Pathfinderrr_20250501 / trainer_state.json
Mathieu-Thomas-JOSSET's picture
Upload fine-tuned model (lora) based on Qwen/Qwen3-0.6B from run 20250501_212414
92ec2fc verified
{
"best_global_step": null,
"best_metric": null,
"best_model_checkpoint": null,
"epoch": 2.64,
"eval_steps": 500,
"global_step": 18,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.16,
"grad_norm": 3.242586612701416,
"learning_rate": 5e-05,
"loss": 3.5477,
"step": 1
},
{
"epoch": 0.32,
"grad_norm": 2.756465435028076,
"learning_rate": 4.962019382530521e-05,
"loss": 3.4875,
"step": 2
},
{
"epoch": 0.48,
"grad_norm": 2.6460156440734863,
"learning_rate": 4.849231551964771e-05,
"loss": 3.5183,
"step": 3
},
{
"epoch": 0.64,
"grad_norm": 2.6053574085235596,
"learning_rate": 4.665063509461097e-05,
"loss": 3.4119,
"step": 4
},
{
"epoch": 0.8,
"grad_norm": 2.42952823638916,
"learning_rate": 4.415111107797445e-05,
"loss": 3.3665,
"step": 5
},
{
"epoch": 0.96,
"grad_norm": 2.268110990524292,
"learning_rate": 4.1069690242163484e-05,
"loss": 3.3536,
"step": 6
},
{
"epoch": 1.0,
"grad_norm": 3.588498115539551,
"learning_rate": 3.7500000000000003e-05,
"loss": 3.7349,
"step": 7
},
{
"epoch": 1.16,
"grad_norm": 2.3541579246520996,
"learning_rate": 3.355050358314172e-05,
"loss": 3.2116,
"step": 8
},
{
"epoch": 1.32,
"grad_norm": 2.076106071472168,
"learning_rate": 2.9341204441673266e-05,
"loss": 3.1976,
"step": 9
},
{
"epoch": 1.48,
"grad_norm": 2.280040740966797,
"learning_rate": 2.5e-05,
"loss": 3.2582,
"step": 10
},
{
"epoch": 1.6400000000000001,
"grad_norm": 2.1664576530456543,
"learning_rate": 2.0658795558326743e-05,
"loss": 3.0772,
"step": 11
},
{
"epoch": 1.8,
"grad_norm": 2.567535400390625,
"learning_rate": 1.6449496416858284e-05,
"loss": 3.1218,
"step": 12
},
{
"epoch": 1.96,
"grad_norm": 2.1526150703430176,
"learning_rate": 1.2500000000000006e-05,
"loss": 3.1159,
"step": 13
},
{
"epoch": 2.0,
"grad_norm": 3.201275587081909,
"learning_rate": 8.930309757836517e-06,
"loss": 3.0268,
"step": 14
},
{
"epoch": 2.16,
"grad_norm": 2.2598965167999268,
"learning_rate": 5.848888922025553e-06,
"loss": 3.1547,
"step": 15
},
{
"epoch": 2.32,
"grad_norm": 2.4263217449188232,
"learning_rate": 3.3493649053890326e-06,
"loss": 3.2911,
"step": 16
},
{
"epoch": 2.48,
"grad_norm": 2.0498416423797607,
"learning_rate": 1.5076844803522922e-06,
"loss": 2.9848,
"step": 17
},
{
"epoch": 2.64,
"grad_norm": 1.9774571657180786,
"learning_rate": 3.7980617469479953e-07,
"loss": 2.9408,
"step": 18
},
{
"epoch": 2.64,
"step": 18,
"total_flos": 52264106459136.0,
"train_loss": 3.2667255534066095,
"train_runtime": 27.387,
"train_samples_per_second": 5.477,
"train_steps_per_second": 0.657
}
],
"logging_steps": 1,
"max_steps": 18,
"num_input_tokens_seen": 0,
"num_train_epochs": 3,
"save_steps": 500,
"stateful_callbacks": {
"TrainerControl": {
"args": {
"should_epoch_stop": false,
"should_evaluate": false,
"should_log": false,
"should_save": true,
"should_training_stop": true
},
"attributes": {}
}
},
"total_flos": 52264106459136.0,
"train_batch_size": 2,
"trial_name": null,
"trial_params": null
}