zlm_b32_le5_s4000 / checkpoint-1000 /trainer_state.json
mikhail-panzo's picture
Training in progress, step 1000, checkpoint
f26d572 verified
{
"best_metric": 0.514046847820282,
"best_model_checkpoint": "mikhail-panzo/zlm_b32_le5_s4000/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": 5.243492126464844,
"learning_rate": 2.4000000000000003e-07,
"loss": 1.1863,
"step": 50
},
{
"epoch": 0.041884816753926704,
"grad_norm": 16.012319564819336,
"learning_rate": 4.900000000000001e-07,
"loss": 1.01,
"step": 100
},
{
"epoch": 0.06282722513089005,
"grad_norm": 3.0757930278778076,
"learning_rate": 7.4e-07,
"loss": 0.9741,
"step": 150
},
{
"epoch": 0.08376963350785341,
"grad_norm": 3.897144079208374,
"learning_rate": 9.85e-07,
"loss": 0.9358,
"step": 200
},
{
"epoch": 0.10471204188481675,
"grad_norm": 3.6317360401153564,
"learning_rate": 1.235e-06,
"loss": 0.8767,
"step": 250
},
{
"epoch": 0.1256544502617801,
"grad_norm": 4.750243186950684,
"learning_rate": 1.485e-06,
"loss": 0.8576,
"step": 300
},
{
"epoch": 0.14659685863874344,
"grad_norm": 3.4898314476013184,
"learning_rate": 1.7350000000000001e-06,
"loss": 0.8472,
"step": 350
},
{
"epoch": 0.16753926701570682,
"grad_norm": 4.961293697357178,
"learning_rate": 1.985e-06,
"loss": 0.7591,
"step": 400
},
{
"epoch": 0.18848167539267016,
"grad_norm": 3.254897117614746,
"learning_rate": 2.235e-06,
"loss": 0.7548,
"step": 450
},
{
"epoch": 0.2094240837696335,
"grad_norm": 3.3995440006256104,
"learning_rate": 2.4850000000000003e-06,
"loss": 0.7211,
"step": 500
},
{
"epoch": 0.2094240837696335,
"eval_loss": 0.6148238182067871,
"eval_runtime": 267.7517,
"eval_samples_per_second": 31.705,
"eval_steps_per_second": 3.966,
"step": 500
},
{
"epoch": 0.23036649214659685,
"grad_norm": 3.2476744651794434,
"learning_rate": 2.7350000000000005e-06,
"loss": 0.6859,
"step": 550
},
{
"epoch": 0.2513089005235602,
"grad_norm": 4.222187042236328,
"learning_rate": 2.9850000000000002e-06,
"loss": 0.6653,
"step": 600
},
{
"epoch": 0.27225130890052357,
"grad_norm": 3.4279122352600098,
"learning_rate": 3.2350000000000004e-06,
"loss": 0.6612,
"step": 650
},
{
"epoch": 0.2931937172774869,
"grad_norm": 2.7588794231414795,
"learning_rate": 3.485e-06,
"loss": 0.6404,
"step": 700
},
{
"epoch": 0.31413612565445026,
"grad_norm": 3.481127977371216,
"learning_rate": 3.7350000000000002e-06,
"loss": 0.647,
"step": 750
},
{
"epoch": 0.33507853403141363,
"grad_norm": 4.153278827667236,
"learning_rate": 3.985000000000001e-06,
"loss": 0.6374,
"step": 800
},
{
"epoch": 0.35602094240837695,
"grad_norm": 5.10654354095459,
"learning_rate": 4.235e-06,
"loss": 0.6225,
"step": 850
},
{
"epoch": 0.3769633507853403,
"grad_norm": 2.7014458179473877,
"learning_rate": 4.485e-06,
"loss": 0.6026,
"step": 900
},
{
"epoch": 0.39790575916230364,
"grad_norm": 2.965581178665161,
"learning_rate": 4.735e-06,
"loss": 0.5963,
"step": 950
},
{
"epoch": 0.418848167539267,
"grad_norm": 2.671508312225342,
"learning_rate": 4.9850000000000006e-06,
"loss": 0.6059,
"step": 1000
},
{
"epoch": 0.418848167539267,
"eval_loss": 0.514046847820282,
"eval_runtime": 268.9563,
"eval_samples_per_second": 31.563,
"eval_steps_per_second": 3.949,
"step": 1000
}
],
"logging_steps": 50,
"max_steps": 4000,
"num_input_tokens_seen": 0,
"num_train_epochs": 2,
"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
}