sugiv/qwen3-8b-tanglish
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: sugiv/qwen3-8b-tanglish
Source: Original Platform
This commit is contained in:
ModelHub XC
2026-09-08 15:34:20 +08:00
commit 5734ffb7af
437 changed files with 4525344 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

565
checkpoints/checkpoint-1000/trainer_state.json Normal file
View File

@@ -0,0 +1,565 @@
{
"best_global_step": 1000,
"best_metric": 1.059064507484436,
"best_model_checkpoint": "/workspace/tanglish/checkpoints/qwen_lora_v1/lora/checkpoint-1000",
"epoch": 0.20725388601036268,
"eval_steps": 500,
"global_step": 1000,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"entropy": 1.4381825268268584,
"epoch": 0.004145077720207254,
"grad_norm": 2.1717000007629395,
"learning_rate": 5.248618784530387e-06,
"loss": 4.0762,
"mean_token_accuracy": 0.4441531464457512,
"num_tokens": 51480.0,
"step": 20
},
{
"entropy": 1.5037598952651023,
"epoch": 0.008290155440414507,
"grad_norm": 2.724255323410034,
"learning_rate": 1.0773480662983426e-05,
"loss": 3.8127,
"mean_token_accuracy": 0.4525611069053411,
"num_tokens": 103460.0,
"step": 40
},
{
"entropy": 1.901321442425251,
"epoch": 0.012435233160621761,
"grad_norm": 1.1939035654067993,
"learning_rate": 1.6298342541436465e-05,
"loss": 3.1292,
"mean_token_accuracy": 0.48335910513997077,
"num_tokens": 154920.0,
"step": 60
},
{
"entropy": 2.358680176734924,
"epoch": 0.016580310880829015,
"grad_norm": 0.7893019318580627,
"learning_rate": 2.1823204419889505e-05,
"loss": 2.3084,
"mean_token_accuracy": 0.554056815057993,
"num_tokens": 207786.0,
"step": 80
},
{
"entropy": 1.830135741829872,
"epoch": 0.02072538860103627,
"grad_norm": 0.5736128091812134,
"learning_rate": 2.7348066298342545e-05,
"loss": 1.708,
"mean_token_accuracy": 0.6866068348288537,
"num_tokens": 256847.0,
"step": 100
},
{
"entropy": 1.4562795095145702,
"epoch": 0.024870466321243522,
"grad_norm": 0.9956451058387756,
"learning_rate": 3.287292817679558e-05,
"loss": 1.4132,
"mean_token_accuracy": 0.7327493987977505,
"num_tokens": 309233.0,
"step": 120
},
{
"entropy": 1.4073016680777073,
"epoch": 0.029015544041450778,
"grad_norm": 0.8126519918441772,
"learning_rate": 3.8397790055248625e-05,
"loss": 1.3879,
"mean_token_accuracy": 0.7353364005684853,
"num_tokens": 361206.0,
"step": 140
},
{
"entropy": 1.28325657248497,
"epoch": 0.03316062176165803,
"grad_norm": 0.6408252120018005,
"learning_rate": 4.392265193370166e-05,
"loss": 1.2517,
"mean_token_accuracy": 0.7543429806828499,
"num_tokens": 412093.0,
"step": 160
},
{
"entropy": 1.219002117216587,
"epoch": 0.03730569948186528,
"grad_norm": 0.6377286911010742,
"learning_rate": 4.94475138121547e-05,
"loss": 1.2238,
"mean_token_accuracy": 0.7544910028576851,
"num_tokens": 463753.0,
"step": 180
},
{
"entropy": 1.2931843183934688,
"epoch": 0.04145077720207254,
"grad_norm": 0.7240749001502991,
"learning_rate": 5.4972375690607735e-05,
"loss": 1.3026,
"mean_token_accuracy": 0.7400020457804203,
"num_tokens": 515356.0,
"step": 200
},
{
"entropy": 1.2196356683969498,
"epoch": 0.04559585492227979,
"grad_norm": 0.7647127509117126,
"learning_rate": 6.049723756906077e-05,
"loss": 1.2411,
"mean_token_accuracy": 0.7544264450669289,
"num_tokens": 565236.0,
"step": 220
},
{
"entropy": 1.171409723162651,
"epoch": 0.049740932642487044,
"grad_norm": 0.7827390432357788,
"learning_rate": 6.602209944751382e-05,
"loss": 1.1776,
"mean_token_accuracy": 0.7626889519393444,
"num_tokens": 615515.0,
"step": 240
},
{
"entropy": 1.1406699948012828,
"epoch": 0.0538860103626943,
"grad_norm": 1.0029950141906738,
"learning_rate": 7.154696132596685e-05,
"loss": 1.1546,
"mean_token_accuracy": 0.766391609609127,
"num_tokens": 666062.0,
"step": 260
},
{
"entropy": 1.1219736263155937,
"epoch": 0.058031088082901555,
"grad_norm": 1.0503102540969849,
"learning_rate": 7.70718232044199e-05,
"loss": 1.1352,
"mean_token_accuracy": 0.7641930885612964,
"num_tokens": 719494.0,
"step": 280
},
{
"entropy": 1.1847783662378788,
"epoch": 0.06217616580310881,
"grad_norm": 0.8290141820907593,
"learning_rate": 8.259668508287292e-05,
"loss": 1.202,
"mean_token_accuracy": 0.7585213594138622,
"num_tokens": 772839.0,
"step": 300
},
{
"entropy": 1.1723797246813774,
"epoch": 0.06632124352331606,
"grad_norm": 0.877815842628479,
"learning_rate": 8.812154696132597e-05,
"loss": 1.1825,
"mean_token_accuracy": 0.7599225588142872,
"num_tokens": 825189.0,
"step": 320
},
{
"entropy": 1.1410918802022934,
"epoch": 0.07046632124352331,
"grad_norm": 0.7032957673072815,
"learning_rate": 9.364640883977901e-05,
"loss": 1.1333,
"mean_token_accuracy": 0.7682646617293358,
"num_tokens": 877294.0,
"step": 340
},
{
"entropy": 1.1239484950900078,
"epoch": 0.07461139896373056,
"grad_norm": 0.8480610847473145,
"learning_rate": 9.917127071823204e-05,
"loss": 1.1536,
"mean_token_accuracy": 0.7669501408934594,
"num_tokens": 927234.0,
"step": 360
},
{
"entropy": 1.110478015244007,
"epoch": 0.07875647668393783,
"grad_norm": 0.7844213843345642,
"learning_rate": 0.0001046961325966851,
"loss": 1.1255,
"mean_token_accuracy": 0.7683746472001076,
"num_tokens": 980398.0,
"step": 380
},
{
"entropy": 1.1618236370384694,
"epoch": 0.08290155440414508,
"grad_norm": 0.7246690988540649,
"learning_rate": 0.00011022099447513811,
"loss": 1.1729,
"mean_token_accuracy": 0.7629252031445504,
"num_tokens": 1030335.0,
"step": 400
},
{
"entropy": 1.1364129185676575,
"epoch": 0.08704663212435233,
"grad_norm": 0.7912014722824097,
"learning_rate": 0.00011574585635359116,
"loss": 1.152,
"mean_token_accuracy": 0.7651066802442074,
"num_tokens": 1083694.0,
"step": 420
},
{
"entropy": 1.0954272992908956,
"epoch": 0.09119170984455958,
"grad_norm": 0.7310007214546204,
"learning_rate": 0.00012127071823204422,
"loss": 1.1039,
"mean_token_accuracy": 0.7700782015919685,
"num_tokens": 1134346.0,
"step": 440
},
{
"entropy": 1.1179637335240842,
"epoch": 0.09533678756476684,
"grad_norm": 0.7415347099304199,
"learning_rate": 0.00012679558011049723,
"loss": 1.1267,
"mean_token_accuracy": 0.7677439145743847,
"num_tokens": 1188485.0,
"step": 460
},
{
"entropy": 1.0841590076684953,
"epoch": 0.09948186528497409,
"grad_norm": 0.8192847371101379,
"learning_rate": 0.00013232044198895027,
"loss": 1.095,
"mean_token_accuracy": 0.7726523540914059,
"num_tokens": 1239371.0,
"step": 480
},
{
"entropy": 1.0643972344696522,
"epoch": 0.10362694300518134,
"grad_norm": 0.8887308239936829,
"learning_rate": 0.00013784530386740332,
"loss": 1.0762,
"mean_token_accuracy": 0.7769059598445892,
"num_tokens": 1290491.0,
"step": 500
},
{
"epoch": 0.10362694300518134,
"eval_entropy": 1.095322513439524,
"eval_loss": 1.1184215545654297,
"eval_mean_token_accuracy": 0.7680243507144958,
"eval_num_tokens": 1290491.0,
"eval_runtime": 198.0973,
"eval_samples_per_second": 20.51,
"eval_steps_per_second": 2.564,
"step": 500
},
{
"entropy": 1.0820845410227775,
"epoch": 0.1077720207253886,
"grad_norm": 0.7347564697265625,
"learning_rate": 0.00014337016574585636,
"loss": 1.0979,
"mean_token_accuracy": 0.7743508994579316,
"num_tokens": 1341205.0,
"step": 520
},
{
"entropy": 1.1001912213861942,
"epoch": 0.11191709844559586,
"grad_norm": 0.690479040145874,
"learning_rate": 0.0001488950276243094,
"loss": 1.1105,
"mean_token_accuracy": 0.7694470949470997,
"num_tokens": 1394461.0,
"step": 540
},
{
"entropy": 1.0999469496309757,
"epoch": 0.11606217616580311,
"grad_norm": 0.6917790174484253,
"learning_rate": 0.00015441988950276245,
"loss": 1.117,
"mean_token_accuracy": 0.7682196438312531,
"num_tokens": 1447828.0,
"step": 560
},
{
"entropy": 1.0735741600394249,
"epoch": 0.12020725388601036,
"grad_norm": 0.833886444568634,
"learning_rate": 0.00015994475138121547,
"loss": 1.086,
"mean_token_accuracy": 0.7741858348250389,
"num_tokens": 1498479.0,
"step": 580
},
{
"entropy": 1.0962664999067784,
"epoch": 0.12435233160621761,
"grad_norm": 0.6647179126739502,
"learning_rate": 0.0001654696132596685,
"loss": 1.0958,
"mean_token_accuracy": 0.7741696208715438,
"num_tokens": 1548137.0,
"step": 600
},
{
"entropy": 1.0694029107689857,
"epoch": 0.12849740932642487,
"grad_norm": 0.6861886382102966,
"learning_rate": 0.00017099447513812155,
"loss": 1.0829,
"mean_token_accuracy": 0.7720654770731926,
"num_tokens": 1599638.0,
"step": 620
},
{
"entropy": 1.075825183838606,
"epoch": 0.13264248704663212,
"grad_norm": 0.7000752687454224,
"learning_rate": 0.0001765193370165746,
"loss": 1.0825,
"mean_token_accuracy": 0.7762658178806305,
"num_tokens": 1650323.0,
"step": 640
},
{
"entropy": 1.083068310469389,
"epoch": 0.13678756476683937,
"grad_norm": 0.7568516731262207,
"learning_rate": 0.00018204419889502764,
"loss": 1.0918,
"mean_token_accuracy": 0.7720131345093251,
"num_tokens": 1701215.0,
"step": 660
},
{
"entropy": 1.0660202227532865,
"epoch": 0.14093264248704662,
"grad_norm": 0.8106701374053955,
"learning_rate": 0.00018756906077348066,
"loss": 1.073,
"mean_token_accuracy": 0.7741113729774952,
"num_tokens": 1751275.0,
"step": 680
},
{
"entropy": 1.00438876375556,
"epoch": 0.14507772020725387,
"grad_norm": 0.6197234988212585,
"learning_rate": 0.0001930939226519337,
"loss": 1.0234,
"mean_token_accuracy": 0.7818360298871994,
"num_tokens": 1803256.0,
"step": 700
},
{
"entropy": 1.1083124563097955,
"epoch": 0.14922279792746113,
"grad_norm": 0.6046066284179688,
"learning_rate": 0.00019861878453038677,
"loss": 1.122,
"mean_token_accuracy": 0.7704763405025006,
"num_tokens": 1855372.0,
"step": 720
},
{
"entropy": 1.097743584215641,
"epoch": 0.1533678756476684,
"grad_norm": 0.6518492102622986,
"learning_rate": 0.00019999941280374287,
"loss": 1.1116,
"mean_token_accuracy": 0.7682732604444027,
"num_tokens": 1906015.0,
"step": 740
},
{
"entropy": 1.1098480015993117,
"epoch": 0.15751295336787566,
"grad_norm": 0.6173190474510193,
"learning_rate": 0.00019999680305650554,
"loss": 1.1173,
"mean_token_accuracy": 0.7627138867974281,
"num_tokens": 1959638.0,
"step": 760
},
{
"entropy": 1.0928526185452938,
"epoch": 0.1616580310880829,
"grad_norm": 0.6099253296852112,
"learning_rate": 0.00019999210556868917,
"loss": 1.1156,
"mean_token_accuracy": 0.7711292572319508,
"num_tokens": 2010476.0,
"step": 780
},
{
"entropy": 1.0808037273585795,
"epoch": 0.16580310880829016,
"grad_norm": 0.6099389791488647,
"learning_rate": 0.0001999853204383683,
"loss": 1.087,
"mean_token_accuracy": 0.7749647930264473,
"num_tokens": 2060494.0,
"step": 800
},
{
"entropy": 1.048035841435194,
"epoch": 0.1699481865284974,
"grad_norm": 0.5170192122459412,
"learning_rate": 0.0001999764478072033,
"loss": 1.0642,
"mean_token_accuracy": 0.7727734588086606,
"num_tokens": 2113260.0,
"step": 820
},
{
"entropy": 1.0665065109729768,
"epoch": 0.17409326424870467,
"grad_norm": 0.6461181640625,
"learning_rate": 0.00019996548786043768,
"loss": 1.0795,
"mean_token_accuracy": 0.777741952240467,
"num_tokens": 2164157.0,
"step": 840
},
{
"entropy": 1.0095063984394073,
"epoch": 0.17823834196891192,
"grad_norm": 0.5376838445663452,
"learning_rate": 0.00019995244082689396,
"loss": 1.0238,
"mean_token_accuracy": 0.7842500291764736,
"num_tokens": 2215501.0,
"step": 860
},
{
"entropy": 1.0482122428715228,
"epoch": 0.18238341968911917,
"grad_norm": 0.5362586379051208,
"learning_rate": 0.00019993730697896912,
"loss": 1.0567,
"mean_token_accuracy": 0.7757732585072518,
"num_tokens": 2265908.0,
"step": 880
},
{
"entropy": 1.0375815689563752,
"epoch": 0.18652849740932642,
"grad_norm": 0.6826100945472717,
"learning_rate": 0.0001999200866326287,
"loss": 1.055,
"mean_token_accuracy": 0.777643445879221,
"num_tokens": 2316578.0,
"step": 900
},
{
"entropy": 1.0592326551675797,
"epoch": 0.19067357512953367,
"grad_norm": 0.4966115355491638,
"learning_rate": 0.00019990078014740036,
"loss": 1.0682,
"mean_token_accuracy": 0.7757245123386383,
"num_tokens": 2367491.0,
"step": 920
},
{
"entropy": 1.0522546865046025,
"epoch": 0.19481865284974093,
"grad_norm": 0.5403271317481995,
"learning_rate": 0.00019987938792636632,
"loss": 1.0597,
"mean_token_accuracy": 0.7824762858450413,
"num_tokens": 2418531.0,
"step": 940
},
{
"entropy": 1.0112147249281407,
"epoch": 0.19896373056994818,
"grad_norm": 0.591804027557373,
"learning_rate": 0.00019985591041615496,
"loss": 1.0179,
"mean_token_accuracy": 0.7824166387319564,
"num_tokens": 2468312.0,
"step": 960
},
{
"entropy": 1.0457121133804321,
"epoch": 0.20310880829015543,
"grad_norm": 0.4792396128177643,
"learning_rate": 0.00019983034810693137,
"loss": 1.0567,
"mean_token_accuracy": 0.7786681324243545,
"num_tokens": 2518124.0,
"step": 980
},
{
"entropy": 1.0062039978802204,
"epoch": 0.20725388601036268,
"grad_norm": 0.6134971976280212,
"learning_rate": 0.00019980270153238742,
"loss": 1.0284,
"mean_token_accuracy": 0.7829495541751385,
"num_tokens": 2569706.0,
"step": 1000
},
{
"epoch": 0.20725388601036268,
"eval_entropy": 1.0445965941496722,
"eval_loss": 1.059064507484436,
"eval_mean_token_accuracy": 0.7763740899055962,
"eval_num_tokens": 2569706.0,
"eval_runtime": 197.7332,
"eval_samples_per_second": 20.548,
"eval_steps_per_second": 2.569,
"step": 1000
}
],
"logging_steps": 20,
"max_steps": 14475,
"num_input_tokens_seen": 0,
"num_train_epochs": 3,
"save_steps": 500,
"stateful_callbacks": {
"EarlyStoppingCallback": {
"args": {
"early_stopping_patience": 4,
"early_stopping_threshold": 0.005
},
"attributes": {
"early_stopping_patience_counter": 0
}
},
"TrainerControl": {
"args": {
"should_epoch_stop": false,
"should_evaluate": false,
"should_log": false,
"should_save": true,
"should_training_stop": false
},
"attributes": {}
}
},
"total_flos": 1.773757424068608e+17,
"train_batch_size": 4,
"trial_name": null,
"trial_params": null
}