302 lines
7.7 KiB
JSON
302 lines
7.7 KiB
JSON
{
|
|
"best_global_step": 1500,
|
|
"best_metric": 0.4548405110836029,
|
|
"best_model_checkpoint": "./qwen3-4b-2507-ru-lora/checkpoint-1500",
|
|
"epoch": 3.4248394004282656,
|
|
"eval_steps": 500,
|
|
"global_step": 3000,
|
|
"is_hyper_param_search": false,
|
|
"is_local_process_zero": true,
|
|
"is_world_process_zero": true,
|
|
"log_history": [
|
|
{
|
|
"epoch": 0.11420413990007137,
|
|
"grad_norm": 0.1101919561624527,
|
|
"learning_rate": 9.900000000000001e-05,
|
|
"loss": 1.1473062133789063,
|
|
"step": 100
|
|
},
|
|
{
|
|
"epoch": 0.22840827980014275,
|
|
"grad_norm": 0.09405623376369476,
|
|
"learning_rate": 9.832203389830509e-05,
|
|
"loss": 0.797086181640625,
|
|
"step": 200
|
|
},
|
|
{
|
|
"epoch": 0.3426124197002141,
|
|
"grad_norm": 0.1505233645439148,
|
|
"learning_rate": 9.66271186440678e-05,
|
|
"loss": 0.8210951232910156,
|
|
"step": 300
|
|
},
|
|
{
|
|
"epoch": 0.4568165596002855,
|
|
"grad_norm": 0.09275370091199875,
|
|
"learning_rate": 9.493220338983051e-05,
|
|
"loss": 0.7762732696533203,
|
|
"step": 400
|
|
},
|
|
{
|
|
"epoch": 0.5710206995003569,
|
|
"grad_norm": 0.15582314133644104,
|
|
"learning_rate": 9.323728813559323e-05,
|
|
"loss": 0.79928955078125,
|
|
"step": 500
|
|
},
|
|
{
|
|
"epoch": 0.5710206995003569,
|
|
"eval_loss": 0.4887394309043884,
|
|
"eval_runtime": 85.2625,
|
|
"eval_samples_per_second": 4.328,
|
|
"eval_steps_per_second": 4.328,
|
|
"step": 500
|
|
},
|
|
{
|
|
"epoch": 0.6852248394004282,
|
|
"grad_norm": 0.06571248918771744,
|
|
"learning_rate": 9.154237288135593e-05,
|
|
"loss": 0.7494968414306641,
|
|
"step": 600
|
|
},
|
|
{
|
|
"epoch": 0.7994289793004996,
|
|
"grad_norm": 0.22434799373149872,
|
|
"learning_rate": 8.984745762711865e-05,
|
|
"loss": 0.7933966827392578,
|
|
"step": 700
|
|
},
|
|
{
|
|
"epoch": 0.913633119200571,
|
|
"grad_norm": 0.14180706441402435,
|
|
"learning_rate": 8.815254237288137e-05,
|
|
"loss": 0.7898688507080078,
|
|
"step": 800
|
|
},
|
|
{
|
|
"epoch": 1.0274089935760171,
|
|
"grad_norm": 0.09555750340223312,
|
|
"learning_rate": 8.645762711864407e-05,
|
|
"loss": 0.753864517211914,
|
|
"step": 900
|
|
},
|
|
{
|
|
"epoch": 1.1416131334760884,
|
|
"grad_norm": 0.13440796732902527,
|
|
"learning_rate": 8.476271186440679e-05,
|
|
"loss": 0.6366143798828126,
|
|
"step": 1000
|
|
},
|
|
{
|
|
"epoch": 1.1416131334760884,
|
|
"eval_loss": 0.4765624701976776,
|
|
"eval_runtime": 85.3281,
|
|
"eval_samples_per_second": 4.324,
|
|
"eval_steps_per_second": 4.324,
|
|
"step": 1000
|
|
},
|
|
{
|
|
"epoch": 1.25581727337616,
|
|
"grad_norm": 0.09627823531627655,
|
|
"learning_rate": 8.30677966101695e-05,
|
|
"loss": 0.6536759948730468,
|
|
"step": 1100
|
|
},
|
|
{
|
|
"epoch": 1.3700214132762314,
|
|
"grad_norm": 0.216465562582016,
|
|
"learning_rate": 8.137288135593221e-05,
|
|
"loss": 0.6386687850952149,
|
|
"step": 1200
|
|
},
|
|
{
|
|
"epoch": 1.4842255531763027,
|
|
"grad_norm": 0.16032801568508148,
|
|
"learning_rate": 7.967796610169492e-05,
|
|
"loss": 0.6274388885498047,
|
|
"step": 1300
|
|
},
|
|
{
|
|
"epoch": 1.598429693076374,
|
|
"grad_norm": 0.3502936065196991,
|
|
"learning_rate": 7.798305084745763e-05,
|
|
"loss": 0.6400037384033204,
|
|
"step": 1400
|
|
},
|
|
{
|
|
"epoch": 1.7126338329764454,
|
|
"grad_norm": 0.21663837134838104,
|
|
"learning_rate": 7.628813559322034e-05,
|
|
"loss": 0.6600186920166016,
|
|
"step": 1500
|
|
},
|
|
{
|
|
"epoch": 1.7126338329764454,
|
|
"eval_loss": 0.4548405110836029,
|
|
"eval_runtime": 85.2775,
|
|
"eval_samples_per_second": 4.327,
|
|
"eval_steps_per_second": 4.327,
|
|
"step": 1500
|
|
},
|
|
{
|
|
"epoch": 1.8268379728765167,
|
|
"grad_norm": 0.1811513453722,
|
|
"learning_rate": 7.459322033898304e-05,
|
|
"loss": 0.6541920471191406,
|
|
"step": 1600
|
|
},
|
|
{
|
|
"epoch": 1.941042112776588,
|
|
"grad_norm": 0.13848915696144104,
|
|
"learning_rate": 7.289830508474576e-05,
|
|
"loss": 0.6397422790527344,
|
|
"step": 1700
|
|
},
|
|
{
|
|
"epoch": 2.0548179871520342,
|
|
"grad_norm": 0.10245347768068314,
|
|
"learning_rate": 7.120338983050848e-05,
|
|
"loss": 0.5696288299560547,
|
|
"step": 1800
|
|
},
|
|
{
|
|
"epoch": 2.1690221270521057,
|
|
"grad_norm": 0.11742890626192093,
|
|
"learning_rate": 6.950847457627118e-05,
|
|
"loss": 0.44962631225585936,
|
|
"step": 1900
|
|
},
|
|
{
|
|
"epoch": 2.283226266952177,
|
|
"grad_norm": 0.20419853925704956,
|
|
"learning_rate": 6.78135593220339e-05,
|
|
"loss": 0.4812296676635742,
|
|
"step": 2000
|
|
},
|
|
{
|
|
"epoch": 2.283226266952177,
|
|
"eval_loss": 0.47145259380340576,
|
|
"eval_runtime": 85.3343,
|
|
"eval_samples_per_second": 4.324,
|
|
"eval_steps_per_second": 4.324,
|
|
"step": 2000
|
|
},
|
|
{
|
|
"epoch": 2.3974304068522483,
|
|
"grad_norm": 0.21591816842556,
|
|
"learning_rate": 6.61186440677966e-05,
|
|
"loss": 0.47923736572265624,
|
|
"step": 2100
|
|
},
|
|
{
|
|
"epoch": 2.51163454675232,
|
|
"grad_norm": 0.29100939631462097,
|
|
"learning_rate": 6.442372881355932e-05,
|
|
"loss": 0.4617042541503906,
|
|
"step": 2200
|
|
},
|
|
{
|
|
"epoch": 2.6258386866523913,
|
|
"grad_norm": 0.15308618545532227,
|
|
"learning_rate": 6.272881355932204e-05,
|
|
"loss": 0.501895751953125,
|
|
"step": 2300
|
|
},
|
|
{
|
|
"epoch": 2.740042826552463,
|
|
"grad_norm": 0.1499933898448944,
|
|
"learning_rate": 6.103389830508475e-05,
|
|
"loss": 0.4711552047729492,
|
|
"step": 2400
|
|
},
|
|
{
|
|
"epoch": 2.854246966452534,
|
|
"grad_norm": 0.24623849987983704,
|
|
"learning_rate": 5.9338983050847465e-05,
|
|
"loss": 0.47890899658203123,
|
|
"step": 2500
|
|
},
|
|
{
|
|
"epoch": 2.854246966452534,
|
|
"eval_loss": 0.47348523139953613,
|
|
"eval_runtime": 85.3423,
|
|
"eval_samples_per_second": 4.324,
|
|
"eval_steps_per_second": 4.324,
|
|
"step": 2500
|
|
},
|
|
{
|
|
"epoch": 2.9684511063526053,
|
|
"grad_norm": 0.14382973313331604,
|
|
"learning_rate": 5.764406779661017e-05,
|
|
"loss": 0.4821274185180664,
|
|
"step": 2600
|
|
},
|
|
{
|
|
"epoch": 3.0822269807280516,
|
|
"grad_norm": 0.17034879326820374,
|
|
"learning_rate": 5.594915254237289e-05,
|
|
"loss": 0.3882878112792969,
|
|
"step": 2700
|
|
},
|
|
{
|
|
"epoch": 3.1964311206281226,
|
|
"grad_norm": 0.22748088836669922,
|
|
"learning_rate": 5.42542372881356e-05,
|
|
"loss": 0.3157962989807129,
|
|
"step": 2800
|
|
},
|
|
{
|
|
"epoch": 3.310635260528194,
|
|
"grad_norm": 0.1980806142091751,
|
|
"learning_rate": 5.25593220338983e-05,
|
|
"loss": 0.3221218490600586,
|
|
"step": 2900
|
|
},
|
|
{
|
|
"epoch": 3.4248394004282656,
|
|
"grad_norm": 0.22527898848056793,
|
|
"learning_rate": 5.086440677966102e-05,
|
|
"loss": 0.32430213928222656,
|
|
"step": 3000
|
|
},
|
|
{
|
|
"epoch": 3.4248394004282656,
|
|
"eval_loss": 0.5339688062667847,
|
|
"eval_runtime": 85.3358,
|
|
"eval_samples_per_second": 4.324,
|
|
"eval_steps_per_second": 4.324,
|
|
"step": 3000
|
|
}
|
|
],
|
|
"logging_steps": 100,
|
|
"max_steps": 6000,
|
|
"num_input_tokens_seen": 0,
|
|
"num_train_epochs": 7,
|
|
"save_steps": 500,
|
|
"stateful_callbacks": {
|
|
"EarlyStoppingCallback": {
|
|
"args": {
|
|
"early_stopping_patience": 3,
|
|
"early_stopping_threshold": 0.0
|
|
},
|
|
"attributes": {
|
|
"early_stopping_patience_counter": 3
|
|
}
|
|
},
|
|
"TrainerControl": {
|
|
"args": {
|
|
"should_epoch_stop": false,
|
|
"should_evaluate": false,
|
|
"should_log": false,
|
|
"should_save": true,
|
|
"should_training_stop": true
|
|
},
|
|
"attributes": {}
|
|
}
|
|
},
|
|
"total_flos": 1.5721880180944896e+17,
|
|
"train_batch_size": 1,
|
|
"trial_name": null,
|
|
"trial_params": null
|
|
}
|