Model save

README.md

@@ -2,15 +2,12 @@
 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: []
@@ -21,7 +18,7 @@ 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
 
@@ -51,7 +48,7 @@ The following hyperparameters were used during training:
 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
 - lr_scheduler_type: cosine
 - lr_scheduler_warmup_ratio: 0.1
-- num_epochs: 1
+- num_epochs: 3.0
 
 ### Training results
 

all_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 1.0,
-    "train_loss": 6.2669631242752075,
-    "train_runtime": 15.755,
+    "epoch": 3.0,
+    "train_loss": 6.028296748797099,
+    "train_runtime": 40.098,
     "train_samples": 91,
-    "train_samples_per_second": 15.931,
-    "train_steps_per_second": 0.508
+    "train_samples_per_second": 18.779,
+    "train_steps_per_second": 0.599
 }

config.json

@@ -34,6 +34,6 @@
   },
   "torch_dtype": "bfloat16",
   "transformers_version": "4.39.0.dev0",
-  "use_cache": true,
+  "use_cache": false,
   "vocab_size": 50257
 }

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:acece083b2f894d0cd48e3130ee0b501b3bb60367c3fbcee5e0f8e570e99ecc5
+oid sha256:50fab2e3ac84c5849e05637af1ba07db099619c06e9f634c104e534fd87e44c8
 size 248894656

train_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 1.0,
-    "train_loss": 6.2669631242752075,
-    "train_runtime": 15.755,
+    "epoch": 3.0,
+    "train_loss": 6.028296748797099,
+    "train_runtime": 40.098,
     "train_samples": 91,
-    "train_samples_per_second": 15.931,
-    "train_steps_per_second": 0.508
+    "train_samples_per_second": 18.779,
+    "train_steps_per_second": 0.599
 }

trainer_state.json

@@ -1,9 +1,9 @@
 {
   "best_metric": null,
   "best_model_checkpoint": null,
-  "epoch": 1.0,
+  "epoch": 3.0,
   "eval_steps": 500,
-  "global_step": 8,
+  "global_step": 24,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
@@ -11,33 +11,54 @@
     {
       "epoch": 0.12,
       "grad_norm": 6.875,
-      "learning_rate": 0.0002,
+      "learning_rate": 6.666666666666667e-05,
       "loss": 6.4878,
       "step": 1
     },
     {
       "epoch": 0.62,
-      "grad_norm": 4.25,
-      "learning_rate": 7.774790660436858e-05,
-      "loss": 6.3165,
+      "grad_norm": 4.15625,
+      "learning_rate": 0.0001955572805786141,
+      "loss": 6.2679,
       "step": 5
     },
     {
-      "epoch": 1.0,
-      "step": 8,
-      "total_flos": 133781520384000.0,
-      "train_loss": 6.2669631242752075,
-      "train_runtime": 15.755,
-      "train_samples_per_second": 15.931,
-      "train_steps_per_second": 0.508
+      "epoch": 1.25,
+      "grad_norm": 3.0,
+      "learning_rate": 0.00015000000000000001,
+      "loss": 6.0743,
+      "step": 10
+    },
+    {
+      "epoch": 1.88,
+      "grad_norm": 2.75,
+      "learning_rate": 7.774790660436858e-05,
+      "loss": 5.924,
+      "step": 15
+    },
+    {
+      "epoch": 2.5,
+      "grad_norm": 2.734375,
+      "learning_rate": 1.7376122568400532e-05,
+      "loss": 5.9329,
+      "step": 20
+    },
+    {
+      "epoch": 3.0,
+      "step": 24,
+      "total_flos": 401344561152000.0,
+      "train_loss": 6.028296748797099,
+      "train_runtime": 40.098,
+      "train_samples_per_second": 18.779,
+      "train_steps_per_second": 0.599
     }
   ],
   "logging_steps": 5,
-  "max_steps": 8,
+  "max_steps": 24,
   "num_input_tokens_seen": 0,
-  "num_train_epochs": 1,
+  "num_train_epochs": 3,
   "save_steps": 100,
-  "total_flos": 133781520384000.0,
+  "total_flos": 401344561152000.0,
   "train_batch_size": 16,
   "trial_name": null,
   "trial_params": null

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:092e442b91773e2a02344d8deebf5757a73512c4fb70805bfb7af2061a7e4c65
+oid sha256:1b5d24a506bf06211115f2794625e7452cf246540d312ae8034060bb4acbca34
 size 4984