Model save

README.md

@@ -5,7 +5,6 @@ base_model: alignment-handbook/zephyr-7b-sft-full
 tags:
 - trl
 - sft
-- alignment-handbook
 - generated_from_trainer
 datasets:
 - generator
@@ -21,7 +20,7 @@ should probably proofread and complete it, then remove this comment. -->
 
 This model is a fine-tuned version of [alignment-handbook/zephyr-7b-sft-full](https://huggingface.co/alignment-handbook/zephyr-7b-sft-full) on the generator dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.4948
+- Loss: 0.4842
 
 ## Model description
 
@@ -59,13 +58,14 @@ The following hyperparameters were used during training:
 |:-------------:|:-----:|:----:|:---------------:|
 | 1.1139        | 1.0   | 1    | 1.1133          |
 | 1.1139        | 2.0   | 2    | 1.2947          |
-| 1.1139        | 3.0   | 3    | 1.0839          |
-| 1.1139        | 4.0   | 4    | 0.8506          |
-| 1.0867        | 5.0   | 5    | 0.7521          |
-| 1.0867        | 6.0   | 6    | 0.6283          |
-| 1.0867        | 7.0   | 7    | 0.5677          |
-| 1.0867        | 8.0   | 8    | 0.5233          |
-| 1.0867        | 9.0   | 9    | 0.4948          |
+| 1.1139        | 3.0   | 3    | 1.0840          |
+| 1.1139        | 4.0   | 4    | 0.8533          |
+| 1.0874        | 5.0   | 5    | 0.7534          |
+| 1.0874        | 6.0   | 6    | 0.6295          |
+| 1.0874        | 7.0   | 7    | 0.5679          |
+| 1.0874        | 8.0   | 8    | 0.5225          |
+| 1.0874        | 9.0   | 9    | 0.4936          |
+| 0.5928        | 10.0  | 10   | 0.4842          |
 
 
 ### Framework versions

all_results.json

@@ -1,14 +1,9 @@
 {
     "epoch": 10.0,
-    "eval_loss": 1.0543607473373413,
-    "eval_runtime": 4.7419,
-    "eval_samples": 424,
-    "eval_samples_per_second": 6.537,
-    "eval_steps_per_second": 0.211,
     "total_flos": 8375186227200.0,
-    "train_loss": 0.0,
-    "train_runtime": 1.4498,
+    "train_loss": 0.8427552580833435,
+    "train_runtime": 304.7513,
     "train_samples": 851,
-    "train_samples_per_second": 331.078,
-    "train_steps_per_second": 6.897
+    "train_samples_per_second": 1.575,
+    "train_steps_per_second": 0.033
 }

train_results.json

@@ -1,9 +1,9 @@
 {
     "epoch": 10.0,
     "total_flos": 8375186227200.0,
-    "train_loss": 0.0,
-    "train_runtime": 1.4498,
+    "train_loss": 0.8427552580833435,
+    "train_runtime": 304.7513,
     "train_samples": 851,
-    "train_samples_per_second": 331.078,
-    "train_steps_per_second": 6.897
+    "train_samples_per_second": 1.575,
+    "train_steps_per_second": 0.033
 }

trainer_state.json

@@ -10,7 +10,7 @@
   "log_history": [
     {
       "epoch": 1.0,
-      "grad_norm": 18.364192038860114,
+      "grad_norm": 18.363699029410846,
       "learning_rate": 1e-05,
       "loss": 1.1139,
       "step": 1
@@ -18,97 +18,105 @@
     {
       "epoch": 1.0,
       "eval_loss": 1.1133455038070679,
-      "eval_runtime": 2.8092,
-      "eval_samples_per_second": 17.087,
-      "eval_steps_per_second": 0.356,
+      "eval_runtime": 2.7908,
+      "eval_samples_per_second": 17.199,
+      "eval_steps_per_second": 0.358,
       "step": 1
     },
     {
       "epoch": 2.0,
-      "eval_loss": 1.2946728467941284,
-      "eval_runtime": 2.6987,
-      "eval_samples_per_second": 17.786,
-      "eval_steps_per_second": 0.371,
+      "eval_loss": 1.2946919202804565,
+      "eval_runtime": 2.6849,
+      "eval_samples_per_second": 17.878,
+      "eval_steps_per_second": 0.372,
       "step": 2
     },
     {
       "epoch": 3.0,
-      "eval_loss": 1.0838984251022339,
-      "eval_runtime": 2.6695,
-      "eval_samples_per_second": 17.981,
-      "eval_steps_per_second": 0.375,
+      "eval_loss": 1.084022879600525,
+      "eval_runtime": 2.6785,
+      "eval_samples_per_second": 17.921,
+      "eval_steps_per_second": 0.373,
       "step": 3
     },
     {
       "epoch": 4.0,
-      "eval_loss": 0.8506386280059814,
-      "eval_runtime": 2.6924,
-      "eval_samples_per_second": 17.828,
+      "eval_loss": 0.8533250689506531,
+      "eval_runtime": 2.6989,
+      "eval_samples_per_second": 17.785,
       "eval_steps_per_second": 0.371,
       "step": 4
     },
     {
       "epoch": 5.0,
-      "grad_norm": 19.328671929154766,
+      "grad_norm": 20.40164947404992,
       "learning_rate": 5.8682408883346535e-06,
-      "loss": 1.0867,
+      "loss": 1.0874,
       "step": 5
     },
     {
       "epoch": 5.0,
-      "eval_loss": 0.7521028518676758,
-      "eval_runtime": 2.6708,
-      "eval_samples_per_second": 17.972,
-      "eval_steps_per_second": 0.374,
+      "eval_loss": 0.7533825039863586,
+      "eval_runtime": 2.6807,
+      "eval_samples_per_second": 17.906,
+      "eval_steps_per_second": 0.373,
       "step": 5
     },
     {
       "epoch": 6.0,
-      "eval_loss": 0.628281831741333,
-      "eval_runtime": 2.7091,
-      "eval_samples_per_second": 17.718,
-      "eval_steps_per_second": 0.369,
+      "eval_loss": 0.6295236945152283,
+      "eval_runtime": 2.68,
+      "eval_samples_per_second": 17.911,
+      "eval_steps_per_second": 0.373,
       "step": 6
     },
     {
       "epoch": 7.0,
-      "eval_loss": 0.5677043795585632,
-      "eval_runtime": 2.6968,
-      "eval_samples_per_second": 17.799,
-      "eval_steps_per_second": 0.371,
+      "eval_loss": 0.567865788936615,
+      "eval_runtime": 2.6867,
+      "eval_samples_per_second": 17.866,
+      "eval_steps_per_second": 0.372,
       "step": 7
     },
     {
       "epoch": 8.0,
-      "eval_loss": 0.5233017802238464,
-      "eval_runtime": 2.6981,
-      "eval_samples_per_second": 17.79,
-      "eval_steps_per_second": 0.371,
+      "eval_loss": 0.5224726796150208,
+      "eval_runtime": 2.6682,
+      "eval_samples_per_second": 17.989,
+      "eval_steps_per_second": 0.375,
       "step": 8
     },
     {
       "epoch": 9.0,
-      "eval_loss": 0.4948171079158783,
-      "eval_runtime": 2.6682,
-      "eval_samples_per_second": 17.99,
-      "eval_steps_per_second": 0.375,
+      "eval_loss": 0.4935952425003052,
+      "eval_runtime": 2.672,
+      "eval_samples_per_second": 17.964,
+      "eval_steps_per_second": 0.374,
       "step": 9
     },
     {
       "epoch": 10.0,
-      "grad_norm": 4.900067571807948,
+      "grad_norm": 4.71394718703185,
       "learning_rate": 0.0,
-      "loss": 0.5927,
+      "loss": 0.5928,
+      "step": 10
+    },
+    {
+      "epoch": 10.0,
+      "eval_loss": 0.48424649238586426,
+      "eval_runtime": 2.6942,
+      "eval_samples_per_second": 17.816,
+      "eval_steps_per_second": 0.371,
       "step": 10
     },
     {
       "epoch": 10.0,
       "step": 10,
       "total_flos": 8375186227200.0,
-      "train_loss": 0.0,
-      "train_runtime": 1.4498,
-      "train_samples_per_second": 331.078,
-      "train_steps_per_second": 6.897
+      "train_loss": 0.8427552580833435,
+      "train_runtime": 304.7513,
+      "train_samples_per_second": 1.575,
+      "train_steps_per_second": 0.033
     }
   ],
   "logging_steps": 5,