{
"best_metric": 0.6912627220153809,
"best_model_checkpoint": "./output/clip-finetuned-csu-p14-336-e3l56-l/checkpoint-15000",
"epoch": 2.7634487840825352,
"eval_steps": 500,
"global_step": 15000,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.09211495946941783,
"grad_norm": 18.028329849243164,
"learning_rate": 4.846475067550971e-06,
"loss": 0.2253,
"step": 500
},
{
"epoch": 0.09211495946941783,
"eval_loss": 1.4693936109542847,
"eval_runtime": 74.2802,
"eval_samples_per_second": 16.249,
"eval_steps_per_second": 2.033,
"step": 500
},
{
"epoch": 0.18422991893883567,
"grad_norm": 0.4609026312828064,
"learning_rate": 4.692950135101941e-06,
"loss": 0.2544,
"step": 1000
},
{
"epoch": 0.18422991893883567,
"eval_loss": 1.5420819520950317,
"eval_runtime": 74.0766,
"eval_samples_per_second": 16.294,
"eval_steps_per_second": 2.038,
"step": 1000
},
{
"epoch": 0.2763448784082535,
"grad_norm": 71.96737670898438,
"learning_rate": 4.5394252026529115e-06,
"loss": 0.3236,
"step": 1500
},
{
"epoch": 0.2763448784082535,
"eval_loss": 1.4642689228057861,
"eval_runtime": 74.2336,
"eval_samples_per_second": 16.259,
"eval_steps_per_second": 2.034,
"step": 1500
},
{
"epoch": 0.36845983787767134,
"grad_norm": 6.366127069945549e-10,
"learning_rate": 4.385900270203882e-06,
"loss": 0.1888,
"step": 2000
},
{
"epoch": 0.36845983787767134,
"eval_loss": 1.3193777799606323,
"eval_runtime": 76.1214,
"eval_samples_per_second": 15.856,
"eval_steps_per_second": 1.984,
"step": 2000
},
{
"epoch": 0.46057479734708917,
"grad_norm": 452.8532409667969,
"learning_rate": 4.232375337754851e-06,
"loss": 0.2563,
"step": 2500
},
{
"epoch": 0.46057479734708917,
"eval_loss": 1.37644362449646,
"eval_runtime": 76.3265,
"eval_samples_per_second": 15.814,
"eval_steps_per_second": 1.978,
"step": 2500
},
{
"epoch": 0.552689756816507,
"grad_norm": 0.0002858435327652842,
"learning_rate": 4.078850405305822e-06,
"loss": 0.2794,
"step": 3000
},
{
"epoch": 0.552689756816507,
"eval_loss": 1.3007307052612305,
"eval_runtime": 76.3512,
"eval_samples_per_second": 15.809,
"eval_steps_per_second": 1.978,
"step": 3000
},
{
"epoch": 0.6448047162859248,
"grad_norm": 0.324462890625,
"learning_rate": 3.925325472856792e-06,
"loss": 0.1749,
"step": 3500
},
{
"epoch": 0.6448047162859248,
"eval_loss": 1.3360542058944702,
"eval_runtime": 74.5379,
"eval_samples_per_second": 16.193,
"eval_steps_per_second": 2.026,
"step": 3500
},
{
"epoch": 0.7369196757553427,
"grad_norm": 76.63452911376953,
"learning_rate": 3.7718005404077624e-06,
"loss": 0.2672,
"step": 4000
},
{
"epoch": 0.7369196757553427,
"eval_loss": 1.268362283706665,
"eval_runtime": 76.6253,
"eval_samples_per_second": 15.752,
"eval_steps_per_second": 1.971,
"step": 4000
},
{
"epoch": 0.8290346352247605,
"grad_norm": 1.9390699863433838,
"learning_rate": 3.6182756079587327e-06,
"loss": 0.218,
"step": 4500
},
{
"epoch": 0.8290346352247605,
"eval_loss": 1.1064503192901611,
"eval_runtime": 75.0503,
"eval_samples_per_second": 16.083,
"eval_steps_per_second": 2.012,
"step": 4500
},
{
"epoch": 0.9211495946941783,
"grad_norm": 3.37814665130054e-10,
"learning_rate": 3.464750675509703e-06,
"loss": 0.1665,
"step": 5000
},
{
"epoch": 0.9211495946941783,
"eval_loss": 1.0620027780532837,
"eval_runtime": 76.9535,
"eval_samples_per_second": 15.685,
"eval_steps_per_second": 1.962,
"step": 5000
},
{
"epoch": 1.013264554163596,
"grad_norm": 8.782347293845305e-09,
"learning_rate": 3.311225743060673e-06,
"loss": 0.1842,
"step": 5500
},
{
"epoch": 1.013264554163596,
"eval_loss": 0.9443314075469971,
"eval_runtime": 75.2095,
"eval_samples_per_second": 16.049,
"eval_steps_per_second": 2.008,
"step": 5500
},
{
"epoch": 1.105379513633014,
"grad_norm": 0.002741970354691148,
"learning_rate": 3.1577008106116434e-06,
"loss": 0.1183,
"step": 6000
},
{
"epoch": 1.105379513633014,
"eval_loss": 0.943097710609436,
"eval_runtime": 74.832,
"eval_samples_per_second": 16.129,
"eval_steps_per_second": 2.018,
"step": 6000
},
{
"epoch": 1.1974944731024317,
"grad_norm": 9.241516818292439e-05,
"learning_rate": 3.004175878162614e-06,
"loss": 0.1066,
"step": 6500
},
{
"epoch": 1.1974944731024317,
"eval_loss": 1.005082607269287,
"eval_runtime": 74.7497,
"eval_samples_per_second": 16.147,
"eval_steps_per_second": 2.02,
"step": 6500
},
{
"epoch": 1.2896094325718497,
"grad_norm": 3.972076889624532e-09,
"learning_rate": 2.850650945713584e-06,
"loss": 0.0901,
"step": 7000
},
{
"epoch": 1.2896094325718497,
"eval_loss": 1.001724362373352,
"eval_runtime": 74.9393,
"eval_samples_per_second": 16.106,
"eval_steps_per_second": 2.015,
"step": 7000
},
{
"epoch": 1.3817243920412676,
"grad_norm": 0.6726647615432739,
"learning_rate": 2.6971260132645544e-06,
"loss": 0.112,
"step": 7500
},
{
"epoch": 1.3817243920412676,
"eval_loss": 1.0029873847961426,
"eval_runtime": 76.2263,
"eval_samples_per_second": 15.834,
"eval_steps_per_second": 1.981,
"step": 7500
},
{
"epoch": 1.4738393515106853,
"grad_norm": 0.0007436455925926566,
"learning_rate": 2.543601080815525e-06,
"loss": 0.1265,
"step": 8000
},
{
"epoch": 1.4738393515106853,
"eval_loss": 0.9462646842002869,
"eval_runtime": 74.6539,
"eval_samples_per_second": 16.168,
"eval_steps_per_second": 2.023,
"step": 8000
},
{
"epoch": 1.565954310980103,
"grad_norm": 3.9008838825793646e-08,
"learning_rate": 2.390076148366495e-06,
"loss": 0.1193,
"step": 8500
},
{
"epoch": 1.565954310980103,
"eval_loss": 1.0377594232559204,
"eval_runtime": 74.635,
"eval_samples_per_second": 16.172,
"eval_steps_per_second": 2.023,
"step": 8500
},
{
"epoch": 1.658069270449521,
"grad_norm": 1.2957306353200693e-05,
"learning_rate": 2.236551215917465e-06,
"loss": 0.1318,
"step": 9000
},
{
"epoch": 1.658069270449521,
"eval_loss": 0.9298674464225769,
"eval_runtime": 74.5158,
"eval_samples_per_second": 16.198,
"eval_steps_per_second": 2.026,
"step": 9000
},
{
"epoch": 1.750184229918939,
"grad_norm": 2.964475670150235e-12,
"learning_rate": 2.0830262834684354e-06,
"loss": 0.1262,
"step": 9500
},
{
"epoch": 1.750184229918939,
"eval_loss": 0.9714127779006958,
"eval_runtime": 74.8339,
"eval_samples_per_second": 16.129,
"eval_steps_per_second": 2.018,
"step": 9500
},
{
"epoch": 1.8422991893883567,
"grad_norm": 1.1816314327006694e-06,
"learning_rate": 1.9295013510194058e-06,
"loss": 0.101,
"step": 10000
},
{
"epoch": 1.8422991893883567,
"eval_loss": 0.8753853440284729,
"eval_runtime": 74.7629,
"eval_samples_per_second": 16.144,
"eval_steps_per_second": 2.02,
"step": 10000
},
{
"epoch": 1.9344141488577744,
"grad_norm": 1.1530125298975236e-08,
"learning_rate": 1.775976418570376e-06,
"loss": 0.1158,
"step": 10500
},
{
"epoch": 1.9344141488577744,
"eval_loss": 0.8074946999549866,
"eval_runtime": 74.5296,
"eval_samples_per_second": 16.195,
"eval_steps_per_second": 2.026,
"step": 10500
},
{
"epoch": 2.026529108327192,
"grad_norm": 2.0235971232018812e-10,
"learning_rate": 1.622451486121346e-06,
"loss": 0.0656,
"step": 11000
},
{
"epoch": 2.026529108327192,
"eval_loss": 0.8281168937683105,
"eval_runtime": 76.1895,
"eval_samples_per_second": 15.842,
"eval_steps_per_second": 1.982,
"step": 11000
},
{
"epoch": 2.1186440677966103,
"grad_norm": 27.123056411743164,
"learning_rate": 1.4689265536723166e-06,
"loss": 0.0854,
"step": 11500
},
{
"epoch": 2.1186440677966103,
"eval_loss": 0.7755604982376099,
"eval_runtime": 74.506,
"eval_samples_per_second": 16.2,
"eval_steps_per_second": 2.027,
"step": 11500
},
{
"epoch": 2.210759027266028,
"grad_norm": 0.017793377861380577,
"learning_rate": 1.3154016212232868e-06,
"loss": 0.0574,
"step": 12000
},
{
"epoch": 2.210759027266028,
"eval_loss": 0.7431201934814453,
"eval_runtime": 74.44,
"eval_samples_per_second": 16.214,
"eval_steps_per_second": 2.028,
"step": 12000
},
{
"epoch": 2.3028739867354457,
"grad_norm": 5.614988327026367,
"learning_rate": 1.1618766887742571e-06,
"loss": 0.0643,
"step": 12500
},
{
"epoch": 2.3028739867354457,
"eval_loss": 0.7556394338607788,
"eval_runtime": 76.2998,
"eval_samples_per_second": 15.819,
"eval_steps_per_second": 1.979,
"step": 12500
},
{
"epoch": 2.3949889462048635,
"grad_norm": 0.1345616579055786,
"learning_rate": 1.0083517563252273e-06,
"loss": 0.0657,
"step": 13000
},
{
"epoch": 2.3949889462048635,
"eval_loss": 0.7818958759307861,
"eval_runtime": 74.4509,
"eval_samples_per_second": 16.212,
"eval_steps_per_second": 2.028,
"step": 13000
},
{
"epoch": 2.4871039056742816,
"grad_norm": 2.2698461634718115e-06,
"learning_rate": 8.548268238761975e-07,
"loss": 0.0372,
"step": 13500
},
{
"epoch": 2.4871039056742816,
"eval_loss": 0.768934428691864,
"eval_runtime": 74.5573,
"eval_samples_per_second": 16.189,
"eval_steps_per_second": 2.025,
"step": 13500
},
{
"epoch": 2.5792188651436994,
"grad_norm": 3.2547378540039062,
"learning_rate": 7.013018914271679e-07,
"loss": 0.0286,
"step": 14000
},
{
"epoch": 2.5792188651436994,
"eval_loss": 0.7623304724693298,
"eval_runtime": 74.5031,
"eval_samples_per_second": 16.201,
"eval_steps_per_second": 2.027,
"step": 14000
},
{
"epoch": 2.671333824613117,
"grad_norm": 5.614094394446392e-14,
"learning_rate": 5.477769589781381e-07,
"loss": 0.0581,
"step": 14500
},
{
"epoch": 2.671333824613117,
"eval_loss": 0.7251203060150146,
"eval_runtime": 74.3803,
"eval_samples_per_second": 16.227,
"eval_steps_per_second": 2.03,
"step": 14500
},
{
"epoch": 2.7634487840825352,
"grad_norm": 9.663383426783412e-10,
"learning_rate": 3.942520265291084e-07,
"loss": 0.0578,
"step": 15000
},
{
"epoch": 2.7634487840825352,
"eval_loss": 0.6912627220153809,
"eval_runtime": 74.5101,
"eval_samples_per_second": 16.199,
"eval_steps_per_second": 2.027,
"step": 15000
}
],
"logging_steps": 500,
"max_steps": 16284,
"num_input_tokens_seen": 0,
"num_train_epochs": 3,
"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": 5397094785602520.0,
"train_batch_size": 2,
"trial_name": null,
"trial_params": null
}