grace-discrim-svamp / trainer_state.json
Muhammad Khalifa
first commit
651204f
{
"best_metric": 0.47151947021484375,
"best_model_checkpoint": "ckpts/discrim/contrastive_models/gen.t5.large/flan_t5_large_svamp-clean_100K_max_align_5/checkpoint-2000",
"epoch": 0.5918035212309514,
"global_step": 2000,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.01,
"learning_rate": 2.958579881656805e-06,
"loss": 1.0002,
"step": 50
},
{
"epoch": 0.03,
"learning_rate": 5.91715976331361e-06,
"loss": 1.0001,
"step": 100
},
{
"epoch": 0.04,
"learning_rate": 8.875739644970414e-06,
"loss": 0.9989,
"step": 150
},
{
"epoch": 0.06,
"learning_rate": 1.183431952662722e-05,
"loss": 0.9972,
"step": 200
},
{
"epoch": 0.07,
"learning_rate": 1.4792899408284025e-05,
"loss": 0.995,
"step": 250
},
{
"epoch": 0.09,
"learning_rate": 1.7751479289940828e-05,
"loss": 0.9897,
"step": 300
},
{
"epoch": 0.1,
"learning_rate": 2.071005917159763e-05,
"loss": 0.9809,
"step": 350
},
{
"epoch": 0.12,
"learning_rate": 2.366863905325444e-05,
"loss": 0.8868,
"step": 400
},
{
"epoch": 0.13,
"learning_rate": 2.6627218934911244e-05,
"loss": 0.7408,
"step": 450
},
{
"epoch": 0.15,
"learning_rate": 2.958579881656805e-05,
"loss": 0.674,
"step": 500
},
{
"epoch": 0.16,
"learning_rate": 3.254437869822485e-05,
"loss": 0.6155,
"step": 550
},
{
"epoch": 0.18,
"learning_rate": 3.5502958579881656e-05,
"loss": 0.5576,
"step": 600
},
{
"epoch": 0.19,
"learning_rate": 3.846153846153846e-05,
"loss": 0.5003,
"step": 650
},
{
"epoch": 0.21,
"learning_rate": 4.142011834319526e-05,
"loss": 0.4514,
"step": 700
},
{
"epoch": 0.22,
"learning_rate": 4.4378698224852076e-05,
"loss": 0.4074,
"step": 750
},
{
"epoch": 0.24,
"learning_rate": 4.733727810650888e-05,
"loss": 0.3495,
"step": 800
},
{
"epoch": 0.25,
"learning_rate": 5.029585798816568e-05,
"loss": 0.3617,
"step": 850
},
{
"epoch": 0.27,
"learning_rate": 5.325443786982249e-05,
"loss": 0.3384,
"step": 900
},
{
"epoch": 0.28,
"learning_rate": 5.6213017751479294e-05,
"loss": 0.2962,
"step": 950
},
{
"epoch": 0.3,
"learning_rate": 5.91715976331361e-05,
"loss": 0.2602,
"step": 1000
},
{
"epoch": 0.3,
"eval_loss": 0.6183895468711853,
"eval_runtime": 115.5435,
"eval_samples_per_second": 156.513,
"eval_steps_per_second": 19.568,
"step": 1000
},
{
"epoch": 0.31,
"learning_rate": 5.9999239256099436e-05,
"loss": 0.2554,
"step": 1050
},
{
"epoch": 0.33,
"learning_rate": 5.9995658680583245e-05,
"loss": 0.2386,
"step": 1100
},
{
"epoch": 0.34,
"learning_rate": 5.9989143572558144e-05,
"loss": 0.213,
"step": 1150
},
{
"epoch": 0.36,
"learning_rate": 5.997969456940961e-05,
"loss": 0.2021,
"step": 1200
},
{
"epoch": 0.37,
"learning_rate": 5.9967312595551613e-05,
"loss": 0.1926,
"step": 1250
},
{
"epoch": 0.38,
"learning_rate": 5.9951998862336304e-05,
"loss": 0.1712,
"step": 1300
},
{
"epoch": 0.4,
"learning_rate": 5.9933754867935396e-05,
"loss": 0.1654,
"step": 1350
},
{
"epoch": 0.41,
"learning_rate": 5.991258239719367e-05,
"loss": 0.1424,
"step": 1400
},
{
"epoch": 0.43,
"learning_rate": 5.9888483521454334e-05,
"loss": 0.1452,
"step": 1450
},
{
"epoch": 0.44,
"learning_rate": 5.986146059835637e-05,
"loss": 0.1485,
"step": 1500
},
{
"epoch": 0.46,
"learning_rate": 5.983151627160388e-05,
"loss": 0.1182,
"step": 1550
},
{
"epoch": 0.47,
"learning_rate": 5.9798653470707503e-05,
"loss": 0.1215,
"step": 1600
},
{
"epoch": 0.49,
"learning_rate": 5.976287541069771e-05,
"loss": 0.1183,
"step": 1650
},
{
"epoch": 0.5,
"learning_rate": 5.972418559181037e-05,
"loss": 0.0983,
"step": 1700
},
{
"epoch": 0.52,
"learning_rate": 5.968258779914428e-05,
"loss": 0.1155,
"step": 1750
},
{
"epoch": 0.53,
"learning_rate": 5.9638086102290845e-05,
"loss": 0.0874,
"step": 1800
},
{
"epoch": 0.55,
"learning_rate": 5.9590684854935956e-05,
"loss": 0.0993,
"step": 1850
},
{
"epoch": 0.56,
"learning_rate": 5.9540388694434064e-05,
"loss": 0.088,
"step": 1900
},
{
"epoch": 0.58,
"learning_rate": 5.948720254135451e-05,
"loss": 0.0849,
"step": 1950
},
{
"epoch": 0.59,
"learning_rate": 5.943113159900009e-05,
"loss": 0.0864,
"step": 2000
},
{
"epoch": 0.59,
"eval_loss": 0.47151947021484375,
"eval_runtime": 115.9067,
"eval_samples_per_second": 156.022,
"eval_steps_per_second": 19.507,
"step": 2000
}
],
"max_steps": 16895,
"num_train_epochs": 5,
"total_flos": 0.0,
"trial_name": null,
"trial_params": null
}