Model save

README.md

@@ -34,11 +34,11 @@ More information needed
 
 The following hyperparameters were used during training:
 - learning_rate: 0.0001
-- train_batch_size: 6
+- train_batch_size: 16
 - eval_batch_size: 8
 - seed: 42
 - gradient_accumulation_steps: 16
-- total_train_batch_size: 96
+- total_train_batch_size: 256
 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
 - lr_scheduler_type: linear
 - lr_scheduler_warmup_steps: 2

adapter_config.json

@@ -20,12 +20,12 @@
   "rank_pattern": {},
   "revision": null,
   "target_modules": [
-    "k_proj",
-    "down_proj",
     "q_proj",
+    "down_proj",
+    "up_proj",
+    "k_proj",
     "o_proj",
     "v_proj",
-    "up_proj",
     "gate_proj"
   ],
   "task_type": "CAUSAL_LM",

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:39ef7b5ed66ff345f822092e69d270531391d309f346e882bac6e246a0b98536
+oid sha256:d209c54a5d95be6e1246f9254828bb85ebf2c9da9c8a5016d8aa65f0d84eb0ca
 size 1635969696

trainer_state.json

@@ -1,172 +1,81 @@
 {
   "best_metric": null,
   "best_model_checkpoint": null,
-  "epoch": 3.999089792017476,
+  "epoch": 4.0,
   "eval_steps": 500,
-  "global_step": 5492,
+  "global_step": 2060,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
-      "epoch": 0.18204159650480134,
-      "grad_norm": 1.178935170173645,
-      "learning_rate": 9.548269581056467e-05,
-      "loss": 1.0653,
+      "epoch": 0.4854368932038835,
+      "grad_norm": 0.5123344659805298,
+      "learning_rate": 8.794946550048592e-05,
+      "loss": 0.9496,
       "step": 250
     },
     {
-      "epoch": 0.3640831930096027,
-      "grad_norm": 1.0788908004760742,
-      "learning_rate": 9.092896174863389e-05,
-      "loss": 0.5664,
+      "epoch": 0.970873786407767,
+      "grad_norm": 0.8029014468193054,
+      "learning_rate": 7.580174927113704e-05,
+      "loss": 0.6426,
       "step": 500
     },
     {
-      "epoch": 0.546124789514404,
-      "grad_norm": 0.8843607902526855,
-      "learning_rate": 8.63752276867031e-05,
-      "loss": 0.4896,
+      "epoch": 1.4563106796116505,
+      "grad_norm": 0.9907676577568054,
+      "learning_rate": 6.365403304178815e-05,
+      "loss": 0.5632,
       "step": 750
     },
     {
-      "epoch": 0.7281663860192054,
-      "grad_norm": 0.8972049951553345,
-      "learning_rate": 8.182149362477231e-05,
-      "loss": 0.4628,
+      "epoch": 1.941747572815534,
+      "grad_norm": 0.9990822672843933,
+      "learning_rate": 5.150631681243926e-05,
+      "loss": 0.5117,
       "step": 1000
     },
     {
-      "epoch": 0.9102079825240067,
-      "grad_norm": 0.7201969027519226,
-      "learning_rate": 7.726775956284153e-05,
-      "loss": 0.4461,
+      "epoch": 2.4271844660194173,
+      "grad_norm": 0.7601630687713623,
+      "learning_rate": 3.9358600583090386e-05,
+      "loss": 0.4772,
       "step": 1250
     },
     {
-      "epoch": 1.092249579028808,
-      "grad_norm": 0.7767462134361267,
-      "learning_rate": 7.271402550091076e-05,
-      "loss": 0.4344,
+      "epoch": 2.912621359223301,
+      "grad_norm": 0.7860077023506165,
+      "learning_rate": 2.72108843537415e-05,
+      "loss": 0.4564,
       "step": 1500
     },
     {
-      "epoch": 1.2742911755336095,
-      "grad_norm": 0.7544079422950745,
-      "learning_rate": 6.816029143897996e-05,
-      "loss": 0.425,
+      "epoch": 3.3980582524271843,
+      "grad_norm": 0.5943681597709656,
+      "learning_rate": 1.5063168124392615e-05,
+      "loss": 0.4392,
       "step": 1750
     },
     {
-      "epoch": 1.4563327720384107,
-      "grad_norm": 0.7420190572738647,
-      "learning_rate": 6.360655737704918e-05,
-      "loss": 0.4211,
+      "epoch": 3.883495145631068,
+      "grad_norm": 0.5319840908050537,
+      "learning_rate": 2.915451895043732e-06,
+      "loss": 0.4318,
       "step": 2000
     },
     {
-      "epoch": 1.6383743685432122,
-      "grad_norm": 0.8790698051452637,
-      "learning_rate": 5.90528233151184e-05,
-      "loss": 0.4094,
-      "step": 2250
-    },
-    {
-      "epoch": 1.8204159650480136,
-      "grad_norm": 0.8602856397628784,
-      "learning_rate": 5.449908925318762e-05,
-      "loss": 0.3953,
-      "step": 2500
-    },
-    {
-      "epoch": 2.0024575615528146,
-      "grad_norm": 0.9606088995933533,
-      "learning_rate": 4.994535519125683e-05,
-      "loss": 0.3816,
-      "step": 2750
-    },
-    {
-      "epoch": 2.184499158057616,
-      "grad_norm": 1.0402580499649048,
-      "learning_rate": 4.539162112932605e-05,
-      "loss": 0.3574,
-      "step": 3000
-    },
-    {
-      "epoch": 2.3665407545624175,
-      "grad_norm": 1.0433388948440552,
-      "learning_rate": 4.083788706739526e-05,
-      "loss": 0.3438,
-      "step": 3250
-    },
-    {
-      "epoch": 2.548582351067219,
-      "grad_norm": 1.0822595357894897,
-      "learning_rate": 3.6284153005464486e-05,
-      "loss": 0.326,
-      "step": 3500
-    },
-    {
-      "epoch": 2.7306239475720204,
-      "grad_norm": 1.1437020301818848,
-      "learning_rate": 3.17304189435337e-05,
-      "loss": 0.3155,
-      "step": 3750
-    },
-    {
-      "epoch": 2.9126655440768214,
-      "grad_norm": 1.1588534116744995,
-      "learning_rate": 2.7176684881602916e-05,
-      "loss": 0.3043,
-      "step": 4000
-    },
-    {
-      "epoch": 3.094707140581623,
-      "grad_norm": 1.0676839351654053,
-      "learning_rate": 2.262295081967213e-05,
-      "loss": 0.2888,
-      "step": 4250
-    },
-    {
-      "epoch": 3.2767487370864243,
-      "grad_norm": 1.149923324584961,
-      "learning_rate": 1.806921675774135e-05,
-      "loss": 0.2762,
-      "step": 4500
-    },
-    {
-      "epoch": 3.458790333591226,
-      "grad_norm": 1.1598175764083862,
-      "learning_rate": 1.3515482695810567e-05,
-      "loss": 0.2721,
-      "step": 4750
-    },
-    {
-      "epoch": 3.640831930096027,
-      "grad_norm": 1.1374166011810303,
-      "learning_rate": 8.961748633879782e-06,
-      "loss": 0.2682,
-      "step": 5000
-    },
-    {
-      "epoch": 3.8228735266008282,
-      "grad_norm": 1.123226523399353,
-      "learning_rate": 4.408014571948998e-06,
-      "loss": 0.2637,
-      "step": 5250
-    },
-    {
-      "epoch": 3.999089792017476,
-      "step": 5492,
-      "total_flos": 1.2110720614916813e+18,
-      "train_loss": 0.3990837633392712,
-      "train_runtime": 19256.2222,
-      "train_samples_per_second": 27.386,
-      "train_steps_per_second": 0.285
+      "epoch": 4.0,
+      "step": 2060,
+      "total_flos": 1.4445804612483994e+18,
+      "train_loss": 0.5552371376926459,
+      "train_runtime": 17453.3458,
+      "train_samples_per_second": 30.215,
+      "train_steps_per_second": 0.118
     }
   ],
   "logging_steps": 250,
-  "max_steps": 5492,
+  "max_steps": 2060,
   "num_input_tokens_seen": 0,
   "num_train_epochs": 4,
   "save_steps": 500,
@@ -182,8 +91,8 @@
       "attributes": {}
     }
   },
-  "total_flos": 1.2110720614916813e+18,
-  "train_batch_size": 6,
+  "total_flos": 1.4445804612483994e+18,
+  "train_batch_size": 16,
   "trial_name": null,
   "trial_params": null
 }

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:c43f3ec46e656601dba5c8176f6da304521a9b5a664289cfb437165d2d3c7259
+oid sha256:617a99532cd91765e0705e15346b67e9640bce2631da172c620125a4f5bbf68b
 size 5240