the-jb/tofu_Llama-3.2-3B-Instruct_forget10_NPO
1
0
Fork 0
Code Issues Pull Requests Actions Projects Releases Wiki Activity

初始化项目，由ModelHub XC社区提供模型

Browse Source 
Model: the-jb/tofu_Llama-3.2-3B-Instruct_forget10_NPO
Source: Original Platform
This commit is contained in:
ModelHub XC
2026-05-12 20:00:56 +08:00
commit 0482ff20b7
23 changed files with 42748 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

211
trainer_state.json Normal file
View File

@@ -0,0 +1,211 @@
{
"best_global_step": null,
"best_metric": null,
"best_model_checkpoint": null,
"epoch": 9.24,
"eval_steps": 500,
"global_step": 120,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.4,
"grad_norm": 496.3458558718414,
"learning_rate": 3.3333333333333333e-06,
"loss": 55.5017,
"step": 5
},
{
"epoch": 0.8,
"grad_norm": 554.6838933038925,
"learning_rate": 7.500000000000001e-06,
"loss": 51.4302,
"step": 10
},
{
"epoch": 1.16,
"grad_norm": 183.3033225111087,
"learning_rate": 9.814814814814815e-06,
"loss": 19.7484,
"step": 15
},
{
"epoch": 1.56,
"grad_norm": 42.17902341744105,
"learning_rate": 9.351851851851854e-06,
"loss": 8.5018,
"step": 20
},
{
"epoch": 1.96,
"grad_norm": 52.424248344875025,
"learning_rate": 8.888888888888888e-06,
"loss": 7.8758,
"step": 25
},
{
"epoch": 2.32,
"grad_norm": 42.07252256082996,
"learning_rate": 8.425925925925926e-06,
"loss": 6.9702,
"step": 30
},
{
"epoch": 2.7199999999999998,
"grad_norm": 40.05814690503554,
"learning_rate": 7.962962962962963e-06,
"loss": 7.5265,
"step": 35
},
{
"epoch": 3.08,
"grad_norm": 37.47864647047457,
"learning_rate": 7.500000000000001e-06,
"loss": 5.9698,
"step": 40
},
{
"epoch": 3.48,
"grad_norm": 35.813336917273894,
"learning_rate": 7.0370370370370375e-06,
"loss": 6.4087,
"step": 45
},
{
"epoch": 3.88,
"grad_norm": 27.282894127394396,
"learning_rate": 6.574074074074075e-06,
"loss": 5.5071,
"step": 50
},
{
"epoch": 4.24,
"grad_norm": 27.802578760956333,
"learning_rate": 6.111111111111112e-06,
"loss": 5.0763,
"step": 55
},
{
"epoch": 4.64,
"grad_norm": 29.641222280470267,
"learning_rate": 5.6481481481481485e-06,
"loss": 5.082,
"step": 60
},
{
"epoch": 5.0,
"grad_norm": 31.083024955586065,
"learning_rate": 5.185185185185185e-06,
"loss": 4.0623,
"step": 65
},
{
"epoch": 5.4,
"grad_norm": 30.714787625937127,
"learning_rate": 4.722222222222222e-06,
"loss": 4.3391,
"step": 70
},
{
"epoch": 5.8,
"grad_norm": 28.346550459414253,
"learning_rate": 4.2592592592592596e-06,
"loss": 4.5051,
"step": 75
},
{
"epoch": 6.16,
"grad_norm": 25.631675639275336,
"learning_rate": 3.796296296296297e-06,
"loss": 3.7155,
"step": 80
},
{
"epoch": 6.5600000000000005,
"grad_norm": 28.27021817073457,
"learning_rate": 3.3333333333333333e-06,
"loss": 3.7265,
"step": 85
},
{
"epoch": 6.96,
"grad_norm": 28.031857316121684,
"learning_rate": 2.8703703703703706e-06,
"loss": 3.8315,
"step": 90
},
{
"epoch": 7.32,
"grad_norm": 28.86130903236077,
"learning_rate": 2.4074074074074075e-06,
"loss": 3.2732,
"step": 95
},
{
"epoch": 7.72,
"grad_norm": 27.21813757431694,
"learning_rate": 1.944444444444445e-06,
"loss": 3.8326,
"step": 100
},
{
"epoch": 8.08,
"grad_norm": 24.550229160966467,
"learning_rate": 1.4814814814814815e-06,
"loss": 3.1823,
"step": 105
},
{
"epoch": 8.48,
"grad_norm": 26.71144192851546,
"learning_rate": 1.0185185185185185e-06,
"loss": 3.487,
"step": 110
},
{
"epoch": 8.88,
"grad_norm": 27.277199175542826,
"learning_rate": 5.555555555555555e-07,
"loss": 3.4419,
"step": 115
},
{
"epoch": 9.24,
"grad_norm": 30.168350611828913,
"learning_rate": 9.259259259259259e-08,
"loss": 3.0733,
"step": 120
},
{
"epoch": 9.24,
"step": 120,
"total_flos": 0.0,
"train_loss": 9.586206404368083,
"train_runtime": 2505.9007,
"train_samples_per_second": 1.596,
"train_steps_per_second": 0.048
}
],
"logging_steps": 5,
"max_steps": 120,
"num_input_tokens_seen": 0,
"num_train_epochs": 10,
"save_steps": 500,
"stateful_callbacks": {
"TrainerControl": {
"args": {
"should_epoch_stop": false,
"should_evaluate": false,
"should_log": false,
"should_save": false,
"should_training_stop": false
},
"attributes": {}
}
},
"total_flos": 0.0,
"train_batch_size": 4,
"trial_name": null,
"trial_params": null
}