SLM_vs_LLM_experiments
/
max_seq_length_128_experiments
/Qwen
/Qwen1.5_1.8B_twitter
/checkpoint-250
/trainer_state.json
| { | |
| "best_metric": 0.5038631558418274, | |
| "best_model_checkpoint": "../../experiments_checkpoints/MAdAiLab/Qwen/Qwen1.5_1.8B_twitter/checkpoint-250", | |
| "epoch": 0.9191176470588235, | |
| "eval_steps": 50, | |
| "global_step": 250, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.04, | |
| "grad_norm": 109.95573425292969, | |
| "learning_rate": 4.938725490196079e-06, | |
| "loss": 1.1177, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.07, | |
| "grad_norm": 117.97087860107422, | |
| "learning_rate": 4.8774509803921576e-06, | |
| "loss": 0.8448, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.11, | |
| "grad_norm": 46.05353546142578, | |
| "learning_rate": 4.816176470588236e-06, | |
| "loss": 0.8301, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.15, | |
| "grad_norm": 41.749080657958984, | |
| "learning_rate": 4.754901960784314e-06, | |
| "loss": 0.7069, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.18, | |
| "grad_norm": 26.182971954345703, | |
| "learning_rate": 4.693627450980393e-06, | |
| "loss": 0.6585, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.18, | |
| "eval_accuracy": 0.7123161764705882, | |
| "eval_f1_macro": 0.5810560943233977, | |
| "eval_f1_micro": 0.7123161764705882, | |
| "eval_loss": 0.6434972286224365, | |
| "eval_runtime": 2.8155, | |
| "eval_samples_per_second": 386.429, | |
| "eval_steps_per_second": 12.076, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.22, | |
| "grad_norm": 118.7502212524414, | |
| "learning_rate": 4.632352941176471e-06, | |
| "loss": 0.6179, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.26, | |
| "grad_norm": 34.24118423461914, | |
| "learning_rate": 4.571078431372549e-06, | |
| "loss": 0.6324, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.29, | |
| "grad_norm": 92.65385437011719, | |
| "learning_rate": 4.509803921568628e-06, | |
| "loss": 0.5743, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.33, | |
| "grad_norm": 44.45353317260742, | |
| "learning_rate": 4.448529411764706e-06, | |
| "loss": 0.4798, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.37, | |
| "grad_norm": 35.91661071777344, | |
| "learning_rate": 4.3872549019607845e-06, | |
| "loss": 0.6396, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.37, | |
| "eval_accuracy": 0.7297794117647058, | |
| "eval_f1_macro": 0.6997657235537935, | |
| "eval_f1_micro": 0.7297794117647058, | |
| "eval_loss": 0.6015912294387817, | |
| "eval_runtime": 2.832, | |
| "eval_samples_per_second": 384.18, | |
| "eval_steps_per_second": 12.006, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.4, | |
| "grad_norm": 56.78458786010742, | |
| "learning_rate": 4.3259803921568635e-06, | |
| "loss": 0.5814, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.44, | |
| "grad_norm": 53.924827575683594, | |
| "learning_rate": 4.264705882352942e-06, | |
| "loss": 0.5131, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.48, | |
| "grad_norm": 74.93573760986328, | |
| "learning_rate": 4.20343137254902e-06, | |
| "loss": 0.5242, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.51, | |
| "grad_norm": 26.097169876098633, | |
| "learning_rate": 4.142156862745099e-06, | |
| "loss": 0.5529, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.55, | |
| "grad_norm": 78.16165161132812, | |
| "learning_rate": 4.080882352941177e-06, | |
| "loss": 0.5108, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.55, | |
| "eval_accuracy": 0.7527573529411765, | |
| "eval_f1_macro": 0.6963299829940972, | |
| "eval_f1_micro": 0.7527573529411765, | |
| "eval_loss": 0.5226907134056091, | |
| "eval_runtime": 2.826, | |
| "eval_samples_per_second": 384.993, | |
| "eval_steps_per_second": 12.031, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.59, | |
| "grad_norm": 19.04606056213379, | |
| "learning_rate": 4.019607843137255e-06, | |
| "loss": 0.5012, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.62, | |
| "grad_norm": 127.43460845947266, | |
| "learning_rate": 3.958333333333333e-06, | |
| "loss": 0.5601, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.66, | |
| "grad_norm": 73.45539093017578, | |
| "learning_rate": 3.897058823529412e-06, | |
| "loss": 0.6134, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.7, | |
| "grad_norm": 59.42979049682617, | |
| "learning_rate": 3.8357843137254904e-06, | |
| "loss": 0.5447, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.74, | |
| "grad_norm": 22.45537757873535, | |
| "learning_rate": 3.774509803921569e-06, | |
| "loss": 0.5065, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.74, | |
| "eval_accuracy": 0.7417279411764706, | |
| "eval_f1_macro": 0.6346563132227484, | |
| "eval_f1_micro": 0.7417279411764706, | |
| "eval_loss": 0.5502642393112183, | |
| "eval_runtime": 2.8852, | |
| "eval_samples_per_second": 377.094, | |
| "eval_steps_per_second": 11.784, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.77, | |
| "grad_norm": 50.938880920410156, | |
| "learning_rate": 3.7132352941176476e-06, | |
| "loss": 0.4589, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.81, | |
| "grad_norm": 84.68132781982422, | |
| "learning_rate": 3.6519607843137257e-06, | |
| "loss": 0.5403, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.85, | |
| "grad_norm": 27.052024841308594, | |
| "learning_rate": 3.5906862745098043e-06, | |
| "loss": 0.4618, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.88, | |
| "grad_norm": 23.497787475585938, | |
| "learning_rate": 3.529411764705883e-06, | |
| "loss": 0.452, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.92, | |
| "grad_norm": 15.185086250305176, | |
| "learning_rate": 3.468137254901961e-06, | |
| "loss": 0.4883, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.92, | |
| "eval_accuracy": 0.7775735294117647, | |
| "eval_f1_macro": 0.7420002194942226, | |
| "eval_f1_micro": 0.7775735294117647, | |
| "eval_loss": 0.5038631558418274, | |
| "eval_runtime": 2.8371, | |
| "eval_samples_per_second": 383.494, | |
| "eval_steps_per_second": 11.984, | |
| "step": 250 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 816, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 3, | |
| "save_steps": 50, | |
| "total_flos": 7461905891328000.0, | |
| "train_batch_size": 16, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |