cs-552-2026-the-transformers/safety_model
1
0
Fork 0
Code Issues Pull Requests Actions Projects Releases Wiki Activity

初始化项目，由ModelHub XC社区提供模型

Browse Source 
Model: cs-552-2026-the-transformers/safety_model
Source: Original Platform
This commit is contained in:
ModelHub XC
2026-06-23 12:32:19 +08:00
commit ba0c6eecaf
175 changed files with 19015 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

404
checkpoint-378/trainer_state.json Normal file
View File

@@ -0,0 +1,404 @@
{
"best_global_step": null,
"best_metric": null,
"best_model_checkpoint": null,
"epoch": 2.0,
"eval_steps": 500,
"global_step": 378,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"entropy": 0.9081946484744549,
"epoch": 0.053015241882041084,
"grad_norm": 51.5,
"learning_rate": 3e-06,
"loss": 2.320839500427246,
"mean_token_accuracy": 0.5708254482597113,
"num_tokens": 30761.0,
"step": 10
},
{
"entropy": 1.032745758444071,
"epoch": 0.10603048376408217,
"grad_norm": 21.125,
"learning_rate": 6.333333333333333e-06,
"loss": 1.9631355285644532,
"mean_token_accuracy": 0.5809337079524994,
"num_tokens": 61277.0,
"step": 20
},
{
"entropy": 1.2370933540165425,
"epoch": 0.15904572564612326,
"grad_norm": 11.0625,
"learning_rate": 9.666666666666667e-06,
"loss": 1.4653306961059571,
"mean_token_accuracy": 0.6311864696443081,
"num_tokens": 91730.0,
"step": 30
},
{
"entropy": 1.1646109029650689,
"epoch": 0.21206096752816433,
"grad_norm": 8.5625,
"learning_rate": 1.3000000000000001e-05,
"loss": 1.200753402709961,
"mean_token_accuracy": 0.6730783767998219,
"num_tokens": 122896.0,
"step": 40
},
{
"entropy": 1.0138436160981654,
"epoch": 0.26507620941020543,
"grad_norm": 7.4375,
"learning_rate": 1.6333333333333335e-05,
"loss": 1.0109354972839355,
"mean_token_accuracy": 0.7080431424081326,
"num_tokens": 153323.0,
"step": 50
},
{
"entropy": 0.8669569574296474,
"epoch": 0.31809145129224653,
"grad_norm": 6.75,
"learning_rate": 1.9666666666666666e-05,
"loss": 0.8859419822692871,
"mean_token_accuracy": 0.7356305047869682,
"num_tokens": 183928.0,
"step": 60
},
{
"entropy": 0.856675973534584,
"epoch": 0.3711066931742876,
"grad_norm": 7.6875,
"learning_rate": 1.9666666666666666e-05,
"loss": 0.8721119880676269,
"mean_token_accuracy": 0.7318881675601006,
"num_tokens": 215054.0,
"step": 70
},
{
"entropy": 0.7872001446783543,
"epoch": 0.42412193505632867,
"grad_norm": 7.46875,
"learning_rate": 1.92962962962963e-05,
"loss": 0.7920477390289307,
"mean_token_accuracy": 0.7553018122911453,
"num_tokens": 245614.0,
"step": 80
},
{
"entropy": 0.7739131577312947,
"epoch": 0.47713717693836977,
"grad_norm": 7.09375,
"learning_rate": 1.8925925925925928e-05,
"loss": 0.7949441432952881,
"mean_token_accuracy": 0.7514781631529331,
"num_tokens": 276651.0,
"step": 90
},
{
"entropy": 0.7763891488313674,
"epoch": 0.5301524188204109,
"grad_norm": 7.8125,
"learning_rate": 1.8555555555555557e-05,
"loss": 0.7867375373840332,
"mean_token_accuracy": 0.7523738898336887,
"num_tokens": 308561.0,
"step": 100
},
{
"entropy": 0.7757978036999702,
"epoch": 0.583167660702452,
"grad_norm": 7.21875,
"learning_rate": 1.8185185185185186e-05,
"loss": 0.7776416301727295,
"mean_token_accuracy": 0.7548527352511882,
"num_tokens": 339537.0,
"step": 110
},
{
"entropy": 0.7528716944158077,
"epoch": 0.6361829025844931,
"grad_norm": 6.28125,
"learning_rate": 1.7814814814814815e-05,
"loss": 0.7566296100616455,
"mean_token_accuracy": 0.7637521453201771,
"num_tokens": 370938.0,
"step": 120
},
{
"entropy": 0.7483146995306015,
"epoch": 0.6891981444665342,
"grad_norm": 7.0,
"learning_rate": 1.7444444444444448e-05,
"loss": 0.7508650302886963,
"mean_token_accuracy": 0.758705198019743,
"num_tokens": 402708.0,
"step": 130
},
{
"entropy": 0.7206243231892586,
"epoch": 0.7422133863485753,
"grad_norm": 6.65625,
"learning_rate": 1.7074074074074077e-05,
"loss": 0.7531259536743165,
"mean_token_accuracy": 0.7641572073101998,
"num_tokens": 433645.0,
"step": 140
},
{
"entropy": 0.7596748031675815,
"epoch": 0.7952286282306164,
"grad_norm": 6.03125,
"learning_rate": 1.6703703703703703e-05,
"loss": 0.7591583728790283,
"mean_token_accuracy": 0.7570627965033054,
"num_tokens": 465791.0,
"step": 150
},
{
"entropy": 0.7133973561227321,
"epoch": 0.8482438701126573,
"grad_norm": 7.125,
"learning_rate": 1.6333333333333335e-05,
"loss": 0.7458220005035401,
"mean_token_accuracy": 0.7626615799963474,
"num_tokens": 497296.0,
"step": 160
},
{
"entropy": 0.7134327009320259,
"epoch": 0.9012591119946984,
"grad_norm": 7.0,
"learning_rate": 1.5962962962962964e-05,
"loss": 0.7290266990661621,
"mean_token_accuracy": 0.7681502625346184,
"num_tokens": 528181.0,
"step": 170
},
{
"entropy": 0.7324420973658562,
"epoch": 0.9542743538767395,
"grad_norm": 7.5,
"learning_rate": 1.5592592592592593e-05,
"loss": 0.731742525100708,
"mean_token_accuracy": 0.7631207950413227,
"num_tokens": 558915.0,
"step": 180
},
{
"entropy": 0.7100333733992144,
"epoch": 1.0053015241882042,
"grad_norm": 8.3125,
"learning_rate": 1.5222222222222223e-05,
"loss": 0.7303330898284912,
"mean_token_accuracy": 0.7633489200047084,
"num_tokens": 588567.0,
"step": 190
},
{
"entropy": 0.6628884050995112,
"epoch": 1.0583167660702453,
"grad_norm": 7.0625,
"learning_rate": 1.4851851851851853e-05,
"loss": 0.6481011867523193,
"mean_token_accuracy": 0.7933171860873699,
"num_tokens": 619081.0,
"step": 200
},
{
"entropy": 0.6257969986647367,
"epoch": 1.1113320079522864,
"grad_norm": 7.25,
"learning_rate": 1.4481481481481483e-05,
"loss": 0.642666244506836,
"mean_token_accuracy": 0.7921255677938461,
"num_tokens": 650402.0,
"step": 210
},
{
"entropy": 0.6399037420749665,
"epoch": 1.1643472498343272,
"grad_norm": 6.03125,
"learning_rate": 1.4111111111111113e-05,
"loss": 0.6525338172912598,
"mean_token_accuracy": 0.7887163788080216,
"num_tokens": 681837.0,
"step": 220
},
{
"entropy": 0.6367121540009976,
"epoch": 1.2173624917163686,
"grad_norm": 7.59375,
"learning_rate": 1.3740740740740741e-05,
"loss": 0.6356833457946778,
"mean_token_accuracy": 0.7893597863614559,
"num_tokens": 712571.0,
"step": 230
},
{
"entropy": 0.6334792256355286,
"epoch": 1.2703777335984094,
"grad_norm": 6.6875,
"learning_rate": 1.3370370370370372e-05,
"loss": 0.6391499996185303,
"mean_token_accuracy": 0.7947010740637779,
"num_tokens": 743396.0,
"step": 240
},
{
"entropy": 0.6417076423764229,
"epoch": 1.3233929754804508,
"grad_norm": 7.4375,
"learning_rate": 1.3000000000000001e-05,
"loss": 0.6563210010528564,
"mean_token_accuracy": 0.787147381901741,
"num_tokens": 774883.0,
"step": 250
},
{
"entropy": 0.6213644035160542,
"epoch": 1.3764082173624916,
"grad_norm": 7.1875,
"learning_rate": 1.2629629629629632e-05,
"loss": 0.6295370578765869,
"mean_token_accuracy": 0.7955153584480286,
"num_tokens": 805520.0,
"step": 260
},
{
"entropy": 0.6552363462746144,
"epoch": 1.4294234592445327,
"grad_norm": 6.84375,
"learning_rate": 1.225925925925926e-05,
"loss": 0.6632986068725586,
"mean_token_accuracy": 0.7826757229864597,
"num_tokens": 836758.0,
"step": 270
},
{
"entropy": 0.6550090927630663,
"epoch": 1.4824387011265738,
"grad_norm": 7.25,
"learning_rate": 1.188888888888889e-05,
"loss": 0.6479949474334716,
"mean_token_accuracy": 0.78950335085392,
"num_tokens": 868056.0,
"step": 280
},
{
"entropy": 0.6275094959884882,
"epoch": 1.535453943008615,
"grad_norm": 6.4375,
"learning_rate": 1.151851851851852e-05,
"loss": 0.6384372234344482,
"mean_token_accuracy": 0.7930210866034031,
"num_tokens": 899420.0,
"step": 290
},
{
"entropy": 0.6052290312945843,
"epoch": 1.588469184890656,
"grad_norm": 6.875,
"learning_rate": 1.114814814814815e-05,
"loss": 0.609743309020996,
"mean_token_accuracy": 0.7990770071744919,
"num_tokens": 929571.0,
"step": 300
},
{
"entropy": 0.6174360021948815,
"epoch": 1.6414844267726971,
"grad_norm": 7.0625,
"learning_rate": 1.0777777777777778e-05,
"loss": 0.6313478946685791,
"mean_token_accuracy": 0.7964761383831501,
"num_tokens": 960928.0,
"step": 310
},
{
"entropy": 0.6127162992954254,
"epoch": 1.6944996686547382,
"grad_norm": 6.5625,
"learning_rate": 1.0407407407407408e-05,
"loss": 0.6150648593902588,
"mean_token_accuracy": 0.8008938111364842,
"num_tokens": 991961.0,
"step": 320
},
{
"entropy": 0.5947048265486956,
"epoch": 1.7475149105367793,
"grad_norm": 7.65625,
"learning_rate": 1.003703703703704e-05,
"loss": 0.6053301334381104,
"mean_token_accuracy": 0.801536674797535,
"num_tokens": 1022013.0,
"step": 330
},
{
"entropy": 0.6275908622890711,
"epoch": 1.8005301524188204,
"grad_norm": 6.9375,
"learning_rate": 9.666666666666667e-06,
"loss": 0.6307861804962158,
"mean_token_accuracy": 0.7918523952364922,
"num_tokens": 1053061.0,
"step": 340
},
{
"entropy": 0.6406017865985631,
"epoch": 1.8535453943008615,
"grad_norm": 7.0,
"learning_rate": 9.296296296296296e-06,
"loss": 0.6534273147583007,
"mean_token_accuracy": 0.789715650677681,
"num_tokens": 1085392.0,
"step": 350
},
{
"entropy": 0.6285702683031559,
"epoch": 1.9065606361829026,
"grad_norm": 7.8125,
"learning_rate": 8.925925925925927e-06,
"loss": 0.6331883907318115,
"mean_token_accuracy": 0.7924636773765087,
"num_tokens": 1117204.0,
"step": 360
},
{
"entropy": 0.615465309098363,
"epoch": 1.9595758780649437,
"grad_norm": 7.25,
"learning_rate": 8.555555555555556e-06,
"loss": 0.6278977870941163,
"mean_token_accuracy": 0.7965366452932358,
"num_tokens": 1147711.0,
"step": 370
}
],
"logging_steps": 10,
"max_steps": 600,
"num_input_tokens_seen": 0,
"num_train_epochs": 4,
"save_steps": 500,
"stateful_callbacks": {
"TrainerControl": {
"args": {
"should_epoch_stop": false,
"should_evaluate": false,
"should_log": false,
"should_save": true,
"should_training_stop": false
},
"attributes": {}
}
},
"total_flos": 1.131034650599424e+16,
"train_batch_size": 2,
"trial_name": null,
"trial_params": null
}