Upload trainer_state.json with huggingface_hub

trainer_state.json

@@ -0,0 +1,217 @@
+{
+  "best_metric": 0.9069767594337463,
+  "best_model_checkpoint": "./llama3/27-06-24-Weni-ZeroShot-Agents-Llama3-4.0.37-DPO_Experiment with DPO and Llama3 8B, zeroshot 4.0.37-2_max_steps-570_batch_16_2024-06-27_ppid_9/checkpoint-80",
+  "epoch": 0.8366013071895425,
+  "eval_steps": 20,
+  "global_step": 80,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.10457516339869281,
+      "grad_norm": 3.655749559402466,
+      "learning_rate": 3.3333333333333337e-06,
+      "logits/chosen": -0.22948360443115234,
+      "logits/rejected": -0.22978875041007996,
+      "logps/chosen": -39.710899353027344,
+      "logps/rejected": -39.52346420288086,
+      "loss": 0.662,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": 0.11001075804233551,
+      "rewards/margins": 0.06790003925561905,
+      "rewards/rejected": 0.04211071878671646,
+      "step": 10
+    },
+    {
+      "epoch": 0.20915032679738563,
+      "grad_norm": 2.4934024810791016,
+      "learning_rate": 5.978260869565218e-06,
+      "logits/chosen": -0.22906163334846497,
+      "logits/rejected": -0.22876068949699402,
+      "logps/chosen": -30.569751739501953,
+      "logps/rejected": -33.86491012573242,
+      "loss": 0.5549,
+      "rewards/accuracies": 0.7437499761581421,
+      "rewards/chosen": 0.9981435537338257,
+      "rewards/margins": 0.38557058572769165,
+      "rewards/rejected": 0.6125729084014893,
+      "step": 20
+    },
+    {
+      "epoch": 0.20915032679738563,
+      "eval_logits/chosen": -0.19793638586997986,
+      "eval_logits/rejected": -0.197507843375206,
+      "eval_logps/chosen": -26.848573684692383,
+      "eval_logps/rejected": -35.13609313964844,
+      "eval_loss": 0.43225446343421936,
+      "eval_rewards/accuracies": 0.8081395626068115,
+      "eval_rewards/chosen": 1.355695366859436,
+      "eval_rewards/margins": 0.861485481262207,
+      "eval_rewards/rejected": 0.4942099153995514,
+      "eval_runtime": 76.6712,
+      "eval_samples_per_second": 2.23,
+      "eval_steps_per_second": 1.122,
+      "step": 20
+    },
+    {
+      "epoch": 0.3137254901960784,
+      "grad_norm": 1.9067373275756836,
+      "learning_rate": 5.869565217391305e-06,
+      "logits/chosen": -0.23883526027202606,
+      "logits/rejected": -0.2374078780412674,
+      "logps/chosen": -30.109268188476562,
+      "logps/rejected": -43.55142593383789,
+      "loss": 0.3732,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.0676757097244263,
+      "rewards/margins": 1.4297826290130615,
+      "rewards/rejected": -0.3621070086956024,
+      "step": 30
+    },
+    {
+      "epoch": 0.41830065359477125,
+      "grad_norm": 3.629997730255127,
+      "learning_rate": 5.760869565217392e-06,
+      "logits/chosen": -0.2035980522632599,
+      "logits/rejected": -0.20044592022895813,
+      "logps/chosen": -28.863727569580078,
+      "logps/rejected": -50.26195526123047,
+      "loss": 0.3276,
+      "rewards/accuracies": 0.84375,
+      "rewards/chosen": 1.200588345527649,
+      "rewards/margins": 2.2083239555358887,
+      "rewards/rejected": -1.0077354907989502,
+      "step": 40
+    },
+    {
+      "epoch": 0.41830065359477125,
+      "eval_logits/chosen": -0.16349415481090546,
+      "eval_logits/rejected": -0.15954892337322235,
+      "eval_logps/chosen": -24.61951446533203,
+      "eval_logps/rejected": -48.91069412231445,
+      "eval_loss": 0.34378084540367126,
+      "eval_rewards/accuracies": 0.8895348906517029,
+      "eval_rewards/chosen": 1.5786010026931763,
+      "eval_rewards/margins": 2.4618515968322754,
+      "eval_rewards/rejected": -0.8832504749298096,
+      "eval_runtime": 76.7395,
+      "eval_samples_per_second": 2.228,
+      "eval_steps_per_second": 1.121,
+      "step": 40
+    },
+    {
+      "epoch": 0.5228758169934641,
+      "grad_norm": 2.5461719036102295,
+      "learning_rate": 5.652173913043479e-06,
+      "logits/chosen": -0.16493651270866394,
+      "logits/rejected": -0.16144290566444397,
+      "logps/chosen": -24.76712989807129,
+      "logps/rejected": -46.993003845214844,
+      "loss": 0.3648,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.5913469791412354,
+      "rewards/margins": 2.25089955329895,
+      "rewards/rejected": -0.6595526933670044,
+      "step": 50
+    },
+    {
+      "epoch": 0.6274509803921569,
+      "grad_norm": 2.0358729362487793,
+      "learning_rate": 5.543478260869566e-06,
+      "logits/chosen": -0.2202371060848236,
+      "logits/rejected": -0.21734721958637238,
+      "logps/chosen": -26.78680419921875,
+      "logps/rejected": -47.03938674926758,
+      "loss": 0.3448,
+      "rewards/accuracies": 0.831250011920929,
+      "rewards/chosen": 1.3999578952789307,
+      "rewards/margins": 2.0680160522460938,
+      "rewards/rejected": -0.6680583357810974,
+      "step": 60
+    },
+    {
+      "epoch": 0.6274509803921569,
+      "eval_logits/chosen": -0.17944073677062988,
+      "eval_logits/rejected": -0.1763658970594406,
+      "eval_logps/chosen": -27.3469295501709,
+      "eval_logps/rejected": -48.937984466552734,
+      "eval_loss": 0.2700035870075226,
+      "eval_rewards/accuracies": 0.9069767594337463,
+      "eval_rewards/chosen": 1.3058594465255737,
+      "eval_rewards/margins": 2.191838026046753,
+      "eval_rewards/rejected": -0.8859787583351135,
+      "eval_runtime": 76.7029,
+      "eval_samples_per_second": 2.229,
+      "eval_steps_per_second": 1.121,
+      "step": 60
+    },
+    {
+      "epoch": 0.7320261437908496,
+      "grad_norm": 1.7749762535095215,
+      "learning_rate": 5.4347826086956525e-06,
+      "logits/chosen": -0.1932402402162552,
+      "logits/rejected": -0.19045117497444153,
+      "logps/chosen": -26.827350616455078,
+      "logps/rejected": -46.88459014892578,
+      "loss": 0.2909,
+      "rewards/accuracies": 0.8812500238418579,
+      "rewards/chosen": 1.3978151082992554,
+      "rewards/margins": 2.0628795623779297,
+      "rewards/rejected": -0.6650643348693848,
+      "step": 70
+    },
+    {
+      "epoch": 0.8366013071895425,
+      "grad_norm": 2.9730594158172607,
+      "learning_rate": 5.326086956521739e-06,
+      "logits/chosen": -0.15139932930469513,
+      "logits/rejected": -0.14830470085144043,
+      "logps/chosen": -26.229084014892578,
+      "logps/rejected": -47.40288162231445,
+      "loss": 0.2844,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": 1.493841290473938,
+      "rewards/margins": 2.2263193130493164,
+      "rewards/rejected": -0.7324780225753784,
+      "step": 80
+    },
+    {
+      "epoch": 0.8366013071895425,
+      "eval_logits/chosen": -0.15539461374282837,
+      "eval_logits/rejected": -0.15173271298408508,
+      "eval_logps/chosen": -30.73025894165039,
+      "eval_logps/rejected": -54.970027923583984,
+      "eval_loss": 0.262494832277298,
+      "eval_rewards/accuracies": 0.9069767594337463,
+      "eval_rewards/chosen": 0.967526912689209,
+      "eval_rewards/margins": 2.4567105770111084,
+      "eval_rewards/rejected": -1.4891836643218994,
+      "eval_runtime": 76.69,
+      "eval_samples_per_second": 2.23,
+      "eval_steps_per_second": 1.121,
+      "step": 80
+    }
+  ],
+  "logging_steps": 10,
+  "max_steps": 570,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 6,
+  "save_steps": 80,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 0.0,
+  "train_batch_size": 2,
+  "trial_name": null,
+  "trial_params": null
+}