Model save

README.md

@@ -2,16 +2,11 @@
 license: mit
 base_model: gpt2
 tags:
-- alignment-handbook
 - trl
 - sft
 - generated_from_trainer
-- trl
-- sft
-- alignment-handbook
-- generated_from_trainer
 datasets:
-- shahadalkhalifa/Crypto_Whitepaper_Labeled
+- generator
 model-index:
 - name: gpt2-cpt-white
   results: []
@@ -22,7 +17,7 @@ should probably proofread and complete it, then remove this comment. -->
 
 # gpt2-cpt-white
 
-This model is a fine-tuned version of [gpt2](https://huggingface.co/gpt2) on the shahadalkhalifa/Crypto_Whitepaper_Labeled dataset.
+This model is a fine-tuned version of [gpt2](https://huggingface.co/gpt2) on the generator dataset.
 
 ## Model description
 
@@ -46,13 +41,10 @@ The following hyperparameters were used during training:
 - eval_batch_size: 8
 - seed: 42
 - distributed_type: multi-GPU
-- num_devices: 2
-- total_train_batch_size: 32
-- total_eval_batch_size: 16
 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
 - lr_scheduler_type: cosine
 - lr_scheduler_warmup_ratio: 0.1
-- num_epochs: 4.0
+- num_epochs: 8.0
 
 ### Training results
 

all_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 4.0,
-    "train_loss": 5.970038026571274,
-    "train_runtime": 52.1743,
+    "epoch": 8.0,
+    "train_loss": 5.653645019978285,
+    "train_runtime": 83.1481,
     "train_samples": 91,
-    "train_samples_per_second": 19.243,
-    "train_steps_per_second": 0.613
+    "train_samples_per_second": 24.15,
+    "train_steps_per_second": 1.539
 }

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:0972128030522cc17565b1e687c530752363c97e94f68f395954b9de7db257c8
+oid sha256:d739b2f6e5ebb5f1d67d6a09c779f5d491a40cebc35a5ccae1f04013479c89f4
 size 248894656

train_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 4.0,
-    "train_loss": 5.970038026571274,
-    "train_runtime": 52.1743,
+    "epoch": 8.0,
+    "train_loss": 5.653645019978285,
+    "train_runtime": 83.1481,
     "train_samples": 91,
-    "train_samples_per_second": 19.243,
-    "train_steps_per_second": 0.613
+    "train_samples_per_second": 24.15,
+    "train_steps_per_second": 1.539
 }

trainer_state.json

@@ -1,78 +1,211 @@
 {
   "best_metric": null,
   "best_model_checkpoint": null,
-  "epoch": 4.0,
+  "epoch": 8.0,
   "eval_steps": 500,
-  "global_step": 32,
+  "global_step": 128,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
-      "epoch": 0.12,
-      "grad_norm": 6.875,
-      "learning_rate": 5e-05,
-      "loss": 6.4878,
+      "epoch": 0.06,
+      "grad_norm": 7.65625,
+      "learning_rate": 1.5384615384615387e-05,
+      "loss": 6.3526,
       "step": 1
     },
     {
-      "epoch": 0.62,
-      "grad_norm": 5.0,
-      "learning_rate": 0.00019937122098932428,
-      "loss": 6.2983,
+      "epoch": 0.31,
+      "grad_norm": 6.3125,
+      "learning_rate": 7.692307692307693e-05,
+      "loss": 6.3681,
       "step": 5
     },
     {
-      "epoch": 1.25,
-      "grad_norm": 3.078125,
-      "learning_rate": 0.000178183148246803,
-      "loss": 6.1058,
+      "epoch": 0.62,
+      "grad_norm": 5.09375,
+      "learning_rate": 0.00015384615384615385,
+      "loss": 6.1993,
       "step": 10
     },
     {
-      "epoch": 1.88,
-      "grad_norm": 2.71875,
-      "learning_rate": 0.00013302790619551674,
-      "loss": 5.931,
+      "epoch": 0.94,
+      "grad_norm": 4.78125,
+      "learning_rate": 0.0001998507803482828,
+      "loss": 6.199,
       "step": 15
     },
     {
-      "epoch": 2.5,
-      "grad_norm": 2.609375,
-      "learning_rate": 7.774790660436858e-05,
-      "loss": 5.9051,
+      "epoch": 1.25,
+      "grad_norm": 3.203125,
+      "learning_rate": 0.00019817716940755586,
+      "loss": 5.9857,
       "step": 20
     },
     {
-      "epoch": 3.12,
-      "grad_norm": 2.4375,
-      "learning_rate": 2.9289321881345254e-05,
-      "loss": 5.806,
+      "epoch": 1.56,
+      "grad_norm": 2.96875,
+      "learning_rate": 0.00019467470355910438,
+      "loss": 5.9041,
       "step": 25
     },
     {
-      "epoch": 3.75,
-      "grad_norm": 2.390625,
-      "learning_rate": 2.5072087818176382e-06,
-      "loss": 5.859,
+      "epoch": 1.88,
+      "grad_norm": 2.859375,
+      "learning_rate": 0.00018940862711476513,
+      "loss": 5.8063,
       "step": 30
     },
     {
-      "epoch": 4.0,
-      "step": 32,
-      "total_flos": 535126081536000.0,
-      "train_loss": 5.970038026571274,
-      "train_runtime": 52.1743,
-      "train_samples_per_second": 19.243,
-      "train_steps_per_second": 0.613
+      "epoch": 2.19,
+      "grad_norm": 2.609375,
+      "learning_rate": 0.00018247703711489686,
+      "loss": 5.8111,
+      "step": 35
+    },
+    {
+      "epoch": 2.5,
+      "grad_norm": 2.3125,
+      "learning_rate": 0.0001740090559661252,
+      "loss": 5.6765,
+      "step": 40
+    },
+    {
+      "epoch": 2.81,
+      "grad_norm": 2.3125,
+      "learning_rate": 0.0001641624261352161,
+      "loss": 5.7295,
+      "step": 45
+    },
+    {
+      "epoch": 3.12,
+      "grad_norm": 2.1875,
+      "learning_rate": 0.00015312057170538035,
+      "loss": 5.5331,
+      "step": 50
+    },
+    {
+      "epoch": 3.44,
+      "grad_norm": 2.171875,
+      "learning_rate": 0.00014108918153265485,
+      "loss": 5.5281,
+      "step": 55
+    },
+    {
+      "epoch": 3.75,
+      "grad_norm": 2.046875,
+      "learning_rate": 0.000128292377651693,
+      "loss": 5.605,
+      "step": 60
+    },
+    {
+      "epoch": 4.06,
+      "grad_norm": 1.9296875,
+      "learning_rate": 0.00011496854030631443,
+      "loss": 5.5033,
+      "step": 65
+    },
+    {
+      "epoch": 4.38,
+      "grad_norm": 1.8671875,
+      "learning_rate": 0.0001013658673765951,
+      "loss": 5.5589,
+      "step": 70
+    },
+    {
+      "epoch": 4.69,
+      "grad_norm": 1.8984375,
+      "learning_rate": 8.773775092197017e-05,
+      "loss": 5.5488,
+      "step": 75
+    },
+    {
+      "epoch": 5.0,
+      "grad_norm": 2.125,
+      "learning_rate": 7.433805696660266e-05,
+      "loss": 5.3959,
+      "step": 80
+    },
+    {
+      "epoch": 5.31,
+      "grad_norm": 1.9140625,
+      "learning_rate": 6.141639645568646e-05,
+      "loss": 5.4935,
+      "step": 85
+    },
+    {
+      "epoch": 5.62,
+      "grad_norm": 1.796875,
+      "learning_rate": 4.921347547581939e-05,
+      "loss": 5.4818,
+      "step": 90
+    },
+    {
+      "epoch": 5.94,
+      "grad_norm": 1.703125,
+      "learning_rate": 3.795661135604319e-05,
+      "loss": 5.4362,
+      "step": 95
+    },
+    {
+      "epoch": 6.25,
+      "grad_norm": 1.8671875,
+      "learning_rate": 2.7855498176104434e-05,
+      "loss": 5.3928,
+      "step": 100
+    },
+    {
+      "epoch": 6.56,
+      "grad_norm": 1.8125,
+      "learning_rate": 1.9098300562505266e-05,
+      "loss": 5.5188,
+      "step": 105
+    },
+    {
+      "epoch": 6.88,
+      "grad_norm": 1.7421875,
+      "learning_rate": 1.1848148537532843e-05,
+      "loss": 5.4886,
+      "step": 110
+    },
+    {
+      "epoch": 7.19,
+      "grad_norm": 1.84375,
+      "learning_rate": 6.240098715597975e-06,
+      "loss": 5.3798,
+      "step": 115
+    },
+    {
+      "epoch": 7.5,
+      "grad_norm": 1.7265625,
+      "learning_rate": 2.3786184540455448e-06,
+      "loss": 5.4869,
+      "step": 120
+    },
+    {
+      "epoch": 7.81,
+      "grad_norm": 1.8984375,
+      "learning_rate": 3.3563982394704266e-07,
+      "loss": 5.4311,
+      "step": 125
+    },
+    {
+      "epoch": 8.0,
+      "step": 128,
+      "total_flos": 1049348792123392.0,
+      "train_loss": 5.653645019978285,
+      "train_runtime": 83.1481,
+      "train_samples_per_second": 24.15,
+      "train_steps_per_second": 1.539
     }
   ],
   "logging_steps": 5,
-  "max_steps": 32,
+  "max_steps": 128,
   "num_input_tokens_seen": 0,
-  "num_train_epochs": 4,
+  "num_train_epochs": 8,
   "save_steps": 100,
-  "total_flos": 535126081536000.0,
+  "total_flos": 1049348792123392.0,
   "train_batch_size": 16,
   "trial_name": null,
   "trial_params": null