Model save

README.md

@@ -2,16 +2,9 @@
 license: mit
 library_name: peft
 tags:
-- alignment-handbook
 - trl
 - sft
 - generated_from_trainer
-- trl
-- sft
-- generated_from_trainer
-datasets:
-- ChenWu98/skills_metaphor_chat
-- ChenWu98/skills_red_herring_chat
 base_model: HuggingFaceH4/zephyr-7b-beta
 model-index:
 - name: skills_metaphor_chat-skills_red_herring_chat-lora
@@ -23,9 +16,9 @@ should probably proofread and complete it, then remove this comment. -->
 
 # skills_metaphor_chat-skills_red_herring_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_metaphor_chat and the ChenWu98/skills_red_herring_chat datasets.
+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: 0.2245
+- Loss: 0.2636
 
 ## Model description
 
@@ -54,14 +47,13 @@ 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: 1.0
 
 ### Training results
 
 | Training Loss | Epoch | Step | Validation Loss |
 |:-------------:|:-----:|:----:|:---------------:|
-| 0.2788        | 0.96  | 18   | 0.2390          |
-| 0.1993        | 1.92  | 36   | 0.2245          |
+| 0.2847        | 0.96  | 18   | 0.2636          |
 
 
 ### Framework versions

all_results.json

@@ -1,13 +1,13 @@
 {
-    "epoch": 1.92,
-    "eval_loss": 0.2245262861251831,
-    "eval_runtime": 10.1729,
+    "epoch": 0.96,
+    "eval_loss": 0.26359862089157104,
+    "eval_runtime": 7.9312,
     "eval_samples": 200,
-    "eval_samples_per_second": 19.66,
-    "eval_steps_per_second": 2.458,
-    "train_loss": 0.5313632095025645,
-    "train_runtime": 554.6126,
+    "eval_samples_per_second": 25.217,
+    "eval_steps_per_second": 3.152,
+    "train_loss": 0.7799174222681258,
+    "train_runtime": 192.3889,
     "train_samples": 600,
-    "train_samples_per_second": 2.164,
-    "train_steps_per_second": 0.065
+    "train_samples_per_second": 3.119,
+    "train_steps_per_second": 0.094
 }

eval_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 1.92,
-    "eval_loss": 0.2245262861251831,
-    "eval_runtime": 10.1729,
+    "epoch": 0.96,
+    "eval_loss": 0.26359862089157104,
+    "eval_runtime": 7.9312,
     "eval_samples": 200,
-    "eval_samples_per_second": 19.66,
-    "eval_steps_per_second": 2.458
+    "eval_samples_per_second": 25.217,
+    "eval_steps_per_second": 3.152
 }

train_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 1.92,
-    "train_loss": 0.5313632095025645,
-    "train_runtime": 554.6126,
+    "epoch": 0.96,
+    "train_loss": 0.7799174222681258,
+    "train_runtime": 192.3889,
     "train_samples": 600,
-    "train_samples_per_second": 2.164,
-    "train_steps_per_second": 0.065
+    "train_samples_per_second": 3.119,
+    "train_steps_per_second": 0.094
 }

trainer_state.json

@@ -1,93 +1,61 @@
 {
   "best_metric": null,
   "best_model_checkpoint": null,
-  "epoch": 1.92,
+  "epoch": 0.96,
   "eval_steps": 500,
-  "global_step": 36,
+  "global_step": 18,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
       "epoch": 0.05,
-      "learning_rate": 5e-05,
+      "learning_rate": 0.0001,
       "loss": 2.272,
       "step": 1
     },
     {
       "epoch": 0.27,
-      "learning_rate": 0.0001995184726672197,
-      "loss": 1.954,
+      "learning_rate": 0.00018314696123025454,
+      "loss": 1.7677,
       "step": 5
     },
     {
       "epoch": 0.53,
-      "learning_rate": 0.00018314696123025454,
-      "loss": 0.6371,
+      "learning_rate": 0.0001,
+      "loss": 0.5037,
       "step": 10
     },
     {
       "epoch": 0.8,
-      "learning_rate": 0.0001471396736825998,
-      "loss": 0.2788,
+      "learning_rate": 1.6853038769745467e-05,
+      "loss": 0.2847,
       "step": 15
     },
     {
       "epoch": 0.96,
-      "eval_loss": 0.23904697597026825,
-      "eval_runtime": 11.1391,
-      "eval_samples_per_second": 17.955,
-      "eval_steps_per_second": 2.244,
+      "eval_loss": 0.26359862089157104,
+      "eval_runtime": 8.8309,
+      "eval_samples_per_second": 22.648,
+      "eval_steps_per_second": 2.831,
       "step": 18
     },
     {
-      "epoch": 1.07,
-      "learning_rate": 0.0001,
-      "loss": 0.2252,
-      "step": 20
-    },
-    {
-      "epoch": 1.33,
-      "learning_rate": 5.286032631740023e-05,
-      "loss": 0.2128,
-      "step": 25
-    },
-    {
-      "epoch": 1.6,
-      "learning_rate": 1.6853038769745467e-05,
-      "loss": 0.2122,
-      "step": 30
-    },
-    {
-      "epoch": 1.87,
-      "learning_rate": 4.815273327803182e-07,
-      "loss": 0.1993,
-      "step": 35
-    },
-    {
-      "epoch": 1.92,
-      "eval_loss": 0.2245262861251831,
-      "eval_runtime": 9.6856,
-      "eval_samples_per_second": 20.649,
-      "eval_steps_per_second": 2.581,
-      "step": 36
-    },
-    {
-      "epoch": 1.92,
-      "step": 36,
-      "total_flos": 40159503319040.0,
-      "train_loss": 0.5313632095025645,
-      "train_runtime": 554.6126,
-      "train_samples_per_second": 2.164,
-      "train_steps_per_second": 0.065
+      "epoch": 0.96,
+      "step": 18,
+      "total_flos": 20001168556032.0,
+      "train_loss": 0.7799174222681258,
+      "train_runtime": 192.3889,
+      "train_samples_per_second": 3.119,
+      "train_steps_per_second": 0.094
     }
   ],
   "logging_steps": 5,
-  "max_steps": 36,
+  "max_steps": 18,
   "num_input_tokens_seen": 0,
-  "num_train_epochs": 2,
+  "num_train_epochs": 1,
   "save_steps": 500,
-  "total_flos": 40159503319040.0,
+  "total_flos": 20001168556032.0,
   "train_batch_size": 4,
   "trial_name": null,
   "trial_params": null