GaetanMichelet's picture
Model save
489df1b verified
raw
history blame
9.45 kB
{
"best_metric": 0.8653401732444763,
"best_model_checkpoint": "data/Mistral-7B_task-3_60-samples_config-2_full_auto/checkpoint-31",
"epoch": 17.73913043478261,
"eval_steps": 500,
"global_step": 51,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.34782608695652173,
"grad_norm": 0.9554396867752075,
"learning_rate": 1e-05,
"loss": 1.2468,
"step": 1
},
{
"epoch": 0.6956521739130435,
"grad_norm": 0.9398545026779175,
"learning_rate": 2e-05,
"loss": 1.2142,
"step": 2
},
{
"epoch": 0.6956521739130435,
"eval_loss": 1.2086182832717896,
"eval_runtime": 13.9404,
"eval_samples_per_second": 0.861,
"eval_steps_per_second": 0.861,
"step": 2
},
{
"epoch": 1.391304347826087,
"grad_norm": 0.824161946773529,
"learning_rate": 4e-05,
"loss": 1.2072,
"step": 4
},
{
"epoch": 1.7391304347826086,
"eval_loss": 1.1264773607254028,
"eval_runtime": 13.9408,
"eval_samples_per_second": 0.861,
"eval_steps_per_second": 0.861,
"step": 5
},
{
"epoch": 2.0869565217391304,
"grad_norm": 0.5290440320968628,
"learning_rate": 6e-05,
"loss": 1.1638,
"step": 6
},
{
"epoch": 2.782608695652174,
"grad_norm": 0.39439627528190613,
"learning_rate": 8e-05,
"loss": 1.0872,
"step": 8
},
{
"epoch": 2.782608695652174,
"eval_loss": 1.0535776615142822,
"eval_runtime": 13.9384,
"eval_samples_per_second": 0.861,
"eval_steps_per_second": 0.861,
"step": 8
},
{
"epoch": 3.4782608695652173,
"grad_norm": 0.38090240955352783,
"learning_rate": 0.0001,
"loss": 1.0158,
"step": 10
},
{
"epoch": 3.8260869565217392,
"eval_loss": 0.9948161244392395,
"eval_runtime": 13.9429,
"eval_samples_per_second": 0.861,
"eval_steps_per_second": 0.861,
"step": 11
},
{
"epoch": 4.173913043478261,
"grad_norm": 0.38717082142829895,
"learning_rate": 9.987820251299122e-05,
"loss": 1.0022,
"step": 12
},
{
"epoch": 4.869565217391305,
"grad_norm": 0.5468046069145203,
"learning_rate": 9.951340343707852e-05,
"loss": 0.9378,
"step": 14
},
{
"epoch": 4.869565217391305,
"eval_loss": 0.9304003715515137,
"eval_runtime": 13.9403,
"eval_samples_per_second": 0.861,
"eval_steps_per_second": 0.861,
"step": 14
},
{
"epoch": 5.565217391304348,
"grad_norm": 0.2843795418739319,
"learning_rate": 9.890738003669029e-05,
"loss": 0.8785,
"step": 16
},
{
"epoch": 5.913043478260869,
"eval_loss": 0.8948807120323181,
"eval_runtime": 13.9373,
"eval_samples_per_second": 0.861,
"eval_steps_per_second": 0.861,
"step": 17
},
{
"epoch": 6.260869565217392,
"grad_norm": 0.2556290030479431,
"learning_rate": 9.806308479691595e-05,
"loss": 0.8462,
"step": 18
},
{
"epoch": 6.956521739130435,
"grad_norm": 0.19840207695960999,
"learning_rate": 9.698463103929542e-05,
"loss": 0.8173,
"step": 20
},
{
"epoch": 6.956521739130435,
"eval_loss": 0.879335343837738,
"eval_runtime": 13.9371,
"eval_samples_per_second": 0.861,
"eval_steps_per_second": 0.861,
"step": 20
},
{
"epoch": 7.6521739130434785,
"grad_norm": 0.18154288828372955,
"learning_rate": 9.567727288213005e-05,
"loss": 0.8163,
"step": 22
},
{
"epoch": 8.0,
"eval_loss": 0.8702660202980042,
"eval_runtime": 13.9391,
"eval_samples_per_second": 0.861,
"eval_steps_per_second": 0.861,
"step": 23
},
{
"epoch": 8.347826086956522,
"grad_norm": 0.17678385972976685,
"learning_rate": 9.414737964294636e-05,
"loss": 0.7922,
"step": 24
},
{
"epoch": 8.695652173913043,
"eval_loss": 0.8672122955322266,
"eval_runtime": 13.9382,
"eval_samples_per_second": 0.861,
"eval_steps_per_second": 0.861,
"step": 25
},
{
"epoch": 9.043478260869565,
"grad_norm": 0.18471793830394745,
"learning_rate": 9.24024048078213e-05,
"loss": 0.7716,
"step": 26
},
{
"epoch": 9.73913043478261,
"grad_norm": 0.21089470386505127,
"learning_rate": 9.045084971874738e-05,
"loss": 0.7598,
"step": 28
},
{
"epoch": 9.73913043478261,
"eval_loss": 0.8668568730354309,
"eval_runtime": 13.9503,
"eval_samples_per_second": 0.86,
"eval_steps_per_second": 0.86,
"step": 28
},
{
"epoch": 10.434782608695652,
"grad_norm": 0.19992633163928986,
"learning_rate": 8.83022221559489e-05,
"loss": 0.7519,
"step": 30
},
{
"epoch": 10.782608695652174,
"eval_loss": 0.8653401732444763,
"eval_runtime": 13.9411,
"eval_samples_per_second": 0.861,
"eval_steps_per_second": 0.861,
"step": 31
},
{
"epoch": 11.130434782608695,
"grad_norm": 0.2341090738773346,
"learning_rate": 8.596699001693255e-05,
"loss": 0.703,
"step": 32
},
{
"epoch": 11.826086956521738,
"grad_norm": 0.22710266709327698,
"learning_rate": 8.345653031794292e-05,
"loss": 0.7182,
"step": 34
},
{
"epoch": 11.826086956521738,
"eval_loss": 0.8702945113182068,
"eval_runtime": 13.9386,
"eval_samples_per_second": 0.861,
"eval_steps_per_second": 0.861,
"step": 34
},
{
"epoch": 12.521739130434783,
"grad_norm": 0.24020184576511383,
"learning_rate": 8.07830737662829e-05,
"loss": 0.6734,
"step": 36
},
{
"epoch": 12.869565217391305,
"eval_loss": 0.8749791979789734,
"eval_runtime": 13.9424,
"eval_samples_per_second": 0.861,
"eval_steps_per_second": 0.861,
"step": 37
},
{
"epoch": 13.217391304347826,
"grad_norm": 0.29616954922676086,
"learning_rate": 7.795964517353735e-05,
"loss": 0.6722,
"step": 38
},
{
"epoch": 13.91304347826087,
"grad_norm": 0.2673910856246948,
"learning_rate": 7.500000000000001e-05,
"loss": 0.6525,
"step": 40
},
{
"epoch": 13.91304347826087,
"eval_loss": 0.8916842341423035,
"eval_runtime": 13.9404,
"eval_samples_per_second": 0.861,
"eval_steps_per_second": 0.861,
"step": 40
},
{
"epoch": 14.608695652173914,
"grad_norm": 0.3332022428512573,
"learning_rate": 7.191855733945387e-05,
"loss": 0.6044,
"step": 42
},
{
"epoch": 14.956521739130435,
"eval_loss": 0.8961283564567566,
"eval_runtime": 13.9377,
"eval_samples_per_second": 0.861,
"eval_steps_per_second": 0.861,
"step": 43
},
{
"epoch": 15.304347826086957,
"grad_norm": 0.3555317223072052,
"learning_rate": 6.873032967079561e-05,
"loss": 0.5988,
"step": 44
},
{
"epoch": 16.0,
"grad_norm": 0.5534977912902832,
"learning_rate": 6.545084971874738e-05,
"loss": 0.5698,
"step": 46
},
{
"epoch": 16.0,
"eval_loss": 0.922257125377655,
"eval_runtime": 13.9365,
"eval_samples_per_second": 0.861,
"eval_steps_per_second": 0.861,
"step": 46
},
{
"epoch": 16.695652173913043,
"grad_norm": 0.5275772213935852,
"learning_rate": 6.209609477998338e-05,
"loss": 0.5067,
"step": 48
},
{
"epoch": 16.695652173913043,
"eval_loss": 0.9410943388938904,
"eval_runtime": 13.937,
"eval_samples_per_second": 0.861,
"eval_steps_per_second": 0.861,
"step": 48
},
{
"epoch": 17.391304347826086,
"grad_norm": 0.4959109127521515,
"learning_rate": 5.868240888334653e-05,
"loss": 0.51,
"step": 50
},
{
"epoch": 17.73913043478261,
"eval_loss": 0.9558847546577454,
"eval_runtime": 13.954,
"eval_samples_per_second": 0.86,
"eval_steps_per_second": 0.86,
"step": 51
},
{
"epoch": 17.73913043478261,
"step": 51,
"total_flos": 8.237884169099674e+16,
"train_loss": 0.8130270990670896,
"train_runtime": 3067.7429,
"train_samples_per_second": 0.75,
"train_steps_per_second": 0.033
}
],
"logging_steps": 2,
"max_steps": 100,
"num_input_tokens_seen": 0,
"num_train_epochs": 50,
"save_steps": 25,
"stateful_callbacks": {
"EarlyStoppingCallback": {
"args": {
"early_stopping_patience": 7,
"early_stopping_threshold": 0.0
},
"attributes": {
"early_stopping_patience_counter": 0
}
},
"TrainerControl": {
"args": {
"should_epoch_stop": false,
"should_evaluate": false,
"should_log": false,
"should_save": true,
"should_training_stop": true
},
"attributes": {}
}
},
"total_flos": 8.237884169099674e+16,
"train_batch_size": 1,
"trial_name": null,
"trial_params": null
}