Model save

README.md

@@ -20,7 +20,7 @@ should probably proofread and complete it, then remove this comment. -->
 
 This model is a fine-tuned version of [google/gemma-7b](https://huggingface.co/google/gemma-7b) on the generator dataset.
 It achieves the following results on the evaluation set:
-- Loss: 2.8222
+- Loss: 2.8123
 
 ## Model description
 
@@ -57,7 +57,7 @@ The following hyperparameters were used during training:
 
 | Training Loss | Epoch  | Step | Validation Loss |
 |:-------------:|:------:|:----:|:---------------:|
-| 1.3179        | 0.9977 | 221  | 2.8222          |
+| 1.3142        | 0.9977 | 221  | 2.8123          |
 
 
 ### Framework versions

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:2a146c48f6da46e4999c4af192732c884617f6953364f04986e7b0412e46a0f8
+oid sha256:330572751706f31589c9562c802bbd155a7191c76cd866552840bf809721303a
 size 12859872

all_results.json

@@ -1,14 +1,9 @@
 {
     "epoch": 0.9977426636568849,
-    "eval_loss": 2.8222014904022217,
-    "eval_runtime": 1.2855,
-    "eval_samples": 25,
-    "eval_samples_per_second": 16.336,
-    "eval_steps_per_second": 0.778,
     "total_flos": 6.738683116954583e+17,
-    "train_loss": 5.562472593730392,
-    "train_runtime": 2738.0196,
+    "train_loss": 5.577704090877895,
+    "train_runtime": 575.1744,
     "train_samples": 129221,
-    "train_samples_per_second": 10.349,
-    "train_steps_per_second": 0.081
+    "train_samples_per_second": 49.265,
+    "train_steps_per_second": 0.384
 }

runs/Nov18_05-23-30_main-lora-gemma7b-summarization-0-0/events.out.tfevents.1731926156.main-lora-gemma7b-summarization-0-0.456.0

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:2dfc8fd61b59b95f5d039092be36ed4be112701dadb4dbd11e406d5e3b8dbc36
-size 14260
+oid sha256:208af1ca59aaf7a9d9779f59cb336c4e796753bc05741fa12223503b7ac1fc72
+size 15729

train_results.json

@@ -1,9 +1,9 @@
 {
     "epoch": 0.9977426636568849,
     "total_flos": 6.738683116954583e+17,
-    "train_loss": 5.562472593730392,
-    "train_runtime": 2738.0196,
+    "train_loss": 5.577704090877895,
+    "train_runtime": 575.1744,
     "train_samples": 129221,
-    "train_samples_per_second": 10.349,
-    "train_steps_per_second": 0.081
+    "train_samples_per_second": 49.265,
+    "train_steps_per_second": 0.384
 }

trainer_state.json

@@ -10,335 +10,335 @@
   "log_history": [
     {
       "epoch": 0.004514672686230248,
-      "grad_norm": 200.3717803955078,
+      "grad_norm": 199.82510375976562,
       "learning_rate": 8.695652173913044e-06,
-      "loss": 40.073,
+      "loss": 39.0472,
       "step": 1
     },
     {
       "epoch": 0.022573363431151242,
-      "grad_norm": 131.19760131835938,
+      "grad_norm": 178.4315185546875,
       "learning_rate": 4.347826086956522e-05,
-      "loss": 39.1612,
+      "loss": 39.3436,
       "step": 5
     },
     {
       "epoch": 0.045146726862302484,
-      "grad_norm": 41.251068115234375,
+      "grad_norm": 40.531532287597656,
       "learning_rate": 8.695652173913044e-05,
-      "loss": 30.2659,
+      "loss": 30.4834,
       "step": 10
     },
     {
       "epoch": 0.06772009029345373,
-      "grad_norm": 12.021024703979492,
+      "grad_norm": 12.844618797302246,
       "learning_rate": 0.00013043478260869567,
-      "loss": 23.7836,
+      "loss": 23.9376,
       "step": 15
     },
     {
       "epoch": 0.09029345372460497,
-      "grad_norm": 13.451888084411621,
+      "grad_norm": 15.092730522155762,
       "learning_rate": 0.00017391304347826088,
-      "loss": 20.8756,
+      "loss": 20.9382,
       "step": 20
     },
     {
       "epoch": 0.11286681715575621,
-      "grad_norm": 4.8614888191223145,
+      "grad_norm": 4.834229469299316,
       "learning_rate": 0.00019994965423831854,
-      "loss": 19.0103,
+      "loss": 19.0458,
       "step": 25
     },
     {
       "epoch": 0.13544018058690746,
-      "grad_norm": 8.32628059387207,
+      "grad_norm": 8.221763610839844,
       "learning_rate": 0.00019938384644612543,
-      "loss": 18.0406,
+      "loss": 18.0883,
       "step": 30
     },
     {
       "epoch": 0.1580135440180587,
-      "grad_norm": 17.078899383544922,
+      "grad_norm": 16.106189727783203,
       "learning_rate": 0.00019819286972627066,
-      "loss": 16.4692,
+      "loss": 16.5366,
       "step": 35
     },
     {
       "epoch": 0.18058690744920994,
-      "grad_norm": 24.7756404876709,
+      "grad_norm": 24.50678825378418,
       "learning_rate": 0.00019638421585599423,
-      "loss": 13.1214,
+      "loss": 13.2306,
       "step": 40
     },
     {
       "epoch": 0.20316027088036118,
-      "grad_norm": 31.732440948486328,
+      "grad_norm": 32.74522399902344,
       "learning_rate": 0.00019396926207859084,
-      "loss": 8.5547,
+      "loss": 8.7225,
       "step": 45
     },
     {
       "epoch": 0.22573363431151242,
-      "grad_norm": 9.871573448181152,
+      "grad_norm": 9.876773834228516,
       "learning_rate": 0.00019096319953545185,
-      "loss": 4.2086,
+      "loss": 4.3489,
       "step": 50
     },
     {
       "epoch": 0.24830699774266365,
-      "grad_norm": 8.936306953430176,
+      "grad_norm": 8.977824211120605,
       "learning_rate": 0.00018738493770697852,
-      "loss": 2.9412,
+      "loss": 2.9443,
       "step": 55
     },
     {
       "epoch": 0.2708803611738149,
-      "grad_norm": 3.45522141456604,
+      "grad_norm": 3.1251025199890137,
       "learning_rate": 0.00018325698546347715,
-      "loss": 2.4618,
+      "loss": 2.4464,
       "step": 60
     },
     {
       "epoch": 0.29345372460496616,
-      "grad_norm": 1.241241693496704,
+      "grad_norm": 1.4109467267990112,
       "learning_rate": 0.00017860530947427875,
-      "loss": 2.1257,
+      "loss": 2.0921,
       "step": 65
     },
     {
       "epoch": 0.3160270880361174,
-      "grad_norm": 1.5993455648422241,
+      "grad_norm": 1.3768062591552734,
       "learning_rate": 0.00017345917086575332,
-      "loss": 1.8936,
+      "loss": 1.8687,
       "step": 70
     },
     {
       "epoch": 0.33860045146726864,
-      "grad_norm": 1.037279486656189,
+      "grad_norm": 1.025557279586792,
       "learning_rate": 0.00016785094115571322,
-      "loss": 1.7682,
+      "loss": 1.7492,
       "step": 75
     },
     {
       "epoch": 0.3611738148984199,
-      "grad_norm": 1.5379009246826172,
+      "grad_norm": 1.717323660850525,
       "learning_rate": 0.00016181589862206052,
-      "loss": 1.6873,
+      "loss": 1.6714,
       "step": 80
     },
     {
       "epoch": 0.3837471783295711,
-      "grad_norm": 1.3465474843978882,
+      "grad_norm": 1.101912021636963,
       "learning_rate": 0.00015539200638661104,
-      "loss": 1.6315,
+      "loss": 1.6189,
       "step": 85
     },
     {
       "epoch": 0.40632054176072235,
-      "grad_norm": 0.7901816964149475,
+      "grad_norm": 0.8903200030326843,
       "learning_rate": 0.00014861967361004687,
-      "loss": 1.5786,
+      "loss": 1.5678,
       "step": 90
     },
     {
       "epoch": 0.4288939051918736,
-      "grad_norm": 1.0765341520309448,
+      "grad_norm": 1.1036828756332397,
       "learning_rate": 0.00014154150130018866,
-      "loss": 1.5464,
+      "loss": 1.5342,
       "step": 95
     },
     {
       "epoch": 0.45146726862302483,
-      "grad_norm": 1.1750948429107666,
+      "grad_norm": 1.4335256814956665,
       "learning_rate": 0.00013420201433256689,
-      "loss": 1.5079,
+      "loss": 1.496,
       "step": 100
     },
     {
       "epoch": 0.47404063205417607,
-      "grad_norm": 1.1422131061553955,
+      "grad_norm": 1.3628050088882446,
       "learning_rate": 0.00012664738136900348,
-      "loss": 1.488,
+      "loss": 1.4813,
       "step": 105
     },
     {
       "epoch": 0.4966139954853273,
-      "grad_norm": 1.0460481643676758,
+      "grad_norm": 1.7230000495910645,
       "learning_rate": 0.00011892512443604102,
-      "loss": 1.4541,
+      "loss": 1.4475,
       "step": 110
     },
     {
       "epoch": 0.5191873589164786,
-      "grad_norm": 1.787310242652893,
+      "grad_norm": 2.127577066421509,
       "learning_rate": 0.00011108381999010111,
-      "loss": 1.4428,
+      "loss": 1.4365,
       "step": 115
     },
     {
       "epoch": 0.5417607223476298,
-      "grad_norm": 0.8201248645782471,
+      "grad_norm": 1.139145851135254,
       "learning_rate": 0.00010317279334980678,
-      "loss": 1.4187,
+      "loss": 1.413,
       "step": 120
     },
     {
       "epoch": 0.5643340857787811,
-      "grad_norm": 0.8423493504524231,
+      "grad_norm": 1.1161638498306274,
       "learning_rate": 9.524180841762577e-05,
-      "loss": 1.4086,
+      "loss": 1.4018,
       "step": 125
     },
     {
       "epoch": 0.5869074492099323,
-      "grad_norm": 2.127373218536377,
+      "grad_norm": 1.6872923374176025,
       "learning_rate": 8.734075464262507e-05,
-      "loss": 1.3947,
+      "loss": 1.3883,
       "step": 130
     },
     {
       "epoch": 0.6094808126410836,
-      "grad_norm": 0.8410171270370483,
+      "grad_norm": 1.0828768014907837,
       "learning_rate": 7.951933319348095e-05,
-      "loss": 1.3942,
+      "loss": 1.3903,
       "step": 135
     },
     {
       "epoch": 0.6320541760722348,
-      "grad_norm": 0.7370259761810303,
+      "grad_norm": 1.0485174655914307,
       "learning_rate": 7.182674431585704e-05,
-      "loss": 1.3738,
+      "loss": 1.3705,
       "step": 140
     },
     {
       "epoch": 0.654627539503386,
-      "grad_norm": 0.6766214370727539,
+      "grad_norm": 0.8391625285148621,
       "learning_rate": 6.431137784081282e-05,
-      "loss": 1.3752,
+      "loss": 1.3695,
       "step": 145
     },
     {
       "epoch": 0.6772009029345373,
-      "grad_norm": 0.7337061762809753,
+      "grad_norm": 0.7003734707832336,
       "learning_rate": 5.702050879108284e-05,
-      "loss": 1.3607,
+      "loss": 1.3572,
       "step": 150
     },
     {
       "epoch": 0.6997742663656885,
-      "grad_norm": 0.835713267326355,
+      "grad_norm": 0.9220963716506958,
       "learning_rate": 5.000000000000002e-05,
-      "loss": 1.3486,
+      "loss": 1.3447,
       "step": 155
     },
     {
       "epoch": 0.7223476297968398,
-      "grad_norm": 0.7193433046340942,
+      "grad_norm": 0.8875526189804077,
       "learning_rate": 4.329401361372294e-05,
-      "loss": 1.3472,
+      "loss": 1.3435,
       "step": 160
     },
     {
       "epoch": 0.744920993227991,
-      "grad_norm": 0.8834360837936401,
+      "grad_norm": 1.013006329536438,
       "learning_rate": 3.694473329154778e-05,
-      "loss": 1.3397,
+      "loss": 1.3352,
       "step": 165
     },
     {
       "epoch": 0.7674943566591422,
-      "grad_norm": 0.9133020639419556,
+      "grad_norm": 0.7449764609336853,
       "learning_rate": 3.099209885178882e-05,
-      "loss": 1.3359,
+      "loss": 1.3322,
       "step": 170
     },
     {
       "epoch": 0.7900677200902935,
-      "grad_norm": 0.6195141673088074,
+      "grad_norm": 0.7195131778717041,
       "learning_rate": 2.5473555032424533e-05,
-      "loss": 1.3314,
+      "loss": 1.3275,
       "step": 175
     },
     {
       "epoch": 0.8126410835214447,
-      "grad_norm": 0.6716246604919434,
+      "grad_norm": 0.6962352395057678,
       "learning_rate": 2.042381594691678e-05,
-      "loss": 1.3376,
+      "loss": 1.3348,
       "step": 180
     },
     {
       "epoch": 0.835214446952596,
-      "grad_norm": 0.8972969055175781,
+      "grad_norm": 0.8499621152877808,
       "learning_rate": 1.587464671688187e-05,
-      "loss": 1.3334,
+      "loss": 1.3293,
       "step": 185
     },
     {
       "epoch": 0.8577878103837472,
-      "grad_norm": 0.6790732145309448,
+      "grad_norm": 0.6392541527748108,
       "learning_rate": 1.1854663655241805e-05,
-      "loss": 1.3305,
+      "loss": 1.325,
       "step": 190
     },
     {
       "epoch": 0.8803611738148984,
-      "grad_norm": 0.7163174748420715,
+      "grad_norm": 0.7239265441894531,
       "learning_rate": 8.38915425679304e-06,
-      "loss": 1.3258,
+      "loss": 1.3204,
       "step": 195
     },
     {
       "epoch": 0.9029345372460497,
-      "grad_norm": 0.7202246189117432,
+      "grad_norm": 0.7242350578308105,
       "learning_rate": 5.499918128533155e-06,
-      "loss": 1.33,
+      "loss": 1.324,
       "step": 200
     },
     {
       "epoch": 0.9255079006772009,
-      "grad_norm": 0.6578797101974487,
+      "grad_norm": 0.6468668580055237,
       "learning_rate": 3.2051298603643753e-06,
-      "loss": 1.3235,
+      "loss": 1.3198,
       "step": 205
     },
     {
       "epoch": 0.9480812641083521,
-      "grad_norm": 0.6833933591842651,
+      "grad_norm": 0.6733763813972473,
       "learning_rate": 1.5192246987791981e-06,
-      "loss": 1.327,
+      "loss": 1.3226,
       "step": 210
     },
     {
       "epoch": 0.9706546275395034,
-      "grad_norm": 1.3088898658752441,
+      "grad_norm": 1.3773088455200195,
       "learning_rate": 4.5280774269154115e-07,
-      "loss": 1.3266,
+      "loss": 1.3211,
       "step": 215
     },
     {
       "epoch": 0.9932279909706546,
-      "grad_norm": 0.6681888699531555,
+      "grad_norm": 0.689209520816803,
       "learning_rate": 1.2587232612493172e-08,
-      "loss": 1.3179,
+      "loss": 1.3142,
       "step": 220
     },
     {
       "epoch": 0.9977426636568849,
-      "eval_loss": 2.8222014904022217,
-      "eval_runtime": 1.269,
-      "eval_samples_per_second": 16.549,
-      "eval_steps_per_second": 0.788,
+      "eval_loss": 2.81227445602417,
+      "eval_runtime": 0.459,
+      "eval_samples_per_second": 45.75,
+      "eval_steps_per_second": 2.179,
       "step": 221
     },
     {
       "epoch": 0.9977426636568849,
       "step": 221,
       "total_flos": 6.738683116954583e+17,
-      "train_loss": 5.562472593730392,
-      "train_runtime": 2738.0196,
-      "train_samples_per_second": 10.349,
-      "train_steps_per_second": 0.081
+      "train_loss": 5.577704090877895,
+      "train_runtime": 575.1744,
+      "train_samples_per_second": 49.265,
+      "train_steps_per_second": 0.384
     }
   ],
   "logging_steps": 5,