zlm_b32_le4_s8000 / checkpoint-1000 /trainer_state.json
mikhail-panzo's picture
Training in progress, step 1000, checkpoint
43c3c7f verified
raw
history blame
4.54 kB
{
"best_metric": 0.4265913963317871,
"best_model_checkpoint": "mikhail-panzo/zlm_b32_le4_s8000/checkpoint-1000",
"epoch": 0.418848167539267,
"eval_steps": 500,
"global_step": 1000,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.020942408376963352,
"grad_norm": 3.1775436401367188,
"learning_rate": 2.35e-06,
"loss": 1.1228,
"step": 50
},
{
"epoch": 0.041884816753926704,
"grad_norm": 4.18980073928833,
"learning_rate": 4.85e-06,
"loss": 0.8304,
"step": 100
},
{
"epoch": 0.06282722513089005,
"grad_norm": 4.5355024337768555,
"learning_rate": 7.35e-06,
"loss": 0.7701,
"step": 150
},
{
"epoch": 0.08376963350785341,
"grad_norm": 3.9590957164764404,
"learning_rate": 9.85e-06,
"loss": 0.7096,
"step": 200
},
{
"epoch": 0.10471204188481675,
"grad_norm": 5.419675350189209,
"learning_rate": 1.235e-05,
"loss": 0.6518,
"step": 250
},
{
"epoch": 0.1256544502617801,
"grad_norm": 3.1052777767181396,
"learning_rate": 1.485e-05,
"loss": 0.6326,
"step": 300
},
{
"epoch": 0.14659685863874344,
"grad_norm": 3.4649453163146973,
"learning_rate": 1.7349999999999998e-05,
"loss": 0.6241,
"step": 350
},
{
"epoch": 0.16753926701570682,
"grad_norm": 4.962624549865723,
"learning_rate": 1.985e-05,
"loss": 0.5723,
"step": 400
},
{
"epoch": 0.18848167539267016,
"grad_norm": 2.7586095333099365,
"learning_rate": 2.235e-05,
"loss": 0.5764,
"step": 450
},
{
"epoch": 0.2094240837696335,
"grad_norm": 3.5610299110412598,
"learning_rate": 2.485e-05,
"loss": 0.5552,
"step": 500
},
{
"epoch": 0.2094240837696335,
"eval_loss": 0.48825809359550476,
"eval_runtime": 277.5892,
"eval_samples_per_second": 30.581,
"eval_steps_per_second": 3.826,
"step": 500
},
{
"epoch": 0.23036649214659685,
"grad_norm": 3.3351809978485107,
"learning_rate": 2.7350000000000004e-05,
"loss": 0.5322,
"step": 550
},
{
"epoch": 0.2513089005235602,
"grad_norm": 3.7222373485565186,
"learning_rate": 2.985e-05,
"loss": 0.5186,
"step": 600
},
{
"epoch": 0.27225130890052357,
"grad_norm": 3.1811156272888184,
"learning_rate": 3.235e-05,
"loss": 0.5157,
"step": 650
},
{
"epoch": 0.2931937172774869,
"grad_norm": 2.2585642337799072,
"learning_rate": 3.485e-05,
"loss": 0.5089,
"step": 700
},
{
"epoch": 0.31413612565445026,
"grad_norm": 3.999460220336914,
"learning_rate": 3.735e-05,
"loss": 0.5117,
"step": 750
},
{
"epoch": 0.33507853403141363,
"grad_norm": 3.739990472793579,
"learning_rate": 3.9850000000000006e-05,
"loss": 0.5031,
"step": 800
},
{
"epoch": 0.35602094240837695,
"grad_norm": 4.251980781555176,
"learning_rate": 4.235e-05,
"loss": 0.5064,
"step": 850
},
{
"epoch": 0.3769633507853403,
"grad_norm": 2.770602226257324,
"learning_rate": 4.4850000000000006e-05,
"loss": 0.4921,
"step": 900
},
{
"epoch": 0.39790575916230364,
"grad_norm": 2.506974220275879,
"learning_rate": 4.735e-05,
"loss": 0.4839,
"step": 950
},
{
"epoch": 0.418848167539267,
"grad_norm": 2.2666189670562744,
"learning_rate": 4.9850000000000006e-05,
"loss": 0.4913,
"step": 1000
},
{
"epoch": 0.418848167539267,
"eval_loss": 0.4265913963317871,
"eval_runtime": 265.1599,
"eval_samples_per_second": 32.015,
"eval_steps_per_second": 4.005,
"step": 1000
}
],
"logging_steps": 50,
"max_steps": 8000,
"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": 4446942197892480.0,
"train_batch_size": 16,
"trial_name": null,
"trial_params": null
}