Model save

README.md

@@ -1,11 +1,8 @@
 ---
 base_model: mistralai/Mistral-7B-Instruct-v0.1
-datasets:
-- generator
 library_name: peft
 license: apache-2.0
 tags:
-- alignment-handbook
 - trl
 - sft
 - generated_from_trainer
@@ -19,9 +16,9 @@ should probably proofread and complete it, then remove this comment. -->
 
 # mistral_default_lora
 
-This model is a fine-tuned version of [mistralai/Mistral-7B-Instruct-v0.1](https://huggingface.co/mistralai/Mistral-7B-Instruct-v0.1) on the generator dataset.
+This model is a fine-tuned version of [mistralai/Mistral-7B-Instruct-v0.1](https://huggingface.co/mistralai/Mistral-7B-Instruct-v0.1) on the None dataset.
 It achieves the following results on the evaluation set:
-- Loss: 1.2007
+- Loss: 1.2705
 
 ## Model description
 
@@ -50,27 +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: 15
+- training_steps: 20
 
 ### Training results
 
 | Training Loss | Epoch | Step | Validation Loss |
 |:-------------:|:-----:|:----:|:---------------:|
-| 2.054         | 1.0   | 1    | 2.0482          |
-| 2.054         | 2.0   | 2    | 1.7447          |
-| 2.054         | 3.0   | 3    | 1.4612          |
-| 2.054         | 4.0   | 4    | 1.3664          |
-| 1.636         | 5.0   | 5    | 1.3182          |
-| 1.636         | 6.0   | 6    | 1.2509          |
-| 1.636         | 7.0   | 7    | 1.2262          |
-| 1.636         | 8.0   | 8    | 1.2207          |
-| 1.636         | 9.0   | 9    | 1.2038          |
-| 0.8947        | 10.0  | 10   | 1.1923          |
-| 0.8947        | 11.0  | 11   | 1.1926          |
-| 0.8947        | 12.0  | 12   | 1.1967          |
-| 0.8947        | 13.0  | 13   | 1.1993          |
-| 0.8947        | 14.0  | 14   | 1.2008          |
-| 0.5319        | 15.0  | 15   | 1.2007          |
+| 2.0498        | 0.4   | 5    | 1.5630          |
+| 1.412         | 0.8   | 10   | 1.3434          |
+| 1.1535        | 1.2   | 15   | 1.2856          |
+| 1.0048        | 1.6   | 20   | 1.2705          |
 
 
 ### Framework versions

all_results.json

@@ -1,14 +1,9 @@
 {
-    "epoch": 15.0,
-    "eval_loss": 1.200675368309021,
-    "eval_runtime": 0.4082,
-    "eval_samples": 20,
-    "eval_samples_per_second": 2.45,
-    "eval_steps_per_second": 2.45,
-    "total_flos": 2787467329536.0,
-    "train_loss": 1.0487482070922851,
-    "train_runtime": 88.2648,
+    "epoch": 1.6,
+    "total_flos": 548330668032.0,
+    "train_loss": 1.421624231338501,
+    "train_runtime": 82.0086,
     "train_samples": 100,
-    "train_samples_per_second": 1.02,
-    "train_steps_per_second": 0.17
+    "train_samples_per_second": 1.951,
+    "train_steps_per_second": 0.244
 }

train_results.json

@@ -1,9 +1,9 @@
 {
-    "epoch": 15.0,
-    "total_flos": 2787467329536.0,
-    "train_loss": 1.0487482070922851,
-    "train_runtime": 88.2648,
+    "epoch": 1.6,
+    "total_flos": 548330668032.0,
+    "train_loss": 1.421624231338501,
+    "train_runtime": 82.0086,
     "train_samples": 100,
-    "train_samples_per_second": 1.02,
-    "train_steps_per_second": 0.17
+    "train_samples_per_second": 1.951,
+    "train_steps_per_second": 0.244
 }

trainer_state.json

@@ -1,175 +1,94 @@
 {
   "best_metric": null,
   "best_model_checkpoint": null,
-  "epoch": 15.0,
-  "eval_steps": 500,
-  "global_step": 15,
+  "epoch": 1.6,
+  "eval_steps": 5,
+  "global_step": 20,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
-      "epoch": 1.0,
-      "grad_norm": 2.775644201582239,
+      "epoch": 0.08,
+      "grad_norm": 1.7204986331972942,
       "learning_rate": 0.0001,
-      "loss": 2.054,
+      "loss": 2.3815,
       "step": 1
     },
     {
-      "epoch": 1.0,
-      "eval_loss": 2.0482397079467773,
-      "eval_runtime": 0.4302,
-      "eval_samples_per_second": 2.324,
-      "eval_steps_per_second": 2.324,
-      "step": 1
-    },
-    {
-      "epoch": 2.0,
-      "eval_loss": 1.7446707487106323,
-      "eval_runtime": 1.5149,
-      "eval_samples_per_second": 0.66,
-      "eval_steps_per_second": 0.66,
-      "step": 2
-    },
-    {
-      "epoch": 3.0,
-      "eval_loss": 1.4611855745315552,
-      "eval_runtime": 0.37,
-      "eval_samples_per_second": 2.703,
-      "eval_steps_per_second": 2.703,
-      "step": 3
-    },
-    {
-      "epoch": 4.0,
-      "eval_loss": 1.3664478063583374,
-      "eval_runtime": 0.3722,
-      "eval_samples_per_second": 2.686,
-      "eval_steps_per_second": 2.686,
-      "step": 4
-    },
-    {
-      "epoch": 5.0,
-      "grad_norm": 0.7397193908660463,
-      "learning_rate": 0.00017485107481711012,
-      "loss": 1.636,
+      "epoch": 0.4,
+      "grad_norm": 3.2293833170556625,
+      "learning_rate": 0.00018660254037844388,
+      "loss": 2.0498,
       "step": 5
     },
     {
-      "epoch": 5.0,
-      "eval_loss": 1.3181551694869995,
-      "eval_runtime": 0.3714,
-      "eval_samples_per_second": 2.693,
-      "eval_steps_per_second": 2.693,
+      "epoch": 0.4,
+      "eval_loss": 1.5629652738571167,
+      "eval_runtime": 3.0527,
+      "eval_samples_per_second": 6.552,
+      "eval_steps_per_second": 1.638,
       "step": 5
     },
     {
-      "epoch": 6.0,
-      "eval_loss": 1.2508817911148071,
-      "eval_runtime": 0.3699,
-      "eval_samples_per_second": 2.704,
-      "eval_steps_per_second": 2.704,
-      "step": 6
-    },
-    {
-      "epoch": 7.0,
-      "eval_loss": 1.2261731624603271,
-      "eval_runtime": 0.3682,
-      "eval_samples_per_second": 2.716,
-      "eval_steps_per_second": 2.716,
-      "step": 7
-    },
-    {
-      "epoch": 8.0,
-      "eval_loss": 1.2207050323486328,
-      "eval_runtime": 0.3695,
-      "eval_samples_per_second": 2.706,
-      "eval_steps_per_second": 2.706,
-      "step": 8
-    },
-    {
-      "epoch": 9.0,
-      "eval_loss": 1.2037922143936157,
-      "eval_runtime": 0.3696,
-      "eval_samples_per_second": 2.706,
-      "eval_steps_per_second": 2.706,
-      "step": 9
-    },
-    {
-      "epoch": 10.0,
-      "grad_norm": 0.7005825834442886,
-      "learning_rate": 6.453951129574644e-05,
-      "loss": 0.8947,
+      "epoch": 0.8,
+      "grad_norm": 0.9620857356197902,
+      "learning_rate": 0.00011736481776669306,
+      "loss": 1.412,
       "step": 10
     },
     {
-      "epoch": 10.0,
-      "eval_loss": 1.1922813653945923,
-      "eval_runtime": 0.37,
-      "eval_samples_per_second": 2.703,
-      "eval_steps_per_second": 2.703,
+      "epoch": 0.8,
+      "eval_loss": 1.3433921337127686,
+      "eval_runtime": 1.8318,
+      "eval_samples_per_second": 10.918,
+      "eval_steps_per_second": 2.73,
       "step": 10
     },
     {
-      "epoch": 11.0,
-      "eval_loss": 1.1926474571228027,
-      "eval_runtime": 0.3699,
-      "eval_samples_per_second": 2.703,
-      "eval_steps_per_second": 2.703,
-      "step": 11
-    },
-    {
-      "epoch": 12.0,
-      "eval_loss": 1.1966757774353027,
-      "eval_runtime": 0.3703,
-      "eval_samples_per_second": 2.7,
-      "eval_steps_per_second": 2.7,
-      "step": 12
-    },
-    {
-      "epoch": 13.0,
-      "eval_loss": 1.1992850303649902,
-      "eval_runtime": 0.3694,
-      "eval_samples_per_second": 2.707,
-      "eval_steps_per_second": 2.707,
-      "step": 13
+      "epoch": 1.2,
+      "grad_norm": 1.0025754462857623,
+      "learning_rate": 3.5721239031346066e-05,
+      "loss": 1.1535,
+      "step": 15
     },
     {
-      "epoch": 14.0,
-      "eval_loss": 1.200769305229187,
-      "eval_runtime": 0.3692,
-      "eval_samples_per_second": 2.709,
-      "eval_steps_per_second": 2.709,
-      "step": 14
+      "epoch": 1.2,
+      "eval_loss": 1.2855589389801025,
+      "eval_runtime": 1.8278,
+      "eval_samples_per_second": 10.942,
+      "eval_steps_per_second": 2.736,
+      "step": 15
     },
     {
-      "epoch": 15.0,
-      "grad_norm": 0.5467216766155176,
+      "epoch": 1.6,
+      "grad_norm": 1.0149567514468123,
       "learning_rate": 0.0,
-      "loss": 0.5319,
-      "step": 15
+      "loss": 1.0048,
+      "step": 20
     },
     {
-      "epoch": 15.0,
-      "eval_loss": 1.200675368309021,
-      "eval_runtime": 0.3831,
-      "eval_samples_per_second": 2.61,
-      "eval_steps_per_second": 2.61,
-      "step": 15
+      "epoch": 1.6,
+      "eval_loss": 1.2704746723175049,
+      "eval_runtime": 1.8362,
+      "eval_samples_per_second": 10.892,
+      "eval_steps_per_second": 2.723,
+      "step": 20
     },
     {
-      "epoch": 15.0,
-      "step": 15,
-      "total_flos": 2787467329536.0,
-      "train_loss": 1.0487482070922851,
-      "train_runtime": 88.2648,
-      "train_samples_per_second": 1.02,
-      "train_steps_per_second": 0.17
+      "epoch": 1.6,
+      "step": 20,
+      "total_flos": 548330668032.0,
+      "train_loss": 1.421624231338501,
+      "train_runtime": 82.0086,
+      "train_samples_per_second": 1.951,
+      "train_steps_per_second": 0.244
     }
   ],
   "logging_steps": 5,
-  "max_steps": 15,
+  "max_steps": 20,
   "num_input_tokens_seen": 0,
-  "num_train_epochs": 15,
+  "num_train_epochs": 2,
   "save_steps": 1000,
   "stateful_callbacks": {
     "TrainerControl": {
@@ -183,7 +102,7 @@
       "attributes": {}
     }
   },
-  "total_flos": 2787467329536.0,
+  "total_flos": 548330668032.0,
   "train_batch_size": 4,
   "trial_name": null,
   "trial_params": null