End of training

README.md

@@ -17,10 +17,10 @@ should probably proofread and complete it, then remove this comment. -->
 
 # test
 
-This model is a fine-tuned version of [meta-llama/Meta-Llama-3-8B-Instruct](https://huggingface.co/meta-llama/Meta-Llama-3-8B-Instruct) on an unknown dataset.
+This model is a fine-tuned version of [meta-llama/Meta-Llama-3-8B-Instruct](https://huggingface.co/meta-llama/Meta-Llama-3-8B-Instruct) on the wsc dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.3585
-- Num Input Tokens Seen: 46944
+- Loss: 0.3497
+- Num Input Tokens Seen: 49376
 
 ## Model description
 

all_results.json

@@ -0,0 +1,13 @@
+{
+    "epoch": 1.0,
+    "eval_loss": 0.3496828079223633,
+    "eval_runtime": 1.4986,
+    "eval_samples_per_second": 37.369,
+    "eval_steps_per_second": 9.342,
+    "num_input_tokens_seen": 49376,
+    "total_flos": 2223378963628032.0,
+    "train_loss": 1.100914571762085,
+    "train_runtime": 86.254,
+    "train_samples_per_second": 5.774,
+    "train_steps_per_second": 1.449
+}

eval_results.json

@@ -0,0 +1,8 @@
+{
+    "epoch": 1.0,
+    "eval_loss": 0.3496828079223633,
+    "eval_runtime": 1.4986,
+    "eval_samples_per_second": 37.369,
+    "eval_steps_per_second": 9.342,
+    "num_input_tokens_seen": 49376
+}

train_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 1.0,
+    "num_input_tokens_seen": 49376,
+    "total_flos": 2223378963628032.0,
+    "train_loss": 1.100914571762085,
+    "train_runtime": 86.254,
+    "train_samples_per_second": 5.774,
+    "train_steps_per_second": 1.449
+}

trainer_state.json

@@ -0,0 +1,397 @@
+{
+  "best_global_step": 98,
+  "best_metric": 0.3496828079223633,
+  "best_model_checkpoint": "saves/test/checkpoint-98",
+  "epoch": 1.0,
+  "eval_steps": 7,
+  "global_step": 125,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.04,
+      "grad_norm": 396.0,
+      "learning_rate": 0.009230769230769232,
+      "loss": 3.6273,
+      "num_input_tokens_seen": 2144,
+      "step": 5
+    },
+    {
+      "epoch": 0.056,
+      "eval_loss": 6.932018280029297,
+      "eval_runtime": 1.4196,
+      "eval_samples_per_second": 39.449,
+      "eval_steps_per_second": 9.862,
+      "num_input_tokens_seen": 2880,
+      "step": 7
+    },
+    {
+      "epoch": 0.08,
+      "grad_norm": 22.5,
+      "learning_rate": 0.02076923076923077,
+      "loss": 6.3813,
+      "num_input_tokens_seen": 4128,
+      "step": 10
+    },
+    {
+      "epoch": 0.112,
+      "eval_loss": 1.6227623224258423,
+      "eval_runtime": 1.3856,
+      "eval_samples_per_second": 40.417,
+      "eval_steps_per_second": 10.104,
+      "num_input_tokens_seen": 5920,
+      "step": 14
+    },
+    {
+      "epoch": 0.12,
+      "grad_norm": 9.9375,
+      "learning_rate": 0.029994099395745794,
+      "loss": 2.5342,
+      "num_input_tokens_seen": 6240,
+      "step": 15
+    },
+    {
+      "epoch": 0.16,
+      "grad_norm": 0.9140625,
+      "learning_rate": 0.029788065277773536,
+      "loss": 1.4507,
+      "num_input_tokens_seen": 8096,
+      "step": 20
+    },
+    {
+      "epoch": 0.168,
+      "eval_loss": 0.4040253460407257,
+      "eval_runtime": 1.3702,
+      "eval_samples_per_second": 40.869,
+      "eval_steps_per_second": 10.217,
+      "num_input_tokens_seen": 8416,
+      "step": 21
+    },
+    {
+      "epoch": 0.2,
+      "grad_norm": 1.140625,
+      "learning_rate": 0.02929162684084344,
+      "loss": 1.7771,
+      "num_input_tokens_seen": 10112,
+      "step": 25
+    },
+    {
+      "epoch": 0.224,
+      "eval_loss": 3.618654489517212,
+      "eval_runtime": 1.3779,
+      "eval_samples_per_second": 40.642,
+      "eval_steps_per_second": 10.161,
+      "num_input_tokens_seen": 11264,
+      "step": 28
+    },
+    {
+      "epoch": 0.24,
+      "grad_norm": 103.5,
+      "learning_rate": 0.028514533018536285,
+      "loss": 3.7935,
+      "num_input_tokens_seen": 12032,
+      "step": 30
+    },
+    {
+      "epoch": 0.28,
+      "grad_norm": 0.12060546875,
+      "learning_rate": 0.02747204418453818,
+      "loss": 0.7848,
+      "num_input_tokens_seen": 13824,
+      "step": 35
+    },
+    {
+      "epoch": 0.28,
+      "eval_loss": 0.36673951148986816,
+      "eval_runtime": 1.3899,
+      "eval_samples_per_second": 40.29,
+      "eval_steps_per_second": 10.072,
+      "num_input_tokens_seen": 13824,
+      "step": 35
+    },
+    {
+      "epoch": 0.32,
+      "grad_norm": 0.11083984375,
+      "learning_rate": 0.026184632473247482,
+      "loss": 0.4314,
+      "num_input_tokens_seen": 15840,
+      "step": 40
+    },
+    {
+      "epoch": 0.336,
+      "eval_loss": 0.3661668598651886,
+      "eval_runtime": 1.3878,
+      "eval_samples_per_second": 40.353,
+      "eval_steps_per_second": 10.088,
+      "num_input_tokens_seen": 16672,
+      "step": 42
+    },
+    {
+      "epoch": 0.36,
+      "grad_norm": 0.251953125,
+      "learning_rate": 0.024677579753131316,
+      "loss": 0.4096,
+      "num_input_tokens_seen": 17920,
+      "step": 45
+    },
+    {
+      "epoch": 0.392,
+      "eval_loss": 0.5265085101127625,
+      "eval_runtime": 1.3816,
+      "eval_samples_per_second": 40.532,
+      "eval_steps_per_second": 10.133,
+      "num_input_tokens_seen": 19296,
+      "step": 49
+    },
+    {
+      "epoch": 0.4,
+      "grad_norm": 0.07373046875,
+      "learning_rate": 0.022980481147730047,
+      "loss": 0.2423,
+      "num_input_tokens_seen": 19712,
+      "step": 50
+    },
+    {
+      "epoch": 0.44,
+      "grad_norm": 0.150390625,
+      "learning_rate": 0.021126663854039943,
+      "loss": 0.5554,
+      "num_input_tokens_seen": 21952,
+      "step": 55
+    },
+    {
+      "epoch": 0.448,
+      "eval_loss": 0.39251771569252014,
+      "eval_runtime": 1.3876,
+      "eval_samples_per_second": 40.359,
+      "eval_steps_per_second": 10.09,
+      "num_input_tokens_seen": 22432,
+      "step": 56
+    },
+    {
+      "epoch": 0.48,
+      "grad_norm": 0.1513671875,
+      "learning_rate": 0.01915253267137274,
+      "loss": 0.4968,
+      "num_input_tokens_seen": 24160,
+      "step": 60
+    },
+    {
+      "epoch": 0.504,
+      "eval_loss": 2.652535915374756,
+      "eval_runtime": 1.3806,
+      "eval_samples_per_second": 40.562,
+      "eval_steps_per_second": 10.141,
+      "num_input_tokens_seen": 25504,
+      "step": 63
+    },
+    {
+      "epoch": 0.52,
+      "grad_norm": 10.0625,
+      "learning_rate": 0.017096855093032493,
+      "loss": 0.7728,
+      "num_input_tokens_seen": 26112,
+      "step": 65
+    },
+    {
+      "epoch": 0.56,
+      "grad_norm": 0.189453125,
+      "learning_rate": 0.015,
+      "loss": 0.3298,
+      "num_input_tokens_seen": 28064,
+      "step": 70
+    },
+    {
+      "epoch": 0.56,
+      "eval_loss": 0.37763792276382446,
+      "eval_runtime": 1.3741,
+      "eval_samples_per_second": 40.753,
+      "eval_steps_per_second": 10.188,
+      "num_input_tokens_seen": 28064,
+      "step": 70
+    },
+    {
+      "epoch": 0.6,
+      "grad_norm": 0.12060546875,
+      "learning_rate": 0.012903144906967513,
+      "loss": 0.3663,
+      "num_input_tokens_seen": 29824,
+      "step": 75
+    },
+    {
+      "epoch": 0.616,
+      "eval_loss": 0.36274200677871704,
+      "eval_runtime": 1.4269,
+      "eval_samples_per_second": 39.245,
+      "eval_steps_per_second": 9.811,
+      "num_input_tokens_seen": 30720,
+      "step": 77
+    },
+    {
+      "epoch": 0.64,
+      "grad_norm": 0.142578125,
+      "learning_rate": 0.01084746732862726,
+      "loss": 0.3654,
+      "num_input_tokens_seen": 31904,
+      "step": 80
+    },
+    {
+      "epoch": 0.672,
+      "eval_loss": 0.35259029269218445,
+      "eval_runtime": 1.4343,
+      "eval_samples_per_second": 39.044,
+      "eval_steps_per_second": 9.761,
+      "num_input_tokens_seen": 33504,
+      "step": 84
+    },
+    {
+      "epoch": 0.68,
+      "grad_norm": 0.08349609375,
+      "learning_rate": 0.008873336145960059,
+      "loss": 0.3696,
+      "num_input_tokens_seen": 33984,
+      "step": 85
+    },
+    {
+      "epoch": 0.72,
+      "grad_norm": 0.039306640625,
+      "learning_rate": 0.007019518852269953,
+      "loss": 0.3495,
+      "num_input_tokens_seen": 35776,
+      "step": 90
+    },
+    {
+      "epoch": 0.728,
+      "eval_loss": 0.3546493351459503,
+      "eval_runtime": 1.4075,
+      "eval_samples_per_second": 39.786,
+      "eval_steps_per_second": 9.946,
+      "num_input_tokens_seen": 36128,
+      "step": 91
+    },
+    {
+      "epoch": 0.76,
+      "grad_norm": 0.017578125,
+      "learning_rate": 0.005322420246868689,
+      "loss": 0.412,
+      "num_input_tokens_seen": 37472,
+      "step": 95
+    },
+    {
+      "epoch": 0.784,
+      "eval_loss": 0.3496828079223633,
+      "eval_runtime": 1.4395,
+      "eval_samples_per_second": 38.902,
+      "eval_steps_per_second": 9.726,
+      "num_input_tokens_seen": 38592,
+      "step": 98
+    },
+    {
+      "epoch": 0.8,
+      "grad_norm": 0.02734375,
+      "learning_rate": 0.003815367526752516,
+      "loss": 0.3676,
+      "num_input_tokens_seen": 39328,
+      "step": 100
+    },
+    {
+      "epoch": 0.84,
+      "grad_norm": 0.047607421875,
+      "learning_rate": 0.0025279558154618197,
+      "loss": 0.349,
+      "num_input_tokens_seen": 41280,
+      "step": 105
+    },
+    {
+      "epoch": 0.84,
+      "eval_loss": 0.35384148359298706,
+      "eval_runtime": 1.4061,
+      "eval_samples_per_second": 39.827,
+      "eval_steps_per_second": 9.957,
+      "num_input_tokens_seen": 41280,
+      "step": 105
+    },
+    {
+      "epoch": 0.88,
+      "grad_norm": 0.037841796875,
+      "learning_rate": 0.0014854669814637145,
+      "loss": 0.3482,
+      "num_input_tokens_seen": 43552,
+      "step": 110
+    },
+    {
+      "epoch": 0.896,
+      "eval_loss": 0.3565780818462372,
+      "eval_runtime": 1.4483,
+      "eval_samples_per_second": 38.666,
+      "eval_steps_per_second": 9.666,
+      "num_input_tokens_seen": 44160,
+      "step": 112
+    },
+    {
+      "epoch": 0.92,
+      "grad_norm": 0.040283203125,
+      "learning_rate": 0.000708373159156555,
+      "loss": 0.3258,
+      "num_input_tokens_seen": 45216,
+      "step": 115
+    },
+    {
+      "epoch": 0.952,
+      "eval_loss": 0.35846200585365295,
+      "eval_runtime": 1.4597,
+      "eval_samples_per_second": 38.364,
+      "eval_steps_per_second": 9.591,
+      "num_input_tokens_seen": 46944,
+      "step": 119
+    },
+    {
+      "epoch": 0.96,
+      "grad_norm": 0.0189208984375,
+      "learning_rate": 0.0002119347222264617,
+      "loss": 0.3217,
+      "num_input_tokens_seen": 47360,
+      "step": 120
+    },
+    {
+      "epoch": 1.0,
+      "grad_norm": 0.0859375,
+      "learning_rate": 5.90060425420702e-06,
+      "loss": 0.3608,
+      "num_input_tokens_seen": 49376,
+      "step": 125
+    },
+    {
+      "epoch": 1.0,
+      "num_input_tokens_seen": 49376,
+      "step": 125,
+      "total_flos": 2223378963628032.0,
+      "train_loss": 1.100914571762085,
+      "train_runtime": 86.254,
+      "train_samples_per_second": 5.774,
+      "train_steps_per_second": 1.449
+    }
+  ],
+  "logging_steps": 5,
+  "max_steps": 125,
+  "num_input_tokens_seen": 49376,
+  "num_train_epochs": 1,
+  "save_steps": 7,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": true
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 2223378963628032.0,
+  "train_batch_size": 4,
+  "trial_name": null,
+  "trial_params": null
+}