Model save

README.md

@@ -27,7 +27,7 @@ print(output["generated_text"])
 
 ## Training procedure
 
-[<img src="https://raw.githubusercontent.com/wandb/assets/main/wandb-github-badge-28.svg" alt="Visualize in Weights & Biases" width="150" height="24"/>](https://wandb.ai/felipealumni-usp/huggingface/runs/14hhk8v2)
+[<img src="https://raw.githubusercontent.com/wandb/assets/main/wandb-github-badge-28.svg" alt="Visualize in Weights & Biases" width="150" height="24"/>](https://wandb.ai/felipealumni-usp/huggingface/runs/j5fc5zrz)
 
 This model was trained with SFT.
 

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:fa6314d207d082c24a0086630dde03167d9e0a9590cf9bf2e427b3b4e8752117
+oid sha256:a76f4c4267621f6756d7efc35412a52e1c431c8020c37ec87679dbaf899e34be
 size 2185294624

all_results.json

@@ -0,0 +1,10 @@
+{
+    "epoch": 0.9992542878448919,
+    "eval_samples": 26,
+    "total_flos": 7.44129634982953e+16,
+    "train_loss": 0.2547340336130626,
+    "train_runtime": 9634.561,
+    "train_samples": 25295,
+    "train_samples_per_second": 0.556,
+    "train_steps_per_second": 0.07
+}

train_results.json

@@ -0,0 +1,10 @@
+{
+    "epoch": 0.9992542878448919,
+    "eval_samples": 26,
+    "total_flos": 7.44129634982953e+16,
+    "train_loss": 0.2547340336130626,
+    "train_runtime": 9634.561,
+    "train_samples": 25295,
+    "train_samples_per_second": 0.556,
+    "train_steps_per_second": 0.07
+}

trainer_state.json

@@ -0,0 +1,432 @@
+{
+  "best_metric": 0.4858,
+  "best_model_checkpoint": "runs/cocoruta2-llama3-1-8b-regex-only-valid/checkpoint-350",
+  "epoch": 0.9992542878448919,
+  "eval_steps": 25,
+  "global_step": 670,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.037285607755406416,
+      "grad_norm": 0.5538629931679852,
+      "learning_rate": 7.46268656716418e-05,
+      "loss": 0.8987,
+      "step": 25
+    },
+    {
+      "epoch": 0.037285607755406416,
+      "eval_loss": 0.6527890563011169,
+      "eval_runtime": 4.7198,
+      "eval_samples_per_second": 1.059,
+      "eval_steps_per_second": 0.636,
+      "step": 25
+    },
+    {
+      "epoch": 0.07457121551081283,
+      "grad_norm": 0.2341721772038344,
+      "learning_rate": 0.0001492537313432836,
+      "loss": 0.6394,
+      "step": 50
+    },
+    {
+      "epoch": 0.07457121551081283,
+      "eval_loss": 0.5486425757408142,
+      "eval_runtime": 4.7295,
+      "eval_samples_per_second": 1.057,
+      "eval_steps_per_second": 0.634,
+      "step": 50
+    },
+    {
+      "epoch": 0.11185682326621924,
+      "grad_norm": 0.19912920431393782,
+      "learning_rate": 0.00019991315351855748,
+      "loss": 0.6007,
+      "step": 75
+    },
+    {
+      "epoch": 0.11185682326621924,
+      "eval_loss": 0.5331323146820068,
+      "eval_runtime": 4.6724,
+      "eval_samples_per_second": 1.07,
+      "eval_steps_per_second": 0.642,
+      "step": 75
+    },
+    {
+      "epoch": 0.14914243102162567,
+      "grad_norm": 0.22498739924316974,
+      "learning_rate": 0.0001985256759242359,
+      "loss": 0.5696,
+      "step": 100
+    },
+    {
+      "epoch": 0.14914243102162567,
+      "eval_loss": 0.49856358766555786,
+      "eval_runtime": 4.7327,
+      "eval_samples_per_second": 1.056,
+      "eval_steps_per_second": 0.634,
+      "step": 100
+    },
+    {
+      "epoch": 0.18642803877703207,
+      "grad_norm": 0.14018935803156768,
+      "learning_rate": 0.00019546910545535558,
+      "loss": 0.5523,
+      "step": 125
+    },
+    {
+      "epoch": 0.18642803877703207,
+      "eval_loss": 0.48138752579689026,
+      "eval_runtime": 4.6757,
+      "eval_samples_per_second": 1.069,
+      "eval_steps_per_second": 0.642,
+      "step": 125
+    },
+    {
+      "epoch": 0.22371364653243847,
+      "grad_norm": 0.16571925714609734,
+      "learning_rate": 0.00019079522252288386,
+      "loss": 0.5779,
+      "step": 150
+    },
+    {
+      "epoch": 0.22371364653243847,
+      "eval_loss": 0.472788006067276,
+      "eval_runtime": 4.7083,
+      "eval_samples_per_second": 1.062,
+      "eval_steps_per_second": 0.637,
+      "step": 150
+    },
+    {
+      "epoch": 0.2609992542878449,
+      "grad_norm": 0.14926948884435004,
+      "learning_rate": 0.00018458320592590975,
+      "loss": 0.5351,
+      "step": 175
+    },
+    {
+      "epoch": 0.2609992542878449,
+      "eval_loss": 0.467672735452652,
+      "eval_runtime": 4.6956,
+      "eval_samples_per_second": 1.065,
+      "eval_steps_per_second": 0.639,
+      "step": 175
+    },
+    {
+      "epoch": 0.29828486204325133,
+      "grad_norm": 0.13621513300078802,
+      "learning_rate": 0.00017693829150820068,
+      "loss": 0.5251,
+      "step": 200
+    },
+    {
+      "epoch": 0.29828486204325133,
+      "eval_loss": 0.4527561664581299,
+      "eval_runtime": 4.7174,
+      "eval_samples_per_second": 1.06,
+      "eval_steps_per_second": 0.636,
+      "step": 200
+    },
+    {
+      "epoch": 0.33557046979865773,
+      "grad_norm": 0.14343687732505184,
+      "learning_rate": 0.00016798998939045895,
+      "loss": 0.5301,
+      "step": 225
+    },
+    {
+      "epoch": 0.33557046979865773,
+      "eval_loss": 0.45074066519737244,
+      "eval_runtime": 4.7651,
+      "eval_samples_per_second": 1.049,
+      "eval_steps_per_second": 0.63,
+      "step": 225
+    },
+    {
+      "epoch": 0.37285607755406414,
+      "grad_norm": 0.14384455598548593,
+      "learning_rate": 0.00015788988997959114,
+      "loss": 0.5554,
+      "step": 250
+    },
+    {
+      "epoch": 0.37285607755406414,
+      "eval_loss": 0.45150551199913025,
+      "eval_runtime": 4.7234,
+      "eval_samples_per_second": 1.059,
+      "eval_steps_per_second": 0.635,
+      "step": 250
+    },
+    {
+      "epoch": 0.41014168530947054,
+      "grad_norm": 0.11661007763800872,
+      "learning_rate": 0.0001468090959227082,
+      "loss": 0.5198,
+      "step": 275
+    },
+    {
+      "epoch": 0.41014168530947054,
+      "eval_loss": 0.4460136294364929,
+      "eval_runtime": 4.7242,
+      "eval_samples_per_second": 1.058,
+      "eval_steps_per_second": 0.635,
+      "step": 275
+    },
+    {
+      "epoch": 0.44742729306487694,
+      "grad_norm": 0.1321962416990341,
+      "learning_rate": 0.0001349353235103232,
+      "loss": 0.5484,
+      "step": 300
+    },
+    {
+      "epoch": 0.44742729306487694,
+      "eval_loss": 0.4421899914741516,
+      "eval_runtime": 4.7511,
+      "eval_samples_per_second": 1.052,
+      "eval_steps_per_second": 0.631,
+      "step": 300
+    },
+    {
+      "epoch": 0.48471290082028334,
+      "grad_norm": 0.13900994098830932,
+      "learning_rate": 0.0001224697226329772,
+      "loss": 0.5223,
+      "step": 325
+    },
+    {
+      "epoch": 0.48471290082028334,
+      "eval_loss": 0.4396364092826843,
+      "eval_runtime": 4.705,
+      "eval_samples_per_second": 1.063,
+      "eval_steps_per_second": 0.638,
+      "step": 325
+    },
+    {
+      "epoch": 0.5219985085756897,
+      "grad_norm": 0.15825647169367713,
+      "learning_rate": 0.00010962346916341903,
+      "loss": 0.4858,
+      "step": 350
+    },
+    {
+      "epoch": 0.5219985085756897,
+      "eval_loss": 0.43353357911109924,
+      "eval_runtime": 4.7849,
+      "eval_samples_per_second": 1.045,
+      "eval_steps_per_second": 0.627,
+      "step": 350
+    },
+    {
+      "epoch": 0.5592841163310962,
+      "grad_norm": 0.11937908855087626,
+      "learning_rate": 9.661418749173467e-05,
+      "loss": 0.5051,
+      "step": 375
+    },
+    {
+      "epoch": 0.5592841163310962,
+      "eval_loss": 0.42787012457847595,
+      "eval_runtime": 4.7427,
+      "eval_samples_per_second": 1.054,
+      "eval_steps_per_second": 0.633,
+      "step": 375
+    },
+    {
+      "epoch": 0.5965697240865027,
+      "grad_norm": 0.1236665974221879,
+      "learning_rate": 8.366226381814697e-05,
+      "loss": 0.489,
+      "step": 400
+    },
+    {
+      "epoch": 0.5965697240865027,
+      "eval_loss": 0.4265735149383545,
+      "eval_runtime": 4.7586,
+      "eval_samples_per_second": 1.051,
+      "eval_steps_per_second": 0.63,
+      "step": 400
+    },
+    {
+      "epoch": 0.633855331841909,
+      "grad_norm": 0.1340681564316269,
+      "learning_rate": 7.09871126588481e-05,
+      "loss": 0.4992,
+      "step": 425
+    },
+    {
+      "epoch": 0.633855331841909,
+      "eval_loss": 0.41966643929481506,
+      "eval_runtime": 4.725,
+      "eval_samples_per_second": 1.058,
+      "eval_steps_per_second": 0.635,
+      "step": 425
+    },
+    {
+      "epoch": 0.6711409395973155,
+      "grad_norm": 0.14955138964358772,
+      "learning_rate": 5.880345981282876e-05,
+      "loss": 0.4985,
+      "step": 450
+    },
+    {
+      "epoch": 0.6711409395973155,
+      "eval_loss": 0.41805362701416016,
+      "eval_runtime": 4.6928,
+      "eval_samples_per_second": 1.065,
+      "eval_steps_per_second": 0.639,
+      "step": 450
+    },
+    {
+      "epoch": 0.7084265473527218,
+      "grad_norm": 0.13281003649358977,
+      "learning_rate": 4.7317704758809946e-05,
+      "loss": 0.511,
+      "step": 475
+    },
+    {
+      "epoch": 0.7084265473527218,
+      "eval_loss": 0.4160347878932953,
+      "eval_runtime": 4.6998,
+      "eval_samples_per_second": 1.064,
+      "eval_steps_per_second": 0.638,
+      "step": 475
+    },
+    {
+      "epoch": 0.7457121551081283,
+      "grad_norm": 0.13212294340504707,
+      "learning_rate": 3.672442410577965e-05,
+      "loss": 0.5103,
+      "step": 500
+    },
+    {
+      "epoch": 0.7457121551081283,
+      "eval_loss": 0.41526561975479126,
+      "eval_runtime": 4.7196,
+      "eval_samples_per_second": 1.059,
+      "eval_steps_per_second": 0.636,
+      "step": 500
+    },
+    {
+      "epoch": 0.7829977628635347,
+      "grad_norm": 0.13255473695033904,
+      "learning_rate": 2.7203075331094017e-05,
+      "loss": 0.4953,
+      "step": 525
+    },
+    {
+      "epoch": 0.7829977628635347,
+      "eval_loss": 0.41363996267318726,
+      "eval_runtime": 4.6994,
+      "eval_samples_per_second": 1.064,
+      "eval_steps_per_second": 0.638,
+      "step": 525
+    },
+    {
+      "epoch": 0.8202833706189411,
+      "grad_norm": 0.11553089711715571,
+      "learning_rate": 1.89149566470915e-05,
+      "loss": 0.5017,
+      "step": 550
+    },
+    {
+      "epoch": 0.8202833706189411,
+      "eval_loss": 0.4126836359500885,
+      "eval_runtime": 4.7366,
+      "eval_samples_per_second": 1.056,
+      "eval_steps_per_second": 0.633,
+      "step": 550
+    },
+    {
+      "epoch": 0.8575689783743475,
+      "grad_norm": 0.1205660312265152,
+      "learning_rate": 1.2000474498175552e-05,
+      "loss": 0.488,
+      "step": 575
+    },
+    {
+      "epoch": 0.8575689783743475,
+      "eval_loss": 0.41148170828819275,
+      "eval_runtime": 4.6986,
+      "eval_samples_per_second": 1.064,
+      "eval_steps_per_second": 0.638,
+      "step": 575
+    },
+    {
+      "epoch": 0.8948545861297539,
+      "grad_norm": 0.11357983134630366,
+      "learning_rate": 6.576764978849004e-06,
+      "loss": 0.4862,
+      "step": 600
+    },
+    {
+      "epoch": 0.8948545861297539,
+      "eval_loss": 0.4110669493675232,
+      "eval_runtime": 4.7813,
+      "eval_samples_per_second": 1.046,
+      "eval_steps_per_second": 0.627,
+      "step": 600
+    },
+    {
+      "epoch": 0.9321401938851603,
+      "grad_norm": 0.11689127563900782,
+      "learning_rate": 2.735709467518699e-06,
+      "loss": 0.4866,
+      "step": 625
+    },
+    {
+      "epoch": 0.9321401938851603,
+      "eval_loss": 0.41068965196609497,
+      "eval_runtime": 4.7375,
+      "eval_samples_per_second": 1.055,
+      "eval_steps_per_second": 0.633,
+      "step": 625
+    },
+    {
+      "epoch": 0.9694258016405667,
+      "grad_norm": 0.12155258793935453,
+      "learning_rate": 5.42378092601481e-07,
+      "loss": 0.4902,
+      "step": 650
+    },
+    {
+      "epoch": 0.9694258016405667,
+      "eval_loss": 0.4103812277317047,
+      "eval_runtime": 4.6763,
+      "eval_samples_per_second": 1.069,
+      "eval_steps_per_second": 0.642,
+      "step": 650
+    },
+    {
+      "epoch": 0.9992542878448919,
+      "step": 670,
+      "total_flos": 7.44129634982953e+16,
+      "train_loss": 0.2547340336130626,
+      "train_runtime": 9634.561,
+      "train_samples_per_second": 0.556,
+      "train_steps_per_second": 0.07
+    }
+  ],
+  "logging_steps": 25,
+  "max_steps": 670,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 1,
+  "save_steps": 25,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": true
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 7.44129634982953e+16,
+  "train_batch_size": 2,
+  "trial_name": null,
+  "trial_params": null
+}