End of training

README.md

@@ -4,6 +4,7 @@ license: gemma
 base_model: google/gemma-3-1b-it
 tags:
 - llama-factory
+- prompt-tuning
 - generated_from_trainer
 model-index:
 - name: train_2025-04-10-10-34-00
@@ -15,7 +16,7 @@ should probably proofread and complete it, then remove this comment. -->
 
 # train_2025-04-10-10-34-00
 
-This model is a fine-tuned version of [google/gemma-3-1b-it](https://huggingface.co/google/gemma-3-1b-it) on an unknown dataset.
+This model is a fine-tuned version of [google/gemma-3-1b-it](https://huggingface.co/google/gemma-3-1b-it) on the mnli_train dataset.
 
 ## Model description
 

all_results.json

@@ -0,0 +1,13 @@
+{
+    "epoch": 2.634920634920635,
+    "eval_mnli_eval_loss": 0.5607374906539917,
+    "eval_mnli_eval_runtime": 10.1933,
+    "eval_mnli_eval_samples_per_second": 98.104,
+    "eval_mnli_eval_steps_per_second": 6.181,
+    "num_input_tokens_seen": 256448,
+    "total_flos": 1073844298924032.0,
+    "train_loss": 5.759574245838892,
+    "train_runtime": 153.0301,
+    "train_samples_per_second": 19.604,
+    "train_steps_per_second": 0.137
+}

eval_results.json

@@ -0,0 +1,8 @@
+{
+    "epoch": 2.634920634920635,
+    "eval_mnli_eval_loss": 0.5607374906539917,
+    "eval_mnli_eval_runtime": 10.1933,
+    "eval_mnli_eval_samples_per_second": 98.104,
+    "eval_mnli_eval_steps_per_second": 6.181,
+    "num_input_tokens_seen": 256448
+}

train_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 2.634920634920635,
+    "num_input_tokens_seen": 256448,
+    "total_flos": 1073844298924032.0,
+    "train_loss": 5.759574245838892,
+    "train_runtime": 153.0301,
+    "train_samples_per_second": 19.604,
+    "train_steps_per_second": 0.137
+}

trainer_state.json

@@ -0,0 +1,139 @@
+{
+  "best_global_step": 21,
+  "best_metric": 0.5607374906539917,
+  "best_model_checkpoint": "saves/prompt-tuning/gemma-3-1b-it/train_2025-04-10-10-34-00/checkpoint-21",
+  "epoch": 2.634920634920635,
+  "eval_steps": 3,
+  "global_step": 21,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.38095238095238093,
+      "eval_mnli_eval_loss": 11.7061185836792,
+      "eval_mnli_eval_runtime": 10.1405,
+      "eval_mnli_eval_samples_per_second": 98.615,
+      "eval_mnli_eval_steps_per_second": 6.213,
+      "num_input_tokens_seen": 39424,
+      "step": 3
+    },
+    {
+      "epoch": 0.6349206349206349,
+      "grad_norm": 13.432448387145996,
+      "learning_rate": 0.2739358161473992,
+      "loss": 11.6642,
+      "num_input_tokens_seen": 62976,
+      "step": 5
+    },
+    {
+      "epoch": 0.7619047619047619,
+      "eval_mnli_eval_loss": 9.128030776977539,
+      "eval_mnli_eval_runtime": 10.2011,
+      "eval_mnli_eval_samples_per_second": 98.029,
+      "eval_mnli_eval_steps_per_second": 6.176,
+      "num_input_tokens_seen": 76288,
+      "step": 6
+    },
+    {
+      "epoch": 1.126984126984127,
+      "eval_mnli_eval_loss": 5.672970294952393,
+      "eval_mnli_eval_runtime": 10.1995,
+      "eval_mnli_eval_samples_per_second": 98.044,
+      "eval_mnli_eval_steps_per_second": 6.177,
+      "num_input_tokens_seen": 110336,
+      "step": 9
+    },
+    {
+      "epoch": 1.253968253968254,
+      "grad_norm": 2.507275342941284,
+      "learning_rate": 0.18337814009344713,
+      "loss": 7.965,
+      "num_input_tokens_seen": 121856,
+      "step": 10
+    },
+    {
+      "epoch": 1.507936507936508,
+      "eval_mnli_eval_loss": 2.735905170440674,
+      "eval_mnli_eval_runtime": 10.2202,
+      "eval_mnli_eval_samples_per_second": 97.845,
+      "eval_mnli_eval_steps_per_second": 6.164,
+      "num_input_tokens_seen": 149120,
+      "step": 12
+    },
+    {
+      "epoch": 1.8888888888888888,
+      "grad_norm": 1.7794915437698364,
+      "learning_rate": 0.07500000000000002,
+      "loss": 3.4004,
+      "num_input_tokens_seen": 187264,
+      "step": 15
+    },
+    {
+      "epoch": 1.8888888888888888,
+      "eval_mnli_eval_loss": 1.661912441253662,
+      "eval_mnli_eval_runtime": 10.2019,
+      "eval_mnli_eval_samples_per_second": 98.021,
+      "eval_mnli_eval_steps_per_second": 6.175,
+      "num_input_tokens_seen": 187264,
+      "step": 15
+    },
+    {
+      "epoch": 2.253968253968254,
+      "eval_mnli_eval_loss": 0.739750325679779,
+      "eval_mnli_eval_runtime": 10.2066,
+      "eval_mnli_eval_samples_per_second": 97.976,
+      "eval_mnli_eval_steps_per_second": 6.172,
+      "num_input_tokens_seen": 219072,
+      "step": 18
+    },
+    {
+      "epoch": 2.507936507936508,
+      "grad_norm": 0.5217874050140381,
+      "learning_rate": 0.0066640791320788815,
+      "loss": 1.0424,
+      "num_input_tokens_seen": 244032,
+      "step": 20
+    },
+    {
+      "epoch": 2.634920634920635,
+      "eval_mnli_eval_loss": 0.5607374906539917,
+      "eval_mnli_eval_runtime": 10.1954,
+      "eval_mnli_eval_samples_per_second": 98.083,
+      "eval_mnli_eval_steps_per_second": 6.179,
+      "num_input_tokens_seen": 256448,
+      "step": 21
+    },
+    {
+      "epoch": 2.634920634920635,
+      "num_input_tokens_seen": 256448,
+      "step": 21,
+      "total_flos": 1073844298924032.0,
+      "train_loss": 5.759574245838892,
+      "train_runtime": 153.0301,
+      "train_samples_per_second": 19.604,
+      "train_steps_per_second": 0.137
+    }
+  ],
+  "logging_steps": 5,
+  "max_steps": 21,
+  "num_input_tokens_seen": 256448,
+  "num_train_epochs": 3,
+  "save_steps": 3,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": true
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 1073844298924032.0,
+  "train_batch_size": 16,
+  "trial_name": null,
+  "trial_params": null
+}