Model save

README.md

@@ -15,7 +15,7 @@ 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 an unknown dataset.
 It achieves the following results on the evaluation set:
-- Loss: 1.0159
+- Loss: 1.0239
 
 ## Model description
 
@@ -40,8 +40,8 @@ The following hyperparameters were used during training:
 - seed: 42
 - distributed_type: multi-GPU
 - num_devices: 4
-- gradient_accumulation_steps: 128
-- total_train_batch_size: 1024
+- gradient_accumulation_steps: 256
+- total_train_batch_size: 2048
 - total_eval_batch_size: 16
 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
 - lr_scheduler_type: cosine
@@ -51,7 +51,7 @@ The following hyperparameters were used during training:
 
 | Training Loss | Epoch | Step | Validation Loss |
 |:-------------:|:-----:|:----:|:---------------:|
-| 1.0105        | 0.67  | 136  | 1.0159          |
+| 1.0224        | 0.67  | 68   | 1.0240          |
 
 
 ### Framework versions

adapter_config.json

@@ -16,10 +16,10 @@
   "rank_pattern": {},
   "revision": null,
   "target_modules": [
-    "v_proj",
-    "q_proj",
     "k_proj",
-    "o_proj"
+    "q_proj",
+    "o_proj",
+    "v_proj"
   ],
   "task_type": "CAUSAL_LM"
 }

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:d1faa1046cc90a5392fa57389464099e2a0d4321a30e0284bbd50592e2872053
+oid sha256:9c0c9f426276e25b3a38839f3edf593998480ae8aaf9a7dd86f4091e745ac676
 size 218138576

all_results.json

@@ -1,13 +1,13 @@
 {
     "epoch": 0.67,
-    "eval_loss": 1.0158774852752686,
-    "eval_runtime": 2342.0762,
+    "eval_loss": 1.0239042043685913,
+    "eval_runtime": 2000.5663,
     "eval_samples": 23110,
-    "eval_samples_per_second": 9.867,
-    "eval_steps_per_second": 0.617,
-    "train_loss": 1.0439463178901112,
-    "train_runtime": 75333.0344,
+    "eval_samples_per_second": 11.552,
+    "eval_steps_per_second": 0.722,
+    "train_loss": 1.0572445532854866,
+    "train_runtime": 62451.9899,
     "train_samples": 207865,
-    "train_samples_per_second": 2.759,
-    "train_steps_per_second": 0.003
+    "train_samples_per_second": 3.328,
+    "train_steps_per_second": 0.002
 }

eval_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 0.67,
-    "eval_loss": 1.0158774852752686,
-    "eval_runtime": 2342.0762,
+    "eval_loss": 1.0239042043685913,
+    "eval_runtime": 2000.5663,
     "eval_samples": 23110,
-    "eval_samples_per_second": 9.867,
-    "eval_steps_per_second": 0.617
+    "eval_samples_per_second": 11.552,
+    "eval_steps_per_second": 0.722
 }

train_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 0.67,
-    "train_loss": 1.0439463178901112,
-    "train_runtime": 75333.0344,
+    "train_loss": 1.0572445532854866,
+    "train_runtime": 62451.9899,
     "train_samples": 207865,
-    "train_samples_per_second": 2.759,
-    "train_steps_per_second": 0.003
+    "train_samples_per_second": 3.328,
+    "train_steps_per_second": 0.002
 }

trainer_state.json

@@ -3,202 +3,118 @@
   "best_model_checkpoint": null,
   "epoch": 0.6699507389162561,
   "eval_steps": 500,
-  "global_step": 136,
+  "global_step": 68,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
-      "epoch": 0.0,
-      "learning_rate": 1.9998802517966852e-05,
-      "loss": 1.1454,
+      "epoch": 0.01,
+      "learning_rate": 1.999516282291988e-05,
+      "loss": 1.1268,
       "step": 1
     },
-    {
-      "epoch": 0.02,
-      "learning_rate": 1.997007728639956e-05,
-      "loss": 1.1416,
-      "step": 5
-    },
     {
       "epoch": 0.05,
-      "learning_rate": 1.9880488219356086e-05,
-      "loss": 1.1347,
-      "step": 10
-    },
-    {
-      "epoch": 0.07,
-      "learning_rate": 1.973176894846855e-05,
-      "loss": 1.108,
-      "step": 15
+      "learning_rate": 1.987930439740757e-05,
+      "loss": 1.1283,
+      "step": 5
     },
     {
       "epoch": 0.1,
-      "learning_rate": 1.9524809490566878e-05,
-      "loss": 1.0977,
-      "step": 20
-    },
-    {
-      "epoch": 0.12,
-      "learning_rate": 1.926084840336821e-05,
-      "loss": 1.0881,
-      "step": 25
+      "learning_rate": 1.95201310753273e-05,
+      "loss": 1.1092,
+      "step": 10
     },
     {
       "epoch": 0.15,
-      "learning_rate": 1.894146537327533e-05,
-      "loss": 1.0713,
-      "step": 30
-    },
-    {
-      "epoch": 0.17,
-      "learning_rate": 1.8568571761675893e-05,
-      "loss": 1.0523,
-      "step": 35
+      "learning_rate": 1.8931150161867917e-05,
+      "loss": 1.0944,
+      "step": 15
     },
     {
       "epoch": 0.2,
-      "learning_rate": 1.814439916631857e-05,
-      "loss": 1.0533,
-      "step": 40
-    },
-    {
-      "epoch": 0.22,
-      "learning_rate": 1.7671486066220965e-05,
-      "loss": 1.0427,
-      "step": 45
+      "learning_rate": 1.8126579138282502e-05,
+      "loss": 1.0769,
+      "step": 20
     },
     {
       "epoch": 0.25,
-      "learning_rate": 1.7152662630033506e-05,
-      "loss": 1.0308,
-      "step": 50
-    },
-    {
-      "epoch": 0.27,
-      "learning_rate": 1.659103377877423e-05,
-      "loss": 1.0395,
-      "step": 55
+      "learning_rate": 1.7125839641475074e-05,
+      "loss": 1.0658,
+      "step": 25
     },
     {
       "epoch": 0.3,
-      "learning_rate": 1.598996060429634e-05,
-      "loss": 1.0268,
-      "step": 60
-    },
-    {
-      "epoch": 0.32,
-      "learning_rate": 1.5353040254690396e-05,
-      "loss": 1.0304,
-      "step": 65
+      "learning_rate": 1.595308864276666e-05,
+      "loss": 1.0595,
+      "step": 30
     },
     {
       "epoch": 0.34,
-      "learning_rate": 1.4684084406997903e-05,
-      "loss": 1.0264,
-      "step": 70
-    },
-    {
-      "epoch": 0.37,
-      "learning_rate": 1.3987096456067236e-05,
-      "loss": 1.0233,
-      "step": 75
+      "learning_rate": 1.4636635319853274e-05,
+      "loss": 1.0497,
+      "step": 35
     },
     {
       "epoch": 0.39,
-      "learning_rate": 1.3266247556066122e-05,
-      "loss": 1.0094,
-      "step": 80
-    },
-    {
-      "epoch": 0.42,
-      "learning_rate": 1.252585165803135e-05,
-      "loss": 1.0238,
-      "step": 85
+      "learning_rate": 1.3208257698153677e-05,
+      "loss": 1.0332,
+      "step": 40
     },
     {
       "epoch": 0.44,
-      "learning_rate": 1.1770339692844484e-05,
-      "loss": 1.013,
-      "step": 90
-    },
-    {
-      "epoch": 0.47,
-      "learning_rate": 1.1004233054136726e-05,
-      "loss": 1.022,
-      "step": 95
+      "learning_rate": 1.1702435557223988e-05,
+      "loss": 1.0312,
+      "step": 45
     },
     {
       "epoch": 0.49,
-      "learning_rate": 1.0232116539815558e-05,
-      "loss": 1.0152,
-      "step": 100
-    },
-    {
-      "epoch": 0.52,
-      "learning_rate": 9.458610914145826e-06,
-      "loss": 1.0143,
-      "step": 105
+      "learning_rate": 1.0155518119203511e-05,
+      "loss": 1.029,
+      "step": 50
     },
     {
       "epoch": 0.54,
-      "learning_rate": 8.688345254588579e-06,
-      "loss": 1.019,
-      "step": 110
-    },
-    {
-      "epoch": 0.57,
-      "learning_rate": 7.92592924888925e-06,
-      "loss": 1.0175,
-      "step": 115
+      "learning_rate": 8.604846610560771e-06,
+      "loss": 1.0258,
+      "step": 55
     },
     {
       "epoch": 0.59,
-      "learning_rate": 7.175925608204428e-06,
-      "loss": 1.0169,
-      "step": 120
-    },
-    {
-      "epoch": 0.62,
-      "learning_rate": 6.442822761362015e-06,
-      "loss": 1.0167,
-      "step": 125
+      "learning_rate": 7.0878528777274814e-06,
+      "loss": 1.0258,
+      "step": 60
     },
     {
       "epoch": 0.64,
-      "learning_rate": 5.731007993667155e-06,
-      "loss": 1.0117,
-      "step": 130
-    },
-    {
-      "epoch": 0.67,
-      "learning_rate": 5.044741191005908e-06,
-      "loss": 1.0105,
-      "step": 135
+      "learning_rate": 5.64115581524629e-06,
+      "loss": 1.0224,
+      "step": 65
     },
     {
       "epoch": 0.67,
-      "eval_loss": 1.0159411430358887,
-      "eval_runtime": 2343.4311,
-      "eval_samples_per_second": 9.862,
-      "eval_steps_per_second": 0.617,
-      "step": 136
+      "eval_loss": 1.0239689350128174,
+      "eval_runtime": 2002.0819,
+      "eval_samples_per_second": 11.543,
+      "eval_steps_per_second": 0.722,
+      "step": 68
     },
     {
       "epoch": 0.67,
-      "step": 136,
-      "total_flos": 1.227605110369026e+19,
-      "train_loss": 1.0439463178901112,
-      "train_runtime": 75333.0344,
-      "train_samples_per_second": 2.759,
-      "train_steps_per_second": 0.003
+      "step": 68,
+      "total_flos": 1.2276051069330522e+19,
+      "train_loss": 1.0572445532854866,
+      "train_runtime": 62451.9899,
+      "train_samples_per_second": 3.328,
+      "train_steps_per_second": 0.002
     }
   ],
   "logging_steps": 5,
-  "max_steps": 203,
+  "max_steps": 101,
   "num_train_epochs": 1,
   "save_steps": 500,
-  "total_flos": 1.227605110369026e+19,
+  "total_flos": 1.2276051069330522e+19,
   "trial_name": null,
   "trial_params": null
 }

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:999658f84171b894d7a9a748797387b84f386965029df0d7ab314d87f0c3fa63
+oid sha256:36facd59ad2755544b27eb1044191870ef084d9fb2b53ff4c6f47efcc03ef423
 size 4155