End of training

README.md

@@ -16,7 +16,7 @@ should probably proofread and complete it, then remove this comment. -->
 
 This model is a fine-tuned version of [mistralai/Mistral-7B-v0.1](https://huggingface.co/mistralai/Mistral-7B-v0.1) on the None dataset.
 It achieves the following results on the evaluation set:
-- Loss: 1.1407
+- Loss: 1.1312
 
 ## Model description
 
@@ -35,7 +35,7 @@ More information needed
 ### Training hyperparameters
 
 The following hyperparameters were used during training:
-- learning_rate: 5e-05
+- learning_rate: 3e-05
 - train_batch_size: 4
 - eval_batch_size: 4
 - seed: 42
@@ -50,11 +50,11 @@ The following hyperparameters were used during training:
 
 | Training Loss | Epoch | Step | Validation Loss |
 |:-------------:|:-----:|:----:|:---------------:|
-| 1.0956        | 0.34  | 50   | 1.1443          |
-| 1.0635        | 0.68  | 100  | 1.1242          |
-| 1.0208        | 1.02  | 150  | 1.1208          |
-| 0.9505        | 1.36  | 200  | 1.1454          |
-| 0.9604        | 1.7   | 250  | 1.1407          |
+| 1.1066        | 0.34  | 50   | 1.1555          |
+| 1.0655        | 0.68  | 100  | 1.1333          |
+| 1.039         | 1.02  | 150  | 1.1279          |
+| 1.0166        | 1.36  | 200  | 1.1301          |
+| 1.0281        | 1.7   | 250  | 1.1312          |
 
 
 ### Framework versions

adapter_config.json

@@ -10,7 +10,7 @@
   "layers_to_transform": null,
   "loftq_config": {},
   "lora_alpha": 16,
-  "lora_dropout": 0.05,
+  "lora_dropout": 0.1,
   "megatron_config": null,
   "megatron_core": "megatron.core",
   "modules_to_save": null,
@@ -20,12 +20,12 @@
   "revision": null,
   "target_modules": [
     "up_proj",
-    "q_proj",
-    "down_proj",
     "o_proj",
+    "v_proj",
+    "q_proj",
     "k_proj",
-    "gate_proj",
-    "v_proj"
+    "down_proj",
+    "gate_proj"
   ],
   "task_type": "CAUSAL_LM"
 }

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:ce3c104d98913b0d1196f95870d59002ba294eb55e9e70d4cc69e72653f71914
+oid sha256:7c4d086cbb272795476d23287af5b86202bac7bff9ae8f867b0da552176415b7
 size 335605144

all_results.json

@@ -1,11 +1,11 @@
 {
     "epoch": 2.0,
-    "eval_loss": 1.1439318656921387,
-    "eval_runtime": 15.5616,
-    "eval_samples_per_second": 16.836,
-    "eval_steps_per_second": 4.241,
-    "train_loss": 1.0086434495692351,
-    "train_runtime": 872.7075,
-    "train_samples_per_second": 5.399,
-    "train_steps_per_second": 0.337
+    "eval_loss": 1.1402091979980469,
+    "eval_runtime": 15.8109,
+    "eval_samples_per_second": 16.571,
+    "eval_steps_per_second": 4.174,
+    "train_loss": 1.0460824836679057,
+    "train_runtime": 890.7293,
+    "train_samples_per_second": 5.29,
+    "train_steps_per_second": 0.33
 }

eval_results.json

@@ -1,7 +1,7 @@
 {
     "epoch": 2.0,
-    "eval_loss": 1.1439318656921387,
-    "eval_runtime": 15.5616,
-    "eval_samples_per_second": 16.836,
-    "eval_steps_per_second": 4.241
+    "eval_loss": 1.1402091979980469,
+    "eval_runtime": 15.8109,
+    "eval_samples_per_second": 16.571,
+    "eval_steps_per_second": 4.174
 }

metrics.json

@@ -1 +1 @@
-{"run_name": "./output", "train_runtime": 861.2997, "train_samples_per_second": 5.471, "train_steps_per_second": 0.341, "train_loss": 1.0092324691564858, "epoch": 2.0, "eval_loss": 1.1439318656921387, "eval_runtime": 15.5616, "eval_samples_per_second": 16.836, "eval_steps_per_second": 4.241}
+{"run_name": "./output", "train_runtime": 872.7075, "train_samples_per_second": 5.399, "train_steps_per_second": 0.337, "train_loss": 1.0086434495692351, "epoch": 2.0, "eval_loss": 1.1402091979980469, "eval_runtime": 15.8109, "eval_samples_per_second": 16.571, "eval_steps_per_second": 4.174}

tokenizer.json

@@ -2,7 +2,7 @@
   "version": "1.0",
   "truncation": {
     "direction": "Right",
-    "max_length": 512,
+    "max_length": 256,
     "strategy": "LongestFirst",
     "stride": 0
   },

train_results.json

@@ -1,7 +1,7 @@
 {
     "epoch": 2.0,
-    "train_loss": 1.0086434495692351,
-    "train_runtime": 872.7075,
-    "train_samples_per_second": 5.399,
-    "train_steps_per_second": 0.337
+    "train_loss": 1.0460824836679057,
+    "train_runtime": 890.7293,
+    "train_samples_per_second": 5.29,
+    "train_steps_per_second": 0.33
 }

trainer_state.json

@@ -10,226 +10,226 @@
   "log_history": [
     {
       "epoch": 0.07,
-      "learning_rate": 5e-05,
-      "loss": 1.1701,
+      "learning_rate": 3e-05,
+      "loss": 1.2187,
       "step": 10
     },
     {
       "epoch": 0.14,
-      "learning_rate": 5e-05,
-      "loss": 1.1349,
+      "learning_rate": 3e-05,
+      "loss": 1.1481,
       "step": 20
     },
     {
       "epoch": 0.2,
-      "learning_rate": 5e-05,
-      "loss": 1.124,
+      "learning_rate": 3e-05,
+      "loss": 1.1437,
       "step": 30
     },
     {
       "epoch": 0.27,
-      "learning_rate": 5e-05,
-      "loss": 1.0961,
+      "learning_rate": 3e-05,
+      "loss": 1.1076,
       "step": 40
     },
     {
       "epoch": 0.34,
-      "learning_rate": 5e-05,
-      "loss": 1.0956,
+      "learning_rate": 3e-05,
+      "loss": 1.1066,
       "step": 50
     },
     {
       "epoch": 0.34,
-      "eval_loss": 1.1443167924880981,
-      "eval_runtime": 15.8615,
-      "eval_samples_per_second": 16.518,
-      "eval_steps_per_second": 4.161,
+      "eval_loss": 1.1555328369140625,
+      "eval_runtime": 15.6804,
+      "eval_samples_per_second": 16.709,
+      "eval_steps_per_second": 4.209,
       "step": 50
     },
     {
       "epoch": 0.41,
-      "learning_rate": 5e-05,
-      "loss": 1.0875,
+      "learning_rate": 3e-05,
+      "loss": 1.0988,
       "step": 60
     },
     {
       "epoch": 0.48,
-      "learning_rate": 5e-05,
-      "loss": 1.0766,
+      "learning_rate": 3e-05,
+      "loss": 1.0917,
       "step": 70
     },
     {
       "epoch": 0.54,
-      "learning_rate": 5e-05,
-      "loss": 1.1265,
+      "learning_rate": 3e-05,
+      "loss": 1.1344,
       "step": 80
     },
     {
       "epoch": 0.61,
-      "learning_rate": 5e-05,
-      "loss": 1.1126,
+      "learning_rate": 3e-05,
+      "loss": 1.1167,
       "step": 90
     },
     {
       "epoch": 0.68,
-      "learning_rate": 5e-05,
-      "loss": 1.0635,
+      "learning_rate": 3e-05,
+      "loss": 1.0655,
       "step": 100
     },
     {
       "epoch": 0.68,
-      "eval_loss": 1.124241828918457,
-      "eval_runtime": 15.8746,
-      "eval_samples_per_second": 16.504,
-      "eval_steps_per_second": 4.158,
+      "eval_loss": 1.133277177810669,
+      "eval_runtime": 15.7203,
+      "eval_samples_per_second": 16.666,
+      "eval_steps_per_second": 4.198,
       "step": 100
     },
     {
       "epoch": 0.75,
-      "learning_rate": 5e-05,
-      "loss": 1.0434,
+      "learning_rate": 3e-05,
+      "loss": 1.0595,
       "step": 110
     },
     {
       "epoch": 0.81,
-      "learning_rate": 5e-05,
-      "loss": 1.1071,
+      "learning_rate": 3e-05,
+      "loss": 1.1085,
       "step": 120
     },
     {
       "epoch": 0.88,
-      "learning_rate": 5e-05,
-      "loss": 1.0926,
+      "learning_rate": 3e-05,
+      "loss": 1.0945,
       "step": 130
     },
     {
       "epoch": 0.95,
-      "learning_rate": 5e-05,
-      "loss": 1.0677,
+      "learning_rate": 3e-05,
+      "loss": 1.0726,
       "step": 140
     },
     {
       "epoch": 1.02,
-      "learning_rate": 5e-05,
-      "loss": 1.0208,
+      "learning_rate": 3e-05,
+      "loss": 1.039,
       "step": 150
     },
     {
       "epoch": 1.02,
-      "eval_loss": 1.1208317279815674,
-      "eval_runtime": 15.9371,
-      "eval_samples_per_second": 16.44,
-      "eval_steps_per_second": 4.141,
+      "eval_loss": 1.1278975009918213,
+      "eval_runtime": 15.7982,
+      "eval_samples_per_second": 16.584,
+      "eval_steps_per_second": 4.178,
       "step": 150
     },
     {
       "epoch": 1.09,
-      "learning_rate": 5e-05,
-      "loss": 0.9435,
+      "learning_rate": 3e-05,
+      "loss": 0.9891,
       "step": 160
     },
     {
       "epoch": 1.15,
-      "learning_rate": 5e-05,
-      "loss": 1.0034,
+      "learning_rate": 3e-05,
+      "loss": 1.0617,
       "step": 170
     },
     {
       "epoch": 1.22,
-      "learning_rate": 5e-05,
-      "loss": 0.9165,
+      "learning_rate": 3e-05,
+      "loss": 0.98,
       "step": 180
     },
     {
       "epoch": 1.29,
-      "learning_rate": 5e-05,
-      "loss": 0.8848,
+      "learning_rate": 3e-05,
+      "loss": 0.9531,
       "step": 190
     },
     {
       "epoch": 1.36,
-      "learning_rate": 5e-05,
-      "loss": 0.9505,
+      "learning_rate": 3e-05,
+      "loss": 1.0166,
       "step": 200
     },
     {
       "epoch": 1.36,
-      "eval_loss": 1.1453500986099243,
-      "eval_runtime": 15.8776,
-      "eval_samples_per_second": 16.501,
-      "eval_steps_per_second": 4.157,
+      "eval_loss": 1.1301459074020386,
+      "eval_runtime": 15.7471,
+      "eval_samples_per_second": 16.638,
+      "eval_steps_per_second": 4.191,
       "step": 200
     },
     {
       "epoch": 1.43,
-      "learning_rate": 5e-05,
-      "loss": 0.9259,
+      "learning_rate": 3e-05,
+      "loss": 0.9869,
       "step": 210
     },
     {
       "epoch": 1.49,
-      "learning_rate": 5e-05,
-      "loss": 0.8383,
+      "learning_rate": 3e-05,
+      "loss": 0.9066,
       "step": 220
     },
     {
       "epoch": 1.56,
-      "learning_rate": 5e-05,
-      "loss": 0.9317,
+      "learning_rate": 3e-05,
+      "loss": 0.9884,
       "step": 230
     },
     {
       "epoch": 1.63,
-      "learning_rate": 5e-05,
-      "loss": 0.9336,
+      "learning_rate": 3e-05,
+      "loss": 0.9885,
       "step": 240
     },
     {
       "epoch": 1.7,
-      "learning_rate": 5e-05,
-      "loss": 0.9604,
+      "learning_rate": 3e-05,
+      "loss": 1.0281,
       "step": 250
     },
     {
       "epoch": 1.7,
-      "eval_loss": 1.1406688690185547,
-      "eval_runtime": 15.8651,
-      "eval_samples_per_second": 16.514,
-      "eval_steps_per_second": 4.16,
+      "eval_loss": 1.1312241554260254,
+      "eval_runtime": 15.6611,
+      "eval_samples_per_second": 16.729,
+      "eval_steps_per_second": 4.214,
       "step": 250
     },
     {
       "epoch": 1.77,
-      "learning_rate": 5e-05,
-      "loss": 0.8884,
+      "learning_rate": 3e-05,
+      "loss": 0.9723,
       "step": 260
     },
     {
       "epoch": 1.83,
-      "learning_rate": 5e-05,
-      "loss": 0.9757,
+      "learning_rate": 3e-05,
+      "loss": 1.0428,
       "step": 270
     },
     {
       "epoch": 1.9,
-      "learning_rate": 5e-05,
-      "loss": 0.8977,
+      "learning_rate": 3e-05,
+      "loss": 0.961,
       "step": 280
     },
     {
       "epoch": 1.97,
-      "learning_rate": 5e-05,
-      "loss": 0.8645,
+      "learning_rate": 3e-05,
+      "loss": 0.9311,
       "step": 290
     },
     {
       "epoch": 2.0,
       "step": 294,
-      "total_flos": 7.119009482145792e+16,
-      "train_loss": 1.0086434495692351,
-      "train_runtime": 872.7075,
-      "train_samples_per_second": 5.399,
-      "train_steps_per_second": 0.337
+      "total_flos": 7.002531209517466e+16,
+      "train_loss": 1.0460824836679057,
+      "train_runtime": 890.7293,
+      "train_samples_per_second": 5.29,
+      "train_steps_per_second": 0.33
     }
   ],
   "logging_steps": 10,
@@ -237,7 +237,7 @@
   "num_input_tokens_seen": 0,
   "num_train_epochs": 2,
   "save_steps": 250,
-  "total_flos": 7.119009482145792e+16,
+  "total_flos": 7.002531209517466e+16,
   "train_batch_size": 4,
   "trial_name": null,
   "trial_params": null

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:cc89040a415c72790d5fd7b4dba8d591833a97e8e25926a0bf37205f0fa84bae
-size 6712
+oid sha256:7354e725785da8b7b0a039b25e28ac3e5e811514125d390c40da0e60c2ef9472
+size 6648