Model save

README.md

@@ -2,15 +2,9 @@
 license: mit
 library_name: peft
 tags:
-- alignment-handbook
 - trl
 - sft
 - generated_from_trainer
-- trl
-- sft
-- generated_from_trainer
-datasets:
-- ChenWu98/skills_red_herring_metaphor_chat
 base_model: HuggingFaceH4/zephyr-7b-beta
 model-index:
 - name: skills_red_herring_metaphor_chat-lora
@@ -22,9 +16,9 @@ should probably proofread and complete it, then remove this comment. -->
 
 # skills_red_herring_metaphor_chat-lora
 
-This model is a fine-tuned version of [HuggingFaceH4/zephyr-7b-beta](https://huggingface.co/HuggingFaceH4/zephyr-7b-beta) on the ChenWu98/skills_red_herring_metaphor_chat dataset.
+This model is a fine-tuned version of [HuggingFaceH4/zephyr-7b-beta](https://huggingface.co/HuggingFaceH4/zephyr-7b-beta) on the None dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.3410
+- Loss: 0.2900
 
 ## Model description
 
@@ -53,14 +47,16 @@ 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: 2.0
+- num_epochs: 4.0
 
 ### Training results
 
 | Training Loss | Epoch | Step | Validation Loss |
 |:-------------:|:-----:|:----:|:---------------:|
-| 1.6966        | 0.96  | 9    | 0.4003          |
-| 0.3419        | 1.92  | 18   | 0.3410          |
+| 1.8767        | 0.96  | 9    | 0.4226          |
+| 0.3349        | 1.92  | 18   | 0.3096          |
+| 0.2678        | 2.99  | 28   | 0.2913          |
+| 0.2538        | 3.84  | 36   | 0.2900          |
 
 
 ### Framework versions

all_results.json

@@ -1,13 +1,13 @@
 {
-    "epoch": 1.92,
-    "eval_loss": 0.3410404324531555,
-    "eval_runtime": 5.319,
+    "epoch": 3.84,
+    "eval_loss": 0.2900236248970032,
+    "eval_runtime": 9.9443,
     "eval_samples": 100,
-    "eval_samples_per_second": 18.8,
-    "eval_steps_per_second": 2.444,
-    "train_loss": 0.8035296764638689,
-    "train_runtime": 289.9665,
+    "eval_samples_per_second": 10.056,
+    "eval_steps_per_second": 1.307,
+    "train_loss": 0.5664321829875311,
+    "train_runtime": 739.8338,
     "train_samples": 300,
-    "train_samples_per_second": 2.069,
-    "train_steps_per_second": 0.062
+    "train_samples_per_second": 1.622,
+    "train_steps_per_second": 0.049
 }

eval_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 1.92,
-    "eval_loss": 0.3410404324531555,
-    "eval_runtime": 5.319,
+    "epoch": 3.84,
+    "eval_loss": 0.2900236248970032,
+    "eval_runtime": 9.9443,
     "eval_samples": 100,
-    "eval_samples_per_second": 18.8,
-    "eval_steps_per_second": 2.444
+    "eval_samples_per_second": 10.056,
+    "eval_steps_per_second": 1.307
 }

train_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 1.92,
-    "train_loss": 0.8035296764638689,
-    "train_runtime": 289.9665,
+    "epoch": 3.84,
+    "train_loss": 0.5664321829875311,
+    "train_runtime": 739.8338,
     "train_samples": 300,
-    "train_samples_per_second": 2.069,
-    "train_steps_per_second": 0.062
+    "train_samples_per_second": 1.622,
+    "train_steps_per_second": 0.049
 }

trainer_state.json

@@ -1,69 +1,109 @@
 {
   "best_metric": null,
   "best_model_checkpoint": null,
-  "epoch": 1.92,
+  "epoch": 3.84,
   "eval_steps": 500,
-  "global_step": 18,
+  "global_step": 36,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
       "epoch": 0.11,
-      "learning_rate": 0.0001,
+      "learning_rate": 5e-05,
       "loss": 2.1556,
       "step": 1
     },
     {
       "epoch": 0.53,
-      "learning_rate": 0.00018314696123025454,
-      "loss": 1.6966,
+      "learning_rate": 0.0001995184726672197,
+      "loss": 1.8767,
       "step": 5
     },
     {
       "epoch": 0.96,
-      "eval_loss": 0.4002748727798462,
-      "eval_runtime": 5.8667,
-      "eval_samples_per_second": 17.045,
-      "eval_steps_per_second": 2.216,
+      "eval_loss": 0.4225959777832031,
+      "eval_runtime": 8.0339,
+      "eval_samples_per_second": 12.447,
+      "eval_steps_per_second": 1.618,
       "step": 9
     },
     {
       "epoch": 1.07,
-      "learning_rate": 0.0001,
-      "loss": 0.5695,
+      "learning_rate": 0.00018314696123025454,
+      "loss": 0.685,
       "step": 10
     },
     {
       "epoch": 1.6,
-      "learning_rate": 1.6853038769745467e-05,
-      "loss": 0.3419,
+      "learning_rate": 0.0001471396736825998,
+      "loss": 0.3349,
       "step": 15
     },
     {
       "epoch": 1.92,
-      "eval_loss": 0.3410404324531555,
-      "eval_runtime": 7.4319,
-      "eval_samples_per_second": 13.456,
-      "eval_steps_per_second": 1.749,
+      "eval_loss": 0.30958235263824463,
+      "eval_runtime": 4.9729,
+      "eval_samples_per_second": 20.109,
+      "eval_steps_per_second": 2.614,
       "step": 18
     },
     {
-      "epoch": 1.92,
-      "step": 18,
-      "total_flos": 27578625622016.0,
-      "train_loss": 0.8035296764638689,
-      "train_runtime": 289.9665,
-      "train_samples_per_second": 2.069,
-      "train_steps_per_second": 0.062
+      "epoch": 2.13,
+      "learning_rate": 0.0001,
+      "loss": 0.2955,
+      "step": 20
+    },
+    {
+      "epoch": 2.67,
+      "learning_rate": 5.286032631740023e-05,
+      "loss": 0.2678,
+      "step": 25
+    },
+    {
+      "epoch": 2.99,
+      "eval_loss": 0.2913205325603485,
+      "eval_runtime": 8.0251,
+      "eval_samples_per_second": 12.461,
+      "eval_steps_per_second": 1.62,
+      "step": 28
+    },
+    {
+      "epoch": 3.2,
+      "learning_rate": 1.6853038769745467e-05,
+      "loss": 0.2584,
+      "step": 30
+    },
+    {
+      "epoch": 3.73,
+      "learning_rate": 4.815273327803182e-07,
+      "loss": 0.2538,
+      "step": 35
+    },
+    {
+      "epoch": 3.84,
+      "eval_loss": 0.2900236248970032,
+      "eval_runtime": 5.7265,
+      "eval_samples_per_second": 17.463,
+      "eval_steps_per_second": 2.27,
+      "step": 36
+    },
+    {
+      "epoch": 3.84,
+      "step": 36,
+      "total_flos": 55410626723840.0,
+      "train_loss": 0.5664321829875311,
+      "train_runtime": 739.8338,
+      "train_samples_per_second": 1.622,
+      "train_steps_per_second": 0.049
     }
   ],
   "logging_steps": 5,
-  "max_steps": 18,
+  "max_steps": 36,
   "num_input_tokens_seen": 0,
-  "num_train_epochs": 2,
+  "num_train_epochs": 4,
   "save_steps": 500,
-  "total_flos": 27578625622016.0,
+  "total_flos": 55410626723840.0,
   "train_batch_size": 4,
   "trial_name": null,
   "trial_params": null