 3750,
-        "valid accuracy": 0.28,
-        "train loss": 0.670322101354599,
-        "train samples": 15000,
-        "train time": 51.18098081322387,
-        "eval time": 6.044083565939218,
-        "tokens / sec": 4234.05328613807,
-        "mem allocated avg": 13762761537.536,
-        "mem reserved avg": 21592427986.944,
-        "elapsed time": 955.9538224241696
-      },
-      {
-        "step": 4000,
-        "valid accuracy": 0.38,
-        "train loss": 0.6844290305376053,
-        "train samples": 16000,
-        "train time": 49.46939493622631,
-        "eval time": 5.15820625401102,
-        "tokens / sec": 4131.301793027151,
-        "mem allocated avg": 13744092170.24,
-        "mem reserved avg": 21539613310.976,
-        "elapsed time": 1011.6463001200464
-      },
-      {
-        "step": 4250,
-        "valid accuracy": 0.4,
-        "train loss": 0.6669189376831055,
-        "train samples": 17000,
-        "train time": 50.603932400466874,
-        "eval time": 6.134445908013731,
-        "tokens / sec": 4177.323578869727,
-        "mem allocated avg": 13753937317.888,
-        "mem reserved avg": 21385682354.176,
-        "elapsed time": 1069.237313566031
-      },
-      {
-        "step": 4500,
-        "valid accuracy": 0.46,
-        "train loss": 0.6736755061149597,
-        "train samples": 18000,
-        "train time": 49.98166696308181,
-        "eval time": 9.266396904131398,
-        "tokens / sec": 4157.884533012906,
-        "mem allocated avg": 13749339627.52,
-        "mem reserved avg": 21371648212.992,
-        "elapsed time": 1129.6838904172182
-      },
-      {
-        "step": 4750,
-        "valid accuracy": 0.3,
-        "train loss": 0.664890606880188,
-        "train samples": 19000,
-        "train time": 50.29726964957081,
-        "eval time": 9.265053241048008,
-        "tokens / sec": 4173.964142838744,
-        "mem allocated avg": 13751988727.808,
-        "mem reserved avg": 21558370238.464,
-        "elapsed time": 1190.1546793282032
-      },
-      {
-        "step": 5000,
-        "valid accuracy": 0.34,
-        "train loss": 0.6717871301174164,
-        "train samples": 20000,
-        "train time": 49.993743317900226,
-        "eval time": 8.915772874141112,
-        "tokens / sec": 4166.121321934009,
-        "mem allocated avg": 13748624703.488,
-        "mem reserved avg": 21181864345.6,
-        "elapsed time": 1250.1847356450744
-      },
-      {
-        "step": 5000,
-        "test accuracy": 0.33813495072024263,
-        "train loss": 0.6717871301174164,
-        "train samples": 20000,
-        "train total tokens": 4198051,
-        "forgetting": 1.6402645111083984
-      }
-    ]
-  },
-  "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": "5.12.1",
-      "transformers-commit-hash": null,
-      "peft-version": "0.19.2.dev0",
-      "peft-commit-hash": "0fecbcd8641b42b87b558030ad419475da669fe1",
-      "datasets-version": "5.0.0",
-      "datasets-commit-hash": null,
-      "bitsandbytes-version": "0.49.2",
-      "bitsandbytes-commit-hash": null,
-      "torch-version": "2.12.0+cu130",
-      "torch-commit-hash": null
-    },
-    "system_info": {
-      "system": "Linux",
-      "release": "6.8.0-90-generic",
-      "version": "#91~22.04.1-Ubuntu SMP PREEMPT_DYNAMIC Thu Nov 20 15:20:45 UTC 2",
-      "machine": "x86_64",
-      "processor": "x86_64",
-      "accelerator": "NVIDIA GeForce RTX 4090"
-    },
-    "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 v3.11.2 (Git Hash 03c022d3ffdcee958cfacbe720048e725fdf644c)\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: AVX512\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 92.0  (built against CUDA 13.2)\n  - Magma 2.6.1\n  - Build settings: BLAS_INFO=mkl, BUILD_TYPE=Release, COMMIT_SHA=7661cd9c6b841b62b7f411aa52ec51f05457263b, 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 -DLIBKINETO_NOROCTRACER -DLIBKINETO_NOXPUPTI=ON -DUSE_FBGEMM -DUSE_MSLK -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=2.12.0, 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"
-  }
-}

MetaMathQA/ptuning--llama-3.2-3B-vt20-MLP-lr5e-4.json

@@ -1,352 +0,0 @@
-{
-  "run_info": {
-    "created_at": "2026-06-25T13:50:54+00:00",
-    "total_time": 1098.8279759250581,
-    "experiment_name": "ptuning/llama-3.2-3B-vt20-MLP-lr5e-4",
-    "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": 0.0005
-      },
-      "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": "P_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": 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": 14115343748,
-    "accelerator_memory_max": 20946354176,
-    "accelerator_memory_reserved_99th": 19216203776,
-    "train_time": 928.1491970848292,
-    "file_size": 245880,
-    "num_trainable_params": 28382208,
-    "num_total_params": 3241132032,
-    "status": "success",
-    "metrics": [
-      {
-        "step": 250,
-        "valid accuracy": 0.24,
-        "train loss": 0.8859602637290954,
-        "train samples": 1000,
-        "train time": 31.65180461597629,
-        "eval time": 9.40215106192045,
-        "tokens / sec": 6689.00249349873,
-        "mem allocated avg": 7265250011.136,
-        "mem reserved avg": 14200797659.136,
-        "elapsed time": 57.33147974405438
-      },
-      {
-        "step": 500,
-        "valid accuracy": 0.28,
-        "train loss": 0.7676331839561462,
-        "train samples": 2000,
-        "train time": 31.13863810757175,
-        "eval time": 9.25840079691261,
-        "tokens / sec": 6679.643447521984,
-        "mem allocated avg": 7257767559.168,
-        "mem reserved avg": 13883649556.48,
-        "elapsed time": 100.10863005090505
-      },
-      {
-        "step": 750,
-        "valid accuracy": 0.34,
-        "train loss": 0.7299470043182373,
-        "train samples": 3000,
-        "train time": 32.39704798278399,
-        "eval time": 9.253981592133641,
-        "tokens / sec": 6617.917784173859,
-        "mem allocated avg": 7268295294.976,
-        "mem reserved avg": 14123421138.944,
-        "elapsed time": 144.0450889300555
-      },
-      {
-        "step": 1000,
-        "valid accuracy": 0.32,
-        "train loss": 0.7054928681850433,
-        "train samples": 4000,
-        "train time": 31.707324855495244,
-        "eval time": 9.251647465862334,
-        "tokens / sec": 6570.595310373305,
-        "mem allocated avg": 7259502956.544,
-        "mem reserved avg": 14061144113.152,
-        "elapsed time": 187.4869728749618
-      },
-      {
-        "step": 1250,
-        "valid accuracy": 0.32,
-        "train loss": 0.7035304119586945,
-        "train samples": 5000,
-        "train time": 31.407017521560192,
-        "eval time": 8.274948939215392,
-        "tokens / sec": 6639.853652351532,
-        "mem allocated avg": 7258852329.472,
-        "mem reserved avg": 14130685673.472,
-        "elapsed time": 229.53703149803914
-      },
-      {
-        "step": 1500,
-        "valid accuracy": 0.38,
-        "train loss": 0.6977702566385269,
-        "train samples": 6000,
-        "train time": 31.798848503269255,
-        "eval time": 5.635441232006997,
-        "tokens / sec": 6582.974222430683,
-        "mem allocated avg": 7260632471.552,
-        "mem reserved avg": 14072930107.392,
-        "elapsed time": 269.7190164339263
-      },
-      {
-        "step": 1750,
-        "valid accuracy": 0.36,
-        "train loss": 0.6920007370710373,
-        "train samples": 7000,
-        "train time": 31.98915307270363,
-        "eval time": 9.249000354902819,
-        "tokens / sec": 6544.562137177767,
-        "mem allocated avg": 7262342055.936,
-        "mem reserved avg": 14363075280.896,
-        "elapsed time": 313.50859929597937
-      },
-      {
-        "step": 2000,
-        "valid accuracy": 0.4,
-        "train loss": 0.6937211427688599,
-        "train samples": 8000,
-        "train time": 31.306376459077,
-        "eval time": 9.252411246998236,
-        "tokens / sec": 6634.303406895257,
-        "mem allocated avg": 7258369662.976,
-        "mem reserved avg": 14144929529.856,
-        "elapsed time": 356.8602735430468
-      },
-      {
-        "step": 2250,
-        "valid accuracy": 0.4,
-        "train loss": 0.6870029544830323,
-        "train samples": 9000,
-        "train time": 32.82327748509124,
-        "eval time": 9.250748039921746,
-        "tokens / sec": 6548.64524414517,
-        "mem allocated avg": 7270847238.144,
-        "mem reserved avg": 14310663258.112,
-        "elapsed time": 401.30239249696024
-      },
-      {
-        "step": 2500,
-        "valid accuracy": 0.46,
-        "train loss": 0.6864559572935104,
-        "train samples": 10000,
-        "train time": 31.70231397682801,
-        "eval time": 6.498107424937189,
-        "tokens / sec": 6496.907454469925,
-        "mem allocated avg": 7255026315.264,
-        "mem reserved avg": 13888665944.064,
-        "elapsed time": 442.19169994909316
-      },
-      {
-        "step": 2750,
-        "valid accuracy": 0.38,
-        "train loss": 0.680001744389534,
-        "train samples": 11000,
-        "train time": 32.469573070760816,
-        "eval time": 9.25120837194845,
-        "tokens / sec": 6525.524666993574,
-        "mem allocated avg": 7265801664.512,
-        "mem reserved avg": 14312777187.328,
-        "elapsed time": 486.26620708499104
-      },
-      {
-        "step": 3000,
-        "valid accuracy": 0.38,
-        "train loss": 0.672161677122116,
-        "train samples": 12000,
-        "train time": 31.73648655274883,
-        "eval time": 5.958063455065712,
-        "tokens / sec": 6577.004031403121,
-        "mem allocated avg": 7260756502.528,
-        "mem reserved avg": 14207953141.76,
-        "elapsed time": 526.7587625021115
-      },
-      {
-        "step": 3250,
-        "valid accuracy": 0.28,
-        "train loss": 0.681042181134224,
-        "train samples": 13000,
-        "train time": 31.950469674542546,
-        "eval time": 6.172452438157052,
-        "tokens / sec": 6600.873231232698,
-        "mem allocated avg": 7262658037.76,
-        "mem reserved avg": 14124192890.88,
-        "elapsed time": 567.2247899330687
-      },
-      {
-        "step": 3500,
-        "valid accuracy": 0.44,
-        "train loss": 0.6662635884284973,
-        "train samples": 14000,
-        "train time": 31.594176259357482,
-        "eval time": 9.235800415975973,
-        "tokens / sec": 6638.881744475828,
-        "mem allocated avg": 7261104484.352,
-        "mem reserved avg": 14062100414.464,
-        "elapsed time": 610.4624923849478
-      },
-      {
-        "step": 3750,
-        "valid accuracy": 0.44,
-        "train loss": 0.6641098285913467,
-        "train samples": 15000,
-        "train time": 32.448025634046644,
-        "eval time": 8.274530452908948,
-        "tokens / sec": 6678.46489163953,
-        "mem allocated avg": 7272199753.728,
-        "mem reserved avg": 14347447304.192,
-        "elapsed time": 653.7124716730323
-      },
-      {
-        "step": 4000,
-        "valid accuracy": 0.32,
-        "train loss": 0.6793096128702164,
-        "train samples": 16000,
-        "train time": 30.9815093385987,
-        "eval time": 9.262340698158368,
-        "tokens / sec": 6596.612120035719,
-        "mem allocated avg": 7254373582.848,
-        "mem reserved avg": 14100260192.256,
-        "elapsed time": 696.7475293360185
-      },
-      {
-        "step": 4250,
-        "valid accuracy": 0.42,
-        "train loss": 0.6622657473087311,
-        "train samples": 17000,
-        "train time": 31.90420175320469,
-        "eval time": 9.248332449002191,
-        "tokens / sec": 6625.741701209201,
-        "mem allocated avg": 7264152272.896,
-        "mem reserved avg": 14078374313.984,
-        "elapsed time": 740.1776858849917
-      },
-      {
-        "step": 4500,
-        "valid accuracy": 0.36,
-        "train loss": 0.6690288419723511,
-        "train samples": 18000,
-        "train time": 31.268609202466905,
-        "eval time": 9.239017388084903,
-        "tokens / sec": 6646.218213747877,
-        "mem allocated avg": 7258554871.808,
-        "mem reserved avg": 14031742042.112,
-        "elapsed time": 783.2471806830727
-      },
-      {
-        "step": 4750,
-        "valid accuracy": 0.34,
-        "train loss": 0.6596376706361771,
-        "train samples": 19000,
-        "train time": 31.710480290697888,
-        "eval time": 8.497989056166261,
-        "tokens / sec": 6620.492596625368,
-        "mem allocated avg": 7261569159.168,
-        "mem reserved avg": 14077401235.456,
-        "elapsed time": 825.7542321349028
-      },
-      {
-        "step": 5000,
-        "valid accuracy": 0.4,
-        "train loss": 0.6666479827165603,
-        "train samples": 20000,
-        "train time": 31.27789579378441,
-        "eval time": 6.808006115956232,
-        "tokens / sec": 6659.01572705507,
-        "mem allocated avg": 7258790572.032,
-        "mem reserved avg": 13784663982.08,
-        "elapsed time": 866.2085020889062
-      },
-      {
-        "step": 5000,
-        "test accuracy": 0.3525398028809704,
-        "train loss": 0.6666479827165603,
-        "train samples": 20000,
-        "train total tokens": 4198051,
-        "forgetting": 3.314358711242676
-      }
-    ]
-  },
-  "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": "5.12.1",
-      "transformers-commit-hash": null,
-      "peft-version": "0.19.2.dev0",
-      "peft-commit-hash": "f8d0501b710e8949ffecfaf580f4053be72fdcda",
-      "datasets-version": "5.0.0",
-      "datasets-commit-hash": null,
-      "bitsandbytes-version": "0.49.2",
-      "bitsandbytes-commit-hash": null,
-      "torch-version": "2.12.0+cu130",
-      "torch-commit-hash": null
-    },
-    "system_info": {
-      "system": "Linux",
-      "release": "6.8.0-90-generic",
-      "version": "#91~22.04.1-Ubuntu SMP PREEMPT_DYNAMIC Thu Nov 20 15:20:45 UTC 2",
-      "machine": "x86_64",
-      "processor": "x86_64",
-      "accelerator": "NVIDIA GeForce RTX 4090"
-    },
-    "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 v3.11.2 (Git Hash 03c022d3ffdcee958cfacbe720048e725fdf644c)\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: AVX512\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 92.0  (built against CUDA 13.2)\n  - Magma 2.6.1\n  - Build settings: BLAS_INFO=mkl, BUILD_TYPE=Release, COMMIT_SHA=7661cd9c6b841b62b7f411aa52ec51f05457263b, 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 -DLIBKINETO_NOROCTRACER -DLIBKINETO_NOXPUPTI=ON -DUSE_FBGEMM -DUSE_MSLK -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=2.12.0, 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"
-  }
-}

MetaMathQA/ptuning--llama-3.2-3B-vt30-MLP.json

@@ -1,353 +0,0 @@
-{
-  "run_info": {
-    "created_at": "2026-07-01T15:50:58+00:00",
-    "total_time": 1492.634669919993,
-    "experiment_name": "ptuning/llama-3.2-3B-vt30-MLP",
-    "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": 0.0001,
-        "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,
-      "init_kv_cache_prefix": null
-    },
-    "peft_config": {
-      "task_type": "CAUSAL_LM",
-      "peft_type": "P_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": 30,
-      "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": 14323580875,
-    "accelerator_memory_max": 21458059264,
-    "accelerator_memory_reserved_99th": 19736297472,
-    "train_time": 1194.9091126141138,
-    "file_size": 368760,
-    "num_trainable_params": 28412928,
-    "num_total_params": 3241162752,
-    "status": "success",
-    "metrics": [
-      {
-        "step": 250,
-        "valid accuracy": 0.08,
-        "train loss": 0.9363302497863769,
-        "train samples": 1000,
-        "train time": 34.00575314508751,
-        "eval time": 9.754597692983225,
-        "tokens / sec": 6225.975913448782,
-        "mem allocated avg": 7281153935.36,
-        "mem reserved avg": 14393391710.208,
-        "elapsed time": 258.6250282860128
-      },
-      {
-        "step": 500,
-        "valid accuracy": 0.14,
-        "train loss": 0.7985411801338196,
-        "train samples": 2000,
-        "train time": 33.73720351737575,
-        "eval time": 9.62405165200471,
-        "tokens / sec": 6165.1523634102,
-        "mem allocated avg": 7274907719.68,
-        "mem reserved avg": 14096116219.904,
-        "elapsed time": 304.5419928780175
-      },
-      {
-        "step": 750,
-        "valid accuracy": 0.22,
-        "train loss": 0.7601689212322235,
-        "train samples": 3000,
-        "train time": 34.6107883961231,
-        "eval time": 9.648138158983784,
-        "tokens / sec": 6194.629187470805,
-        "mem allocated avg": 7283639179.264,
-        "mem reserved avg": 14317038600.192,
-        "elapsed time": 351.2987851499929
-      },
-      {
-        "step": 1000,
-        "valid accuracy": 0.28,
-        "train loss": 0.7293364794254303,
-        "train samples": 4000,
-        "train time": 34.106910182803404,
-        "eval time": 9.637816358997952,
-        "tokens / sec": 6108.322298425096,
-        "mem allocated avg": 7276656787.456,
-        "mem reserved avg": 14334067474.432,
-        "elapsed time": 397.75173320499016
-      },
-      {
-        "step": 1250,
-        "valid accuracy": 0.32,
-        "train loss": 0.7207725659608841,
-        "train samples": 5000,
-        "train time": 33.97287149247131,
-        "eval time": 9.646640760998707,
-        "tokens / sec": 6138.368375667445,
-        "mem allocated avg": 7275547160.576,
-        "mem reserved avg": 14281647063.04,
-        "elapsed time": 444.0086419559957
-      },
-      {
-        "step": 1500,
-        "valid accuracy": 0.38,
-        "train loss": 0.7102297298908233,
-        "train samples": 6000,
-        "train time": 34.26030106123653,
-        "eval time": 9.622574922977947,
-        "tokens / sec": 6110.016360505525,
-        "mem allocated avg": 7277322196.992,
-        "mem reserved avg": 14224101212.16,
-        "elapsed time": 490.5009809990006
-      },
-      {
-        "step": 1750,
-        "valid accuracy": 0.28,
-        "train loss": 0.7035963997840882,
-        "train samples": 7000,
-        "train time": 34.326304641144816,
-        "eval time": 9.675930661993334,
-        "tokens / sec": 6098.967022190298,
-        "mem allocated avg": 7277080172.544,
-        "mem reserved avg": 14590775656.448,
-        "elapsed time": 537.2490275430027
-      },
-      {
-        "step": 2000,
-        "valid accuracy": 0.36,
-        "train loss": 0.7042187093496323,
-        "train samples": 8000,
-        "train time": 33.91284731691121,
-        "eval time": 7.696966222982155,
-        "tokens / sec": 6124.404655825785,
-        "mem allocated avg": 7274753998.848,
-        "mem reserved avg": 14271337463.808,
-        "elapsed time": 581.5433360229945
-      },
-      {
-        "step": 2250,
-        "valid accuracy": 0.38,
-        "train loss": 0.6965467965602875,
-        "train samples": 9000,
-        "train time": 35.00853349498357,
-        "eval time": 9.622937389998697,
-        "tokens / sec": 6139.874440350387,
-        "mem allocated avg": 7287149615.104,
-        "mem reserved avg": 14514305105.92,
-        "elapsed time": 628.8494034899923
-      },
-      {
-        "step": 2500,
-        "valid accuracy": 0.34,
-        "train loss": 0.6948007898330688,
-        "train samples": 10000,
-        "train time": 33.72064900404075,
-        "eval time": 5.387909529992612,
-        "tokens / sec": 6108.037836855362,
-        "mem allocated avg": 7272138412.032,
-        "mem reserved avg": 14066789646.336,
-        "elapsed time": 670.5936161220015
-      },
-      {
-        "step": 2750,
-        "valid accuracy": 0.32,
-        "train loss": 0.6874945278167724,
-        "train samples": 11000,
-        "train time": 34.540225339180324,
-        "eval time": 9.61607060598908,
-        "tokens / sec": 6134.325932137308,
-        "mem allocated avg": 7281916778.496,
-        "mem reserved avg": 14505924886.528,
-        "elapsed time": 717.2573533540126
-      },
-      {
-        "step": 3000,
-        "valid accuracy": 0.3,
-        "train loss": 0.6796259649991989,
-        "train samples": 12000,
-        "train time": 34.26134033527342,
-        "eval time": 9.50968838101835,
-        "tokens / sec": 6092.318571235321,
-        "mem allocated avg": 7276503754.752,
-        "mem reserved avg": 14410059874.304,
-        "elapsed time": 763.75420076499
-      },
-      {
-        "step": 3250,
-        "valid accuracy": 0.32,
-        "train loss": 0.68837134039402,
-        "train samples": 13000,
-        "train time": 34.255930323008215,
-        "eval time": 9.612164622027194,
-        "tokens / sec": 6156.627422211534,
-        "mem allocated avg": 7278948605.952,
-        "mem reserved avg": 14338605711.36,
-        "elapsed time": 810.1472433689923
-      },
-      {
-        "step": 3500,
-        "valid accuracy": 0.34,
-        "train loss": 0.6742574932575226,
-        "train samples": 14000,
-        "train time": 34.28129725190229,
-        "eval time": 6.6055597880040295,
-        "tokens / sec": 6118.49658018297,
-        "mem allocated avg": 7278363557.888,
-        "mem reserved avg": 14407107084.288,
-        "elapsed time": 853.6943467550154
-      },
-      {
-        "step": 3750,
-        "valid accuracy": 0.4,
-        "train loss": 0.6714792121648788,
-        "train samples": 15000,
-        "train time": 35.124471633374924,
-        "eval time": 8.83637132399599,
-        "tokens / sec": 6169.573232643049,
-        "mem allocated avg": 7288971911.168,
-        "mem reserved avg": 14592495321.088,
-        "elapsed time": 900.199428027001
-      },
-      {
-        "step": 4000,
-        "valid accuracy": 0.32,
-        "train loss": 0.6868616794347763,
-        "train samples": 16000,
-        "train time": 33.620701181003824,
-        "eval time": 9.615348125982564,
-        "tokens / sec": 6078.784582740162,
-        "mem allocated avg": 7270337929.216,
-        "mem reserved avg": 14266983776.256,
-        "elapsed time": 946.0162565729988
-      },
-      {
-        "step": 4250,
-        "valid accuracy": 0.38,
-        "train loss": 0.6681802560091019,
-        "train samples": 17000,
-        "train time": 34.270348728779936,
-        "eval time": 6.986196043988457,
-        "tokens / sec": 6168.276887783093,
-        "mem allocated avg": 7279699822.592,
-        "mem reserved avg": 14257932468.224,
-        "elapsed time": 989.7307399639976
-      },
-      {
-        "step": 4500,
-        "valid accuracy": 0.4,
-        "train loss": 0.6755580611228943,
-        "train samples": 18000,
-        "train time": 34.087618150311755,
-        "eval time": 8.943682793003973,
-        "tokens / sec": 6096.583195798893,
-        "mem allocated avg": 7274260011.008,
-        "mem reserved avg": 14269450027.008,
-        "elapsed time": 1035.466415124014
-      },
-      {
-        "step": 4750,
-        "valid accuracy": 0.38,
-        "train loss": 0.6677044332027435,
-        "train samples": 19000,
-        "train time": 34.23474362070556,
-        "eval time": 9.606772648985498,
-        "tokens / sec": 6132.337438421082,
-        "mem allocated avg": 7277436657.664,
-        "mem reserved avg": 14318162673.664,
-        "elapsed time": 1081.8005726340052
-      },
-      {
-        "step": 5000,
-        "valid accuracy": 0.44,
-        "train loss": 0.673769939661026,
-        "train samples": 20000,
-        "train time": 33.93519900739193,
-        "eval time": 9.618838055990636,
-        "tokens / sec": 6137.580037607307,
-        "mem allocated avg": 7274845607.936,
-        "mem reserved avg": 14015325536.256,
-        "elapsed time": 1127.8809662260173
-      },
-      {
-        "step": 5000,
-        "test accuracy": 0.3419257012888552,
-        "train loss": 0.673769939661026,
-        "train samples": 20000,
-        "train total tokens": 4198051,
-        "forgetting": 1.8561630249023438
-      }
-    ]
-  },
-  "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": "5.12.1",
-      "transformers-commit-hash": null,
-      "peft-version": "0.19.2.dev0",
-      "peft-commit-hash": "0fba0c34a9900c5bb11ce7c6964f9d14cf7a5c4b",
-      "datasets-version": "5.0.0",
-      "datasets-commit-hash": null,
-      "bitsandbytes-version": "0.49.2",
-      "bitsandbytes-commit-hash": null,
-      "torch-version": "2.12.0+cu130",
-      "torch-commit-hash": null
-    },
-    "system_info": {
-      "system": "Linux",
-      "release": "5.15.0-185-generic",
-      "version": "#195-Ubuntu SMP Fri Jun 19 17:11:50 UTC 2026",
-      "machine": "x86_64",
-      "processor": "x86_64",
-      "accelerator": "NVIDIA GeForce RTX 4090"
-    },
-    "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 v3.11.2 (Git Hash 03c022d3ffdcee958cfacbe720048e725fdf644c)\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 92.0  (built against CUDA 13.2)\n  - Magma 2.6.1\n  - Build settings: BLAS_INFO=mkl, BUILD_TYPE=Release, COMMIT_SHA=7661cd9c6b841b62b7f411aa52ec51f05457263b, 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 -DLIBKINETO_NOROCTRACER -DLIBKINETO_NOXPUPTI=ON -DUSE_FBGEMM -DUSE_MSLK -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=2.12.0, 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"
-  }
-}

MetaMathQA/ptuning--llama-3.2-3B-vt50-MLP-lr1e-3.json

@@ -1,352 +0,0 @@
-{
-  "run_info": {
-    "created_at": "2026-07-01T06:31:13+00:00",
-    "total_time": 1213.1375317359925,
-    "experiment_name": "ptuning/llama-3.2-3B-vt50-MLP-lr1e-3",
-    "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": 0.001
-      },
-      "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": "P_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": 50,
-      "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": 14750776190,
-    "accelerator_memory_max": 22204645376,
-    "accelerator_memory_reserved_99th": 20428357632,
-    "train_time": 1036.8485383950465,
-    "file_size": 614520,
-    "num_trainable_params": 28474368,
-    "num_total_params": 3241224192,
-    "status": "success",
-    "metrics": [
-      {
-        "step": 250,
-        "valid accuracy": 0.3,
-        "train loss": 0.8939944224357605,
-        "train samples": 1000,
-        "train time": 35.82512437502737,
-        "eval time": 10.073844177997671,
-        "tokens / sec": 5909.791066840874,
-        "mem allocated avg": 7309444769.792,
-        "mem reserved avg": 14892866207.744,
-        "elapsed time": 64.39889488599147
-      },
-      {
-        "step": 500,
-        "valid accuracy": 0.28,
-        "train loss": 0.7657690613269806,
-        "train samples": 2000,
-        "train time": 35.75427145490539,
-        "eval time": 10.098102464020485,
-        "tokens / sec": 5817.346894127908,
-        "mem allocated avg": 7303010777.088,
-        "mem reserved avg": 14513097146.368,
-        "elapsed time": 112.63863867899636
-      },
-      {
-        "step": 750,
-        "valid accuracy": 0.34,
-        "train loss": 0.728571077823639,
-        "train samples": 3000,
-        "train time": 36.53094967507059,
-        "eval time": 9.911620947008487,
-        "tokens / sec": 5869.023441958622,
-        "mem allocated avg": 7313580834.816,
-        "mem reserved avg": 14636963332.096,
-        "elapsed time": 161.38983537000604
-      },
-      {
-        "step": 1000,
-        "valid accuracy": 0.4,
-        "train loss": 0.7045905925035477,
-        "train samples": 4000,
-        "train time": 35.944540872907965,
-        "eval time": 8.00756757799536,
-        "tokens / sec": 5796.040092336429,
-        "mem allocated avg": 7304926347.264,
-        "mem reserved avg": 14708165836.8,
-        "elapsed time": 207.77592266799184
-      },
-      {
-        "step": 1250,
-        "valid accuracy": 0.3,
-        "train loss": 0.7039546686410904,
-        "train samples": 5000,
-        "train time": 35.91604287101654,
-        "eval time": 7.098663012002362,
-        "tokens / sec": 5806.263255362288,
-        "mem allocated avg": 7305019461.632,
-        "mem reserved avg": 14845118251.008,
-        "elapsed time": 253.18471807800233
-      },
-      {
-        "step": 1500,
-        "valid accuracy": 0.38,
-        "train loss": 0.6957033377885818,
-        "train samples": 6000,
-        "train time": 36.25313519843621,
-        "eval time": 9.914249792986084,
-        "tokens / sec": 5774.148879929964,
-        "mem allocated avg": 7306861111.296,
-        "mem reserved avg": 14650443825.152,
-        "elapsed time": 301.8694867900049
-      },
-      {
-        "step": 1750,
-        "valid accuracy": 0.32,
-        "train loss": 0.6876964381933213,
-        "train samples": 7000,
-        "train time": 36.3216554413375,
-        "eval time": 6.474363713990897,
-        "tokens / sec": 5763.916799941175,
-        "mem allocated avg": 7307783305.216,
-        "mem reserved avg": 15022319206.4,
-        "elapsed time": 347.12181472199154
-      },
-      {
-        "step": 2000,
-        "valid accuracy": 0.34,
-        "train loss": 0.6909883782863617,
-        "train samples": 8000,
-        "train time": 36.12862586573465,
-        "eval time": 10.079137530992739,
-        "tokens / sec": 5748.793235919455,
-        "mem allocated avg": 7304383127.552,
-        "mem reserved avg": 14724288741.376,
-        "elapsed time": 395.86313636400155
-      },
-      {
-        "step": 2250,
-        "valid accuracy": 0.32,
-        "train loss": 0.683233140707016,
-        "train samples": 9000,
-        "train time": 37.00546802487224,
-        "eval time": 9.95705728197936,
-        "tokens / sec": 5808.546992447938,
-        "mem allocated avg": 7315459004.416,
-        "mem reserved avg": 14968799887.36,
-        "elapsed time": 445.25343624298694
-      },
-      {
-        "step": 2500,
-        "valid accuracy": 0.34,
-        "train loss": 0.6816288355588913,
-        "train samples": 10000,
-        "train time": 35.75455765525112,
-        "eval time": 10.050994627003092,
-        "tokens / sec": 5760.580287021129,
-        "mem allocated avg": 7301004111.872,
-        "mem reserved avg": 14506419814.4,
-        "elapsed time": 493.5193583490036
-      },
-      {
-        "step": 2750,
-        "valid accuracy": 0.36,
-        "train loss": 0.6739110689163208,
-        "train samples": 11000,
-        "train time": 36.65771040355321,
-        "eval time": 6.461153475975152,
-        "tokens / sec": 5779.984556249386,
-        "mem allocated avg": 7311758159.872,
-        "mem reserved avg": 14959303983.104,
-        "elapsed time": 539.1099112940137
-      },
-      {
-        "step": 3000,
-        "valid accuracy": 0.36,
-        "train loss": 0.666211401462555,
-        "train samples": 12000,
-        "train time": 36.22247461407096,
-        "eval time": 9.90093484299723,
-        "tokens / sec": 5762.4721177640495,
-        "mem allocated avg": 7306837633.024,
-        "mem reserved avg": 14876651028.48,
-        "elapsed time": 587.7428188080085
-      },
-      {
-        "step": 3250,
-        "valid accuracy": 0.4,
-        "train loss": 0.6753370553255081,
-        "train samples": 13000,
-        "train time": 36.267065661319066,
-        "eval time": 6.997407575021498,
-        "tokens / sec": 5815.22094920787,
-        "mem allocated avg": 7308324284.416,
-        "mem reserved avg": 14763698421.76,
-        "elapsed time": 633.3630509219947
-      },
-      {
-        "step": 3500,
-        "valid accuracy": 0.36,
-        "train loss": 0.6598643933534623,
-        "train samples": 14000,
-        "train time": 36.3332851199375,
-        "eval time": 8.388575849996414,
-        "tokens / sec": 5772.943440363501,
-        "mem allocated avg": 7307028889.6,
-        "mem reserved avg": 14673487331.328,
-        "elapsed time": 680.5796030820056
-      },
-      {
-        "step": 3750,
-        "valid accuracy": 0.36,
-        "train loss": 0.657373480796814,
-        "train samples": 15000,
-        "train time": 37.1786705829727,
-        "eval time": 5.97776805300964,
-        "tokens / sec": 5828.691467500908,
-        "mem allocated avg": 7318995214.336,
-        "mem reserved avg": 14927376941.056,
-        "elapsed time": 726.1502790729865
-      },
-      {
-        "step": 4000,
-        "valid accuracy": 0.42,
-        "train loss": 0.6716070612668991,
-        "train samples": 16000,
-        "train time": 35.672596014803275,
-        "eval time": 7.174363945989171,
-        "tokens / sec": 5729.131681787053,
-        "mem allocated avg": 7300470390.784,
-        "mem reserved avg": 14715841413.12,
-        "elapsed time": 771.4763409139996
-      },
-      {
-        "step": 4250,
-        "valid accuracy": 0.36,
-        "train loss": 0.654036686182022,
-        "train samples": 17000,
-        "train time": 36.42060951096937,
-        "eval time": 10.011166092008352,
-        "tokens / sec": 5804.103853240914,
-        "mem allocated avg": 7309944072.192,
-        "mem reserved avg": 14709793226.752,
-        "elapsed time": 820.2551248139935
-      },
-      {
-        "step": 4500,
-        "valid accuracy": 0.38,
-        "train loss": 0.6614194496870041,
-        "train samples": 18000,
-        "train time": 35.99739089809009,
-        "eval time": 9.344737695006188,
-        "tokens / sec": 5773.140630895729,
-        "mem allocated avg": 7303833925.632,
-        "mem reserved avg": 14722778791.936,
-        "elapsed time": 868.1181147210009
-      },
-      {
-        "step": 4750,
-        "valid accuracy": 0.42,
-        "train loss": 0.6527535218000412,
-        "train samples": 19000,
-        "train time": 36.4203205309459,
-        "eval time": 7.0326879719796125,
-        "tokens / sec": 5764.336967370109,
-        "mem allocated avg": 7307458248.704,
-        "mem reserved avg": 14665165832.192,
-        "elapsed time": 913.9405549239891
-      },
-      {
-        "step": 5000,
-        "valid accuracy": 0.48,
-        "train loss": 0.6592182725667953,
-        "train samples": 20000,
-        "train time": 35.92553637881065,
-        "eval time": 9.915976137999678,
-        "tokens / sec": 5797.547399260162,
-        "mem allocated avg": 7303994013.696,
-        "mem reserved avg": 14532944592.896,
-        "elapsed time": 962.1836511259899
-      },
-      {
-        "step": 5000,
-        "test accuracy": 0.36694465504169826,
-        "train loss": 0.6592182725667953,
-        "train samples": 20000,
-        "train total tokens": 4198051,
-        "forgetting": 4.0564727783203125
-      }
-    ]
-  },
-  "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": "5.12.1",
-      "transformers-commit-hash": null,
-      "peft-version": "0.19.2.dev0",
-      "peft-commit-hash": "02973be18fad9e2583027264b60233218f312ecd",
-      "datasets-version": "5.0.0",
-      "datasets-commit-hash": null,
-      "bitsandbytes-version": "0.49.2",
-      "bitsandbytes-commit-hash": null,
-      "torch-version": "2.12.0+cu130",
-      "torch-commit-hash": null
-    },
-    "system_info": {
-      "system": "Linux",
-      "release": "6.8.0-90-generic",
-      "version": "#91~22.04.1-Ubuntu SMP PREEMPT_DYNAMIC Thu Nov 20 15:20:45 UTC 2",
-      "machine": "x86_64",
-      "processor": "x86_64",
-      "accelerator": "NVIDIA GeForce RTX 4090"
-    },
-    "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 v3.11.2 (Git Hash 03c022d3ffdcee958cfacbe720048e725fdf644c)\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 92.0  (built against CUDA 13.2)\n  - Magma 2.6.1\n  - Build settings: BLAS_INFO=mkl, BUILD_TYPE=Release, COMMIT_SHA=7661cd9c6b841b62b7f411aa52ec51f05457263b, 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 -DLIBKINETO_NOROCTRACER -DLIBKINETO_NOXPUPTI=ON -DUSE_FBGEMM -DUSE_MSLK -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=2.12.0, 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"
-  }
-}

MetaMathQA/ptuning--llama-3.2-3B-vt50-MLP-lr5e-3.json

@@ -1,352 +0,0 @@
-{
-  "run_info": {
-    "created_at": "2026-07-01T07:00:52+00:00",
-    "total_time": 1217.8010101249965,
-    "experiment_name": "ptuning/llama-3.2-3B-vt50-MLP-lr5e-3",
-    "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": "P_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": 50,
-      "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": 14753152263,
-    "accelerator_memory_max": 22204645376,
-    "accelerator_memory_reserved_99th": 20428357632,
-    "train_time": 1029.7820923980034,
-    "file_size": 614520,
-    "num_trainable_params": 28474368,
-    "num_total_params": 3241224192,
-    "status": "success",
-    "metrics": [
-      {
-        "step": 250,
-        "valid accuracy": 0.12,
-        "train loss": 0.8555094628334046,
-        "train samples": 1000,
-        "train time": 35.85137379690423,
-        "eval time": 10.199397451011464,
-        "tokens / sec": 5905.464075083281,
-        "mem allocated avg": 7309444769.792,
-        "mem reserved avg": 14892866207.744,
-        "elapsed time": 64.8755607100029
-      },
-      {
-        "step": 500,
-        "valid accuracy": 0.32,
-        "train loss": 0.7614331521987915,
-        "train samples": 2000,
-        "train time": 35.56839666242013,
-        "eval time": 9.910280911979498,
-        "tokens / sec": 5847.747425167398,
-        "mem allocated avg": 7303010777.088,
-        "mem reserved avg": 14513097146.368,
-        "elapsed time": 112.7274315149989
-      },
-      {
-        "step": 750,
-        "valid accuracy": 0.3,
-        "train loss": 0.7213665727376938,
-        "train samples": 3000,
-        "train time": 36.56926804804243,
-        "eval time": 9.971691955986898,
-        "tokens / sec": 5862.873703633699,
-        "mem allocated avg": 7313580834.816,
-        "mem reserved avg": 14636963332.096,
-        "elapsed time": 161.59096823597793
-      },
-      {
-        "step": 1000,
-        "valid accuracy": 0.3,
-        "train loss": 0.7033745880126954,
-        "train samples": 4000,
-        "train time": 35.96135858510388,
-        "eval time": 9.92101417799131,
-        "tokens / sec": 5793.329512481158,
-        "mem allocated avg": 7304926347.264,
-        "mem reserved avg": 14708165836.8,
-        "elapsed time": 209.91149750500335
-      },
-      {
-        "step": 1250,
-        "valid accuracy": 0.36,
-        "train loss": 0.7029182990789413,
-        "train samples": 5000,
-        "train time": 35.83780645477236,
-        "eval time": 10.019560440996429,
-        "tokens / sec": 5818.93873061614,
-        "mem allocated avg": 7305019461.632,
-        "mem reserved avg": 14845118251.008,
-        "elapsed time": 258.15060135300155
-      },
-      {
-        "step": 1500,
-        "valid accuracy": 0.4,
-        "train loss": 0.6964308704137803,
-        "train samples": 6000,
-        "train time": 36.21187603412545,
-        "eval time": 10.001915690983878,
-        "tokens / sec": 5780.7278419579825,
-        "mem allocated avg": 7306861111.296,
-        "mem reserved avg": 14650443825.152,
-        "elapsed time": 306.8930205789802
-      },
-      {
-        "step": 1750,
-        "valid accuracy": 0.32,
-        "train loss": 0.6913636869192124,
-        "train samples": 7000,
-        "train time": 36.299369902320905,
-        "eval time": 9.905490600998746,
-        "tokens / sec": 5767.455483755223,
-        "mem allocated avg": 7307783305.216,
-        "mem reserved avg": 15022319206.4,
-        "elapsed time": 355.55441494999104
-      },
-      {
-        "step": 2000,
-        "valid accuracy": 0.28,
-        "train loss": 0.6942930264472962,
-        "train samples": 8000,
-        "train time": 35.98092826092034,
-        "eval time": 9.915651136019733,
-        "tokens / sec": 5772.391376172001,
-        "mem allocated avg": 7304320778.24,
-        "mem reserved avg": 14774016409.6,
-        "elapsed time": 403.95992209698306
-      },
-      {
-        "step": 2250,
-        "valid accuracy": 0.36,
-        "train loss": 0.6876053595542908,
-        "train samples": 9000,
-        "train time": 37.06889635126572,
-        "eval time": 10.021580134984106,
-        "tokens / sec": 5798.608028767509,
-        "mem allocated avg": 7315318763.52,
-        "mem reserved avg": 14955420057.6,
-        "elapsed time": 453.478056174994
-      },
-      {
-        "step": 2500,
-        "valid accuracy": 0.38,
-        "train loss": 0.6850743471384049,
-        "train samples": 10000,
-        "train time": 35.693450052058324,
-        "eval time": 9.98401257902151,
-        "tokens / sec": 5770.442467724482,
-        "mem allocated avg": 7301149323.264,
-        "mem reserved avg": 14488627576.832,
-        "elapsed time": 501.58775900499313
-      },
-      {
-        "step": 2750,
-        "valid accuracy": 0.36,
-        "train loss": 0.6785946444272996,
-        "train samples": 11000,
-        "train time": 36.630218399630394,
-        "eval time": 10.036885927984258,
-        "tokens / sec": 5784.322596398659,
-        "mem allocated avg": 7311674300.416,
-        "mem reserved avg": 14942551932.928,
-        "elapsed time": 550.7442475239804
-      },
-      {
-        "step": 3000,
-        "valid accuracy": 0.3,
-        "train loss": 0.6711985548734665,
-        "train samples": 12000,
-        "train time": 36.30712951908936,
-        "eval time": 9.915486376004992,
-        "tokens / sec": 5749.036147026016,
-        "mem allocated avg": 7306745915.392,
-        "mem reserved avg": 14922368942.08,
-        "elapsed time": 599.502809875994
-      },
-      {
-        "step": 3250,
-        "valid accuracy": 0.42,
-        "train loss": 0.6797376515865325,
-        "train samples": 13000,
-        "train time": 36.31324738584226,
-        "eval time": 9.915928879985586,
-        "tokens / sec": 5807.8253855706025,
-        "mem allocated avg": 7308324284.416,
-        "mem reserved avg": 14763698421.76,
-        "elapsed time": 648.1008756539959
-      },
-      {
-        "step": 3500,
-        "valid accuracy": 0.32,
-        "train loss": 0.6659102200269699,
-        "train samples": 14000,
-        "train time": 36.316436834138585,
-        "eval time": 9.99347340699751,
-        "tokens / sec": 5775.621682213836,
-        "mem allocated avg": 7307028889.6,
-        "mem reserved avg": 14673487331.328,
-        "elapsed time": 696.9008167689899
-      },
-      {
-        "step": 3750,
-        "valid accuracy": 0.4,
-        "train loss": 0.6623009448051452,
-        "train samples": 15000,
-        "train time": 37.26567734207492,
-        "eval time": 6.6307652829855215,
-        "tokens / sec": 5815.0828176503,
-        "mem allocated avg": 7318995214.336,
-        "mem reserved avg": 14927376941.056,
-        "elapsed time": 743.2372308159829
-      },
-      {
-        "step": 4000,
-        "valid accuracy": 0.36,
-        "train loss": 0.6768936680555344,
-        "train samples": 16000,
-        "train time": 35.644785020820564,
-        "eval time": 6.08822312799748,
-        "tokens / sec": 5733.601700238147,
-        "mem allocated avg": 7300470390.784,
-        "mem reserved avg": 14715841413.12,
-        "elapsed time": 787.420050811983
-      },
-      {
-        "step": 4250,
-        "valid accuracy": 0.42,
-        "train loss": 0.6587289227247238,
-        "train samples": 17000,
-        "train time": 36.425392148201354,
-        "eval time": 9.920218591025332,
-        "tokens / sec": 5803.341777075094,
-        "mem allocated avg": 7309944072.192,
-        "mem reserved avg": 14709793226.752,
-        "elapsed time": 836.1627081699844
-      },
-      {
-        "step": 4500,
-        "valid accuracy": 0.42,
-        "train loss": 0.665844162940979,
-        "train samples": 18000,
-        "train time": 36.031896430038614,
-        "eval time": 8.943657095020171,
-        "tokens / sec": 5767.612049049656,
-        "mem allocated avg": 7303833925.632,
-        "mem reserved avg": 14722778791.936,
-        "elapsed time": 883.662610319996
-      },
-      {
-        "step": 4750,
-        "valid accuracy": 0.42,
-        "train loss": 0.6578277311325074,
-        "train samples": 19000,
-        "train time": 36.29169078628183,
-        "eval time": 6.231622669001808,
-        "tokens / sec": 5784.767682396226,
-        "mem allocated avg": 7307458248.704,
-        "mem reserved avg": 14665165832.192,
-        "elapsed time": 928.5394251639955
-      },
-      {
-        "step": 5000,
-        "valid accuracy": 0.4,
-        "train loss": 0.6639133603572845,
-        "train samples": 20000,
-        "train time": 36.014907898090314,
-        "eval time": 7.127454289002344,
-        "tokens / sec": 5783.160700823118,
-        "mem allocated avg": 7303994013.696,
-        "mem reserved avg": 14532944592.896,
-        "elapsed time": 974.0815745609871
-      },
-      {
-        "step": 5000,
-        "test accuracy": 0.354814253222138,
-        "train loss": 0.6639133603572845,
-        "train samples": 20000,
-        "train total tokens": 4198051,
-        "forgetting": 3.9880266189575195
-      }
-    ]
-  },
-  "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": "5.12.1",
-      "transformers-commit-hash": null,
-      "peft-version": "0.19.2.dev0",
-      "peft-commit-hash": "e390c22b84bf20c718b3605ebaf141cab8202642",
-      "datasets-version": "5.0.0",
-      "datasets-commit-hash": null,
-      "bitsandbytes-version": "0.49.2",
-      "bitsandbytes-commit-hash": null,
-      "torch-version": "2.12.0+cu130",
-      "torch-commit-hash": null
-    },
-    "system_info": {
-      "system": "Linux",
-      "release": "6.8.0-90-generic",
-      "version": "#91~22.04.1-Ubuntu SMP PREEMPT_DYNAMIC Thu Nov 20 15:20:45 UTC 2",
-      "machine": "x86_64",
-      "processor": "x86_64",
-      "accelerator": "NVIDIA GeForce RTX 4090"
-    },
-    "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 v3.11.2 (Git Hash 03c022d3ffdcee958cfacbe720048e725fdf644c)\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 92.0  (built against CUDA 13.2)\n  - Magma 2.6.1\n  - Build settings: BLAS_INFO=mkl, BUILD_TYPE=Release, COMMIT_SHA=7661cd9c6b841b62b7f411aa52ec51f05457263b, 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 -DLIBKINETO_NOROCTRACER -DLIBKINETO_NOXPUPTI=ON -DUSE_FBGEMM -DUSE_MSLK -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=2.12.0, 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"
-  }
-}

MetaMathQA/ptuning--llama-3.2-3B-vt50-MLP-lr5e-5.json

@@ -1,352 +0,0 @@
-{
-  "run_info": {
-    "created_at": "2026-07-01T05:54:56+00:00",
-    "total_time": 1325.5959430259827,
-    "experiment_name": "ptuning/llama-3.2-3B-vt50-MLP-lr5e-5",
-    "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": 5e-05
-      },
-      "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": "P_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": 50,
-      "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": 14753152263,
-    "accelerator_memory_max": 22204645376,
-    "accelerator_memory_reserved_99th": 20428357632,
-    "train_time": 1113.6049115190108,
-    "file_size": 614520,
-    "num_trainable_params": 28474368,
-    "num_total_params": 3241224192,
-    "status": "success",
-    "metrics": [
-      {
-        "step": 250,
-        "valid accuracy": 0.08,
-        "train loss": 1.0313549892902374,
-        "train samples": 1000,
-        "train time": 35.723904531943845,
-        "eval time": 10.031031877006171,
-        "tokens / sec": 5926.53582451167,
-        "mem allocated avg": 7309444769.792,
-        "mem reserved avg": 14892866207.744,
-        "elapsed time": 115.88588901300682
-      },
-      {
-        "step": 500,
-        "valid accuracy": 0.02,
-        "train loss": 0.8095087568759918,
-        "train samples": 2000,
-        "train time": 35.544265604636166,
-        "eval time": 9.914330987987341,
-        "tokens / sec": 5851.717470085258,
-        "mem allocated avg": 7303010777.088,
-        "mem reserved avg": 14513097146.368,
-        "elapsed time": 163.7232813849987
-      },
-      {
-        "step": 750,
-        "valid accuracy": 0.1,
-        "train loss": 0.7724783256053924,
-        "train samples": 3000,
-        "train time": 36.6061558209185,
-        "eval time": 9.900999550998677,
-        "tokens / sec": 5856.965725897966,
-        "mem allocated avg": 7313580834.816,
-        "mem reserved avg": 14636963332.096,
-        "elapsed time": 212.54644100999576
-      },
-      {
-        "step": 1000,
-        "valid accuracy": 0.0,
-        "train loss": 0.7462414875030517,
-        "train samples": 4000,
-        "train time": 35.97310379197006,
-        "eval time": 10.092258746008156,
-        "tokens / sec": 5791.437992250891,
-        "mem allocated avg": 7304926347.264,
-        "mem reserved avg": 14708165836.8,
-        "elapsed time": 261.0330090589996
-      },
-      {
-        "step": 1250,
-        "valid accuracy": 0.12,
-        "train loss": 0.7396039123535156,
-        "train samples": 5000,
-        "train time": 35.96643319912255,
-        "eval time": 9.919816087989602,
-        "tokens / sec": 5798.1284617649435,
-        "mem allocated avg": 7305019461.632,
-        "mem reserved avg": 14845118251.008,
-        "elapsed time": 309.33319281201693
-      },
-      {
-        "step": 1500,
-        "valid accuracy": 0.1,
-        "train loss": 0.7258432639837264,
-        "train samples": 6000,
-        "train time": 36.29043613522663,
-        "eval time": 10.010930726013612,
-        "tokens / sec": 5768.213950914888,
-        "mem allocated avg": 7306861111.296,
-        "mem reserved avg": 14650443825.152,
-        "elapsed time": 358.1597148909932
-      },
-      {
-        "step": 1750,
-        "valid accuracy": 0.18,
-        "train loss": 0.7163837015628814,
-        "train samples": 7000,
-        "train time": 36.33962430100655,
-        "eval time": 9.916795682976954,
-        "tokens / sec": 5761.066715106385,
-        "mem allocated avg": 7307783305.216,
-        "mem reserved avg": 15022319206.4,
-        "elapsed time": 406.85831744299503
-      },
-      {
-        "step": 2000,
-        "valid accuracy": 0.3,
-        "train loss": 0.7135137581825256,
-        "train samples": 8000,
-        "train time": 36.04737422184553,
-        "eval time": 9.908883731986862,
-        "tokens / sec": 5761.751153406661,
-        "mem allocated avg": 7304320778.24,
-        "mem reserved avg": 14774016409.6,
-        "elapsed time": 455.3342518869904
-      },
-      {
-        "step": 2250,
-        "valid accuracy": 0.28,
-        "train loss": 0.7062074838876724,
-        "train samples": 9000,
-        "train time": 36.97540899395244,
-        "eval time": 10.030416852008784,
-        "tokens / sec": 5813.269030645641,
-        "mem allocated avg": 7315318763.52,
-        "mem reserved avg": 14955420057.6,
-        "elapsed time": 504.7872484539985
-      },
-      {
-        "step": 2500,
-        "valid accuracy": 0.28,
-        "train loss": 0.7037104018926621,
-        "train samples": 10000,
-        "train time": 35.75819702979061,
-        "eval time": 9.925909476005472,
-        "tokens / sec": 5759.993990424246,
-        "mem allocated avg": 7301149323.264,
-        "mem reserved avg": 14488627576.832,
-        "elapsed time": 552.9260756930162
-      },
-      {
-        "step": 2750,
-        "valid accuracy": 0.1,
-        "train loss": 0.695045065164566,
-        "train samples": 11000,
-        "train time": 36.76619625516469,
-        "eval time": 9.944942692003679,
-        "tokens / sec": 5762.9295815510495,
-        "mem allocated avg": 7311674300.416,
-        "mem reserved avg": 14942551932.928,
-        "elapsed time": 602.1679049750092
-      },
-      {
-        "step": 3000,
-        "valid accuracy": 0.06,
-        "train loss": 0.6885745447874069,
-        "train samples": 12000,
-        "train time": 36.17643003768171,
-        "eval time": 9.89490892199683,
-        "tokens / sec": 5769.806467431525,
-        "mem allocated avg": 7306745915.392,
-        "mem reserved avg": 14922368942.08,
-        "elapsed time": 650.7650499030133
-      },
-      {
-        "step": 3250,
-        "valid accuracy": 0.24,
-        "train loss": 0.6949721579551696,
-        "train samples": 13000,
-        "train time": 36.288303180306684,
-        "eval time": 7.917344357003458,
-        "tokens / sec": 5811.817624871861,
-        "mem allocated avg": 7308324284.416,
-        "mem reserved avg": 14763698421.76,
-        "elapsed time": 697.3301032750169
-      },
-      {
-        "step": 3500,
-        "valid accuracy": 0.18,
-        "train loss": 0.6815348329544068,
-        "train samples": 14000,
-        "train time": 36.27739872696111,
-        "eval time": 9.895855661015958,
-        "tokens / sec": 5781.836828452511,
-        "mem allocated avg": 7307028889.6,
-        "mem reserved avg": 14673487331.328,
-        "elapsed time": 745.9850295080105
-      },
-      {
-        "step": 3750,
-        "valid accuracy": 0.2,
-        "train loss": 0.6767042907476425,
-        "train samples": 15000,
-        "train time": 37.175234996073414,
-        "eval time": 9.897059137001634,
-        "tokens / sec": 5829.230131911445,
-        "mem allocated avg": 7318995214.336,
-        "mem reserved avg": 14927376941.056,
-        "elapsed time": 795.4740971440042
-      },
-      {
-        "step": 4000,
-        "valid accuracy": 0.3,
-        "train loss": 0.6919913440942764,
-        "train samples": 16000,
-        "train time": 35.65538451215252,
-        "eval time": 10.031130649003899,
-        "tokens / sec": 5731.897237858787,
-        "mem allocated avg": 7300470390.784,
-        "mem reserved avg": 14715841413.12,
-        "elapsed time": 843.6182897959952
-      },
-      {
-        "step": 4250,
-        "valid accuracy": 0.38,
-        "train loss": 0.6734701672792435,
-        "train samples": 17000,
-        "train time": 36.324834284285316,
-        "eval time": 9.906978653016267,
-        "tokens / sec": 5819.407140184811,
-        "mem allocated avg": 7309944072.192,
-        "mem reserved avg": 14709793226.752,
-        "elapsed time": 892.1824630350166
-      },
-      {
-        "step": 4500,
-        "valid accuracy": 0.22,
-        "train loss": 0.6816293342113495,
-        "train samples": 18000,
-        "train time": 36.00603464478627,
-        "eval time": 9.91252873200574,
-        "tokens / sec": 5771.75470862611,
-        "mem allocated avg": 7303833925.632,
-        "mem reserved avg": 14722778791.936,
-        "elapsed time": 940.6340228409972
-      },
-      {
-        "step": 4750,
-        "valid accuracy": 0.32,
-        "train loss": 0.6723749524354935,
-        "train samples": 19000,
-        "train time": 36.34486310466309,
-        "eval time": 9.924655996990623,
-        "tokens / sec": 5776.304601710402,
-        "mem allocated avg": 7307458248.704,
-        "mem reserved avg": 14665165832.192,
-        "elapsed time": 989.2675137660117
-      },
-      {
-        "step": 5000,
-        "valid accuracy": 0.38,
-        "train loss": 0.678697298526764,
-        "train samples": 20000,
-        "train time": 35.95660388888791,
-        "eval time": 10.044164422986796,
-        "tokens / sec": 5792.53815637375,
-        "mem allocated avg": 7303994013.696,
-        "mem reserved avg": 14532944592.896,
-        "elapsed time": 1037.6695654990035
-      },
-      {
-        "step": 5000,
-        "test accuracy": 0.3055344958301744,
-        "train loss": 0.678697298526764,
-        "train samples": 20000,
-        "train total tokens": 4198051,
-        "forgetting": 3.8238143920898438
-      }
-    ]
-  },
-  "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": "5.12.1",
-      "transformers-commit-hash": null,
-      "peft-version": "0.19.2.dev0",
-      "peft-commit-hash": "9a7f8da7db81de998ad5d4802ca8852d70cead01",
-      "datasets-version": "5.0.0",
-      "datasets-commit-hash": null,
-      "bitsandbytes-version": "0.49.2",
-      "bitsandbytes-commit-hash": null,
-      "torch-version": "2.12.0+cu130",
-      "torch-commit-hash": null
-    },
-    "system_info": {
-      "system": "Linux",
-      "release": "6.8.0-90-generic",
-      "version": "#91~22.04.1-Ubuntu SMP PREEMPT_DYNAMIC Thu Nov 20 15:20:45 UTC 2",
-      "machine": "x86_64",
-      "processor": "x86_64",
-      "accelerator": "NVIDIA GeForce RTX 4090"
-    },
-    "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 v3.11.2 (Git Hash 03c022d3ffdcee958cfacbe720048e725fdf644c)\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 92.0  (built against CUDA 13.2)\n  - Magma 2.6.1\n  - Build settings: BLAS_INFO=mkl, BUILD_TYPE=Release, COMMIT_SHA=7661cd9c6b841b62b7f411aa52ec51f05457263b, 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 -DLIBKINETO_NOROCTRACER -DLIBKINETO_NOXPUPTI=ON -DUSE_FBGEMM -DUSE_MSLK -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=2.12.0, 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"
-  }
-}

MetaMathQA/ptuning--llama-3.2-3B-vt50-MLP.json

@@ -1,353 +0,0 @@
-{
-  "run_info": {
-    "created_at": "2026-06-25T14:16:41+00:00",
-    "total_time": 1187.580871684011,
-    "experiment_name": "ptuning/llama-3.2-3B-vt50-MLP",
-    "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": 0.0001,
-        "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,
-      "init_kv_cache_prefix": null
-    },
-    "peft_config": {
-      "task_type": "CAUSAL_LM",
-      "peft_type": "P_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": 50,
-      "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": 14750776190,
-    "accelerator_memory_max": 22204645376,
-    "accelerator_memory_reserved_99th": 20428357632,
-    "train_time": 1010.8743848598097,
-    "file_size": 614520,
-    "num_trainable_params": 28474368,
-    "num_total_params": 3241224192,
-    "status": "success",
-    "metrics": [
-      {
-        "step": 250,
-        "valid accuracy": 0.14,
-        "train loss": 0.9848673729896545,
-        "train samples": 1000,
-        "train time": 34.98505723965354,
-        "eval time": 9.99676610995084,
-        "tokens / sec": 6051.697973500204,
-        "mem allocated avg": 7309444769.792,
-        "mem reserved avg": 14892866207.744,
-        "elapsed time": 62.244462973903865
-      },
-      {
-        "step": 500,
-        "valid accuracy": 0.22,
-        "train loss": 0.799057240486145,
-        "train samples": 2000,
-        "train time": 34.88942794734612,
-        "eval time": 9.849827846046537,
-        "tokens / sec": 5961.5480171786885,
-        "mem allocated avg": 7303010777.088,
-        "mem reserved avg": 14513097146.368,
-        "elapsed time": 109.26378811709583
-      },
-      {
-        "step": 750,
-        "valid accuracy": 0.36,
-        "train loss": 0.758115061879158,
-        "train samples": 3000,
-        "train time": 35.77429200033657,
-        "eval time": 9.828197933966294,
-        "tokens / sec": 5993.158438970165,
-        "mem allocated avg": 7313580834.816,
-        "mem reserved avg": 14636963332.096,
-        "elapsed time": 157.5186764819082
-      },
-      {
-        "step": 1000,
-        "valid accuracy": 0.34,
-        "train loss": 0.7289376347064972,
-        "train samples": 4000,
-        "train time": 35.046438679564744,
-        "eval time": 9.845785265089944,
-        "tokens / sec": 5944.569772262732,
-        "mem allocated avg": 7304926347.264,
-        "mem reserved avg": 14708165836.8,
-        "elapsed time": 204.7628760500811
-      },
-      {
-        "step": 1250,
-        "valid accuracy": 0.26,
-        "train loss": 0.7213420255184173,
-        "train samples": 5000,
-        "train time": 35.13815849763341,
-        "eval time": 9.83495028782636,
-        "tokens / sec": 5934.801620695212,
-        "mem allocated avg": 7305019461.632,
-        "mem reserved avg": 14845118251.008,
-        "elapsed time": 252.16518334206194
-      },
-      {
-        "step": 1500,
-        "valid accuracy": 0.28,
-        "train loss": 0.7093797615766525,
-        "train samples": 6000,
-        "train time": 35.23495648638345,
-        "eval time": 8.178562826942652,
-        "tokens / sec": 5941.003505450502,
-        "mem allocated avg": 7306861111.296,
-        "mem reserved avg": 14650443825.152,
-        "elapsed time": 298.704996796092
-      },
-      {
-        "step": 1750,
-        "valid accuracy": 0.42,
-        "train loss": 0.7009247547388077,
-        "train samples": 7000,
-        "train time": 35.448243242455646,
-        "eval time": 6.152927630813792,
-        "tokens / sec": 5905.93442298601,
-        "mem allocated avg": 7307783305.216,
-        "mem reserved avg": 15022319206.4,
-        "elapsed time": 342.65977702406235
-      },
-      {
-        "step": 2000,
-        "valid accuracy": 0.36,
-        "train loss": 0.7022378945350647,
-        "train samples": 8000,
-        "train time": 35.023517513182014,
-        "eval time": 9.823164875851944,
-        "tokens / sec": 5930.186764417029,
-        "mem allocated avg": 7304383127.552,
-        "mem reserved avg": 14724288741.376,
-        "elapsed time": 390.2323130860459
-      },
-      {
-        "step": 2250,
-        "valid accuracy": 0.4,
-        "train loss": 0.6948372975587845,
-        "train samples": 9000,
-        "train time": 35.97670590644702,
-        "eval time": 9.842984584858641,
-        "tokens / sec": 5974.643719715355,
-        "mem allocated avg": 7315459004.416,
-        "mem reserved avg": 14968799887.36,
-        "elapsed time": 438.72397272707894
-      },
-      {
-        "step": 2500,
-        "valid accuracy": 0.32,
-        "train loss": 0.6927290350198746,
-        "train samples": 10000,
-        "train time": 35.08757257997058,
-        "eval time": 7.987744936021045,
-        "tokens / sec": 5870.084045585256,
-        "mem allocated avg": 7301004111.872,
-        "mem reserved avg": 14506419814.4,
-        "elapsed time": 484.4672777580563
-      },
-      {
-        "step": 2750,
-        "valid accuracy": 0.36,
-        "train loss": 0.6847438111305236,
-        "train samples": 11000,
-        "train time": 35.630042768549174,
-        "eval time": 6.188074267935008,
-        "tokens / sec": 5946.6950791041,
-        "mem allocated avg": 7311758159.872,
-        "mem reserved avg": 14959303983.104,
-        "elapsed time": 528.6565973460674
-      },
-      {
-        "step": 3000,
-        "valid accuracy": 0.34,
-        "train loss": 0.6785378177165985,
-        "train samples": 12000,
-        "train time": 35.22064670943655,
-        "eval time": 8.96801636298187,
-        "tokens / sec": 5926.381810135116,
-        "mem allocated avg": 7306837633.024,
-        "mem reserved avg": 14876651028.48,
-        "elapsed time": 575.5715269569773
-      },
-      {
-        "step": 3250,
-        "valid accuracy": 0.32,
-        "train loss": 0.6858672415018081,
-        "train samples": 13000,
-        "train time": 35.62791796633974,
-        "eval time": 7.249630860053003,
-        "tokens / sec": 5919.5432132535325,
-        "mem allocated avg": 7308324284.416,
-        "mem reserved avg": 14763698421.76,
-        "elapsed time": 621.039221069077
-      },
-      {
-        "step": 3500,
-        "valid accuracy": 0.48,
-        "train loss": 0.671637071609497,
-        "train samples": 14000,
-        "train time": 35.37562209949829,
-        "eval time": 6.312963296892121,
-        "tokens / sec": 5929.2243514489255,
-        "mem allocated avg": 7307028889.6,
-        "mem reserved avg": 14673487331.328,
-        "elapsed time": 665.5405941770878
-      },
-      {
-        "step": 3750,
-        "valid accuracy": 0.32,
-        "train loss": 0.667889571070671,
-        "train samples": 15000,
-        "train time": 36.57185547542758,
-        "eval time": 9.82020484493114,
-        "tokens / sec": 5925.4034880894,
-        "mem allocated avg": 7318995214.336,
-        "mem reserved avg": 14927376941.056,
-        "elapsed time": 714.2830616689753
-      },
-      {
-        "step": 4000,
-        "valid accuracy": 0.44,
-        "train loss": 0.6836470769643783,
-        "train samples": 16000,
-        "train time": 34.66988317412324,
-        "eval time": 7.146935621043667,
-        "tokens / sec": 5894.828055046319,
-        "mem allocated avg": 7300470390.784,
-        "mem reserved avg": 14715841413.12,
-        "elapsed time": 758.958012880059
-      },
-      {
-        "step": 4250,
-        "valid accuracy": 0.44,
-        "train loss": 0.6656210062503815,
-        "train samples": 17000,
-        "train time": 35.417583748465404,
-        "eval time": 9.825491457013413,
-        "tokens / sec": 5968.476040073151,
-        "mem allocated avg": 7309944072.192,
-        "mem reserved avg": 14709793226.752,
-        "elapsed time": 807.20357554988
-      },
-      {
-        "step": 4500,
-        "valid accuracy": 0.36,
-        "train loss": 0.6724705603122711,
-        "train samples": 18000,
-        "train time": 35.21988031337969,
-        "eval time": 6.9338154550641775,
-        "tokens / sec": 5900.587910886567,
-        "mem allocated avg": 7303833925.632,
-        "mem reserved avg": 14722778791.936,
-        "elapsed time": 851.8069666360971
-      },
-      {
-        "step": 4750,
-        "valid accuracy": 0.44,
-        "train loss": 0.6646332576274872,
-        "train samples": 19000,
-        "train time": 35.34395982767455,
-        "eval time": 9.840549977030605,
-        "tokens / sec": 5939.883392341805,
-        "mem allocated avg": 7307458248.704,
-        "mem reserved avg": 14665165832.192,
-        "elapsed time": 899.864475436043
-      },
-      {
-        "step": 5000,
-        "valid accuracy": 0.34,
-        "train loss": 0.6710220981836319,
-        "train samples": 20000,
-        "train time": 34.99154016003013,
-        "eval time": 9.884989776881412,
-        "tokens / sec": 5952.295870586243,
-        "mem allocated avg": 7303994013.696,
-        "mem reserved avg": 14532944592.896,
-        "elapsed time": 947.2323914049193
-      },
-      {
-        "step": 5000,
-        "test accuracy": 0.3419257012888552,
-        "train loss": 0.6710220981836319,
-        "train samples": 20000,
-        "train total tokens": 4198051,
-        "forgetting": 3.807856559753418
-      }
-    ]
-  },
-  "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": "5.12.1",
-      "transformers-commit-hash": null,
-      "peft-version": "0.19.2.dev0",
-      "peft-commit-hash": "ed69ccdb3a42fff849f01431eed1388177443388",
-      "datasets-version": "5.0.0",
-      "datasets-commit-hash": null,
-      "bitsandbytes-version": "0.49.2",
-      "bitsandbytes-commit-hash": null,
-      "torch-version": "2.12.0+cu130",
-      "torch-commit-hash": null
-    },
-    "system_info": {
-      "system": "Linux",
-      "release": "6.8.0-90-generic",
-      "version": "#91~22.04.1-Ubuntu SMP PREEMPT_DYNAMIC Thu Nov 20 15:20:45 UTC 2",
-      "machine": "x86_64",
-      "processor": "x86_64",
-      "accelerator": "NVIDIA GeForce RTX 4090"
-    },
-    "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 v3.11.2 (Git Hash 03c022d3ffdcee958cfacbe720048e725fdf644c)\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: AVX512\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 92.0  (built against CUDA 13.2)\n  - Magma 2.6.1\n  - Build settings: BLAS_INFO=mkl, BUILD_TYPE=Release, COMMIT_SHA=7661cd9c6b841b62b7f411aa52ec51f05457263b, 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 -DLIBKINETO_NOROCTRACER -DLIBKINETO_NOXPUPTI=ON -DUSE_FBGEMM -DUSE_MSLK -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=2.12.0, 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"
-  }
-}

imagegen/vera--flux2-klein-rank16.json

@@ -1,20 +0,0 @@
-{
-  "peft_method": "vera",
-  "base_model": "flux2-klein",
-  "variant": "rank16",
-  "rank": 16,
-  "batch_size": 2,
-  "d_initial": 0.1,
-  "vera_dropout": 0.0,
-  "lr": 1e-3,
-  "test_dino_similarity": 0.6504,
-  "test_drift": 0.2368,
-  "comparison_baseline": {
-    "config": "vera/flux2-klein-default (r=32)",
-    "test_dino_similarity": 0.6627,
-    "test_drift": 0.231
-  },
-  "conclusion": "No improvement over default rank 32. Similarity between rank64 and baseline, drift worse than both.",
-  "note": "Full auto-generated metadata unavailable due to a get_git_hash failure on Kaggle (transformers installed to dist-packages, not a git checkout). Test metrics recovered manually from run logs. Also observed anomalously slow training (~36-40s/it vs normal pace) on Kaggle T4, reproduced across two separate runs; suspected Kaggle-specific infra issue, not confirmed to be rank-related (Benjamin could not reproduce the slowdown locally). Related fix: https://github.com/huggingface/peft/pull/3395",
-  "environment": "Kaggle, single T4 GPU (16GB), PEFT branch pr-3395 (OOM fix applied)"
-}

imagegen/vera--flux2-klein-rank64.json

@@ -1,21 +0,0 @@
-{
-  "peft_method": "vera",
-  "base_model": "flux2-klein",
-  "variant": "rank64",
-  "rank": 64,
-  "batch_size": 2,
-  "d_initial": 0.1,
-  "vera_dropout": 0.0,
-  "lr": 1e-3,
-  "test_dino_similarity": 0.6469,
-  "test_drift": 0.2272,
-  "num_trainable_params": 896000,
-  "comparison_baseline": {
-    "config": "vera/flux2-klein-default (r=32)",
-    "test_dino_similarity": 0.6627,
-    "test_drift": 0.231
-  },
-  "conclusion": "No improvement over default rank 32. Similarity slightly lower, drift slightly better, but not a clear win.",
-  "note": "Full auto-generated metadata (package/git commit info) unavailable due to a get_git_hash failure on Kaggle (transformers installed to dist-packages, not a git checkout). Test metrics recovered manually from run logs. Related fix: https://github.com/huggingface/peft/pull/3395",
-  "environment": "Kaggle, single T4 GPU (16GB), PEFT branch pr-3395 (OOM fix applied)"
-}