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: 1.6581
+- Loss: 1.6532
 
 ## Model description
 
@@ -57,7 +57,7 @@ The following hyperparameters were used during training:
 
 | Training Loss | Epoch | Step | Validation Loss |
 |:-------------:|:-----:|:----:|:---------------:|
-| 1.6561        | 1.0   | 140  | 1.6581          |
+| 1.6467        | 1.0   | 140  | 1.6532          |
 
 
 ### Framework versions

all_results.json

@@ -1,14 +1,9 @@
 {
     "epoch": 1.0,
-    "eval_loss": 1.6581236124038696,
-    "eval_runtime": 3.846,
-    "eval_samples": 518,
-    "eval_samples_per_second": 46.282,
-    "eval_steps_per_second": 0.78,
     "total_flos": 4.268849030789857e+17,
-    "train_loss": 5.955788305827549,
-    "train_runtime": 1733.7968,
+    "train_loss": 5.926357351030622,
+    "train_runtime": 363.5109,
     "train_samples": 51241,
-    "train_samples_per_second": 10.326,
-    "train_steps_per_second": 0.081
+    "train_samples_per_second": 49.253,
+    "train_steps_per_second": 0.385
 }

runs/Nov17_22-39-49_main-lora-gemma7b-closedqa-0-0/events.out.tfevents.1731901785.main-lora-gemma7b-closedqa-0-0.457.0

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:d09fcf6e9494fffe057c363976c86a106756c39545e2ddfc8819d9251fb7eae3
-size 11716
+oid sha256:8dd40eb05c9b5e3735b26cf56af1cd7ecbaa13529310ad40b014624d5b17d151
+size 12341

train_results.json

@@ -1,9 +1,9 @@
 {
     "epoch": 1.0,
     "total_flos": 4.268849030789857e+17,
-    "train_loss": 5.955788305827549,
-    "train_runtime": 1733.7968,
+    "train_loss": 5.926357351030622,
+    "train_runtime": 363.5109,
     "train_samples": 51241,
-    "train_samples_per_second": 10.326,
-    "train_steps_per_second": 0.081
+    "train_samples_per_second": 49.253,
+    "train_steps_per_second": 0.385
 }

trainer_state.json

@@ -10,223 +10,223 @@
   "log_history": [
     {
       "epoch": 0.007142857142857143,
-      "grad_norm": 106.10701751708984,
+      "grad_norm": 104.33383178710938,
       "learning_rate": 1.4285714285714285e-05,
-      "loss": 27.4831,
+      "loss": 26.8049,
       "step": 1
     },
     {
       "epoch": 0.03571428571428571,
-      "grad_norm": 46.149723052978516,
+      "grad_norm": 53.856834411621094,
       "learning_rate": 7.142857142857143e-05,
-      "loss": 26.2992,
+      "loss": 26.2527,
       "step": 5
     },
     {
       "epoch": 0.07142857142857142,
-      "grad_norm": 16.458189010620117,
+      "grad_norm": 16.46755599975586,
       "learning_rate": 0.00014285714285714287,
-      "loss": 20.4929,
+      "loss": 20.5587,
       "step": 10
     },
     {
       "epoch": 0.10714285714285714,
-      "grad_norm": 9.764805793762207,
+      "grad_norm": 9.823058128356934,
       "learning_rate": 0.00019996891820008164,
-      "loss": 16.0513,
+      "loss": 16.1016,
       "step": 15
     },
     {
       "epoch": 0.14285714285714285,
-      "grad_norm": 3.3296561241149902,
+      "grad_norm": 3.4154016971588135,
       "learning_rate": 0.00019888308262251285,
-      "loss": 13.4526,
+      "loss": 13.4739,
       "step": 20
     },
     {
       "epoch": 0.17857142857142858,
-      "grad_norm": 3.235478162765503,
+      "grad_norm": 3.293179988861084,
       "learning_rate": 0.0001962624246950012,
-      "loss": 12.6172,
+      "loss": 12.6248,
       "step": 25
     },
     {
       "epoch": 0.21428571428571427,
-      "grad_norm": 5.432589530944824,
+      "grad_norm": 5.406515121459961,
       "learning_rate": 0.00019214762118704076,
-      "loss": 11.9808,
+      "loss": 11.9858,
       "step": 30
     },
     {
       "epoch": 0.25,
-      "grad_norm": 10.644349098205566,
+      "grad_norm": 10.7940034866333,
       "learning_rate": 0.00018660254037844388,
-      "loss": 11.0065,
+      "loss": 11.0004,
       "step": 35
     },
     {
       "epoch": 0.2857142857142857,
-      "grad_norm": 14.282907485961914,
+      "grad_norm": 14.75281047821045,
       "learning_rate": 0.00017971325072229226,
-      "loss": 9.3131,
+      "loss": 9.2523,
       "step": 40
     },
     {
       "epoch": 0.32142857142857145,
-      "grad_norm": 21.60712432861328,
+      "grad_norm": 22.644073486328125,
       "learning_rate": 0.00017158668492597186,
-      "loss": 7.1723,
+      "loss": 7.0092,
       "step": 45
     },
     {
       "epoch": 0.35714285714285715,
-      "grad_norm": 15.418853759765625,
+      "grad_norm": 14.965047836303711,
       "learning_rate": 0.00016234898018587337,
-      "loss": 4.6599,
+      "loss": 4.4455,
       "step": 50
     },
     {
       "epoch": 0.39285714285714285,
-      "grad_norm": 5.335650444030762,
+      "grad_norm": 5.274498462677002,
       "learning_rate": 0.0001521435203379498,
-      "loss": 2.855,
+      "loss": 2.7643,
       "step": 55
     },
     {
       "epoch": 0.42857142857142855,
-      "grad_norm": 4.913125038146973,
+      "grad_norm": 4.607834339141846,
       "learning_rate": 0.00014112871031306119,
-      "loss": 2.3938,
+      "loss": 2.3451,
       "step": 60
     },
     {
       "epoch": 0.4642857142857143,
-      "grad_norm": 3.951244592666626,
+      "grad_norm": 3.502431631088257,
       "learning_rate": 0.00012947551744109043,
-      "loss": 2.1646,
+      "loss": 2.1217,
       "step": 65
     },
     {
       "epoch": 0.5,
-      "grad_norm": 1.6933783292770386,
+      "grad_norm": 1.5408211946487427,
       "learning_rate": 0.00011736481776669306,
-      "loss": 2.0061,
+      "loss": 1.979,
       "step": 70
     },
     {
       "epoch": 0.5357142857142857,
-      "grad_norm": 0.9344208836555481,
+      "grad_norm": 1.2710477113723755,
       "learning_rate": 0.00010498458856606972,
-      "loss": 1.9092,
+      "loss": 1.8913,
       "step": 75
     },
     {
       "epoch": 0.5714285714285714,
-      "grad_norm": 0.9437263011932373,
+      "grad_norm": 1.2541048526763916,
       "learning_rate": 9.252699064135758e-05,
-      "loss": 1.8086,
+      "loss": 1.7956,
       "step": 80
     },
     {
       "epoch": 0.6071428571428571,
-      "grad_norm": 1.5205684900283813,
+      "grad_norm": 1.3496887683868408,
       "learning_rate": 8.018538568006027e-05,
-      "loss": 1.7838,
+      "loss": 1.7699,
       "step": 85
     },
     {
       "epoch": 0.6428571428571429,
-      "grad_norm": 0.8009554743766785,
+      "grad_norm": 1.127426028251648,
       "learning_rate": 6.815133497483157e-05,
-      "loss": 1.7433,
+      "loss": 1.7305,
       "step": 90
     },
     {
       "epoch": 0.6785714285714286,
-      "grad_norm": 0.9393540620803833,
+      "grad_norm": 1.5040632486343384,
       "learning_rate": 5.6611626088244194e-05,
-      "loss": 1.736,
+      "loss": 1.728,
       "step": 95
     },
     {
       "epoch": 0.7142857142857143,
-      "grad_norm": 1.6479978561401367,
+      "grad_norm": 1.3170948028564453,
       "learning_rate": 4.574537361342407e-05,
-      "loss": 1.6976,
+      "loss": 1.6858,
       "step": 100
     },
     {
       "epoch": 0.75,
-      "grad_norm": 0.7829992771148682,
+      "grad_norm": 0.6884583830833435,
       "learning_rate": 3.5721239031346066e-05,
-      "loss": 1.6793,
+      "loss": 1.6739,
       "step": 105
     },
     {
       "epoch": 0.7857142857142857,
-      "grad_norm": 0.7236832976341248,
+      "grad_norm": 0.6821992993354797,
       "learning_rate": 2.669481281701739e-05,
-      "loss": 1.7,
+      "loss": 1.6896,
       "step": 110
     },
     {
       "epoch": 0.8214285714285714,
-      "grad_norm": 0.7881686687469482,
+      "grad_norm": 0.8719048500061035,
       "learning_rate": 1.880619942841435e-05,
-      "loss": 1.6778,
+      "loss": 1.6684,
       "step": 115
     },
     {
       "epoch": 0.8571428571428571,
-      "grad_norm": 1.7755261659622192,
+      "grad_norm": 1.8157048225402832,
       "learning_rate": 1.2177842662977135e-05,
-      "loss": 1.6757,
+      "loss": 1.6667,
       "step": 120
     },
     {
       "epoch": 0.8928571428571429,
-      "grad_norm": 1.0512515306472778,
+      "grad_norm": 1.1482408046722412,
       "learning_rate": 6.9126251355795864e-06,
-      "loss": 1.6766,
+      "loss": 1.6691,
       "step": 125
     },
     {
       "epoch": 0.9285714285714286,
-      "grad_norm": 0.7070327997207642,
+      "grad_norm": 0.786839485168457,
       "learning_rate": 3.092271377092215e-06,
-      "loss": 1.67,
+      "loss": 1.6592,
       "step": 130
     },
     {
       "epoch": 0.9642857142857143,
-      "grad_norm": 0.7091158628463745,
+      "grad_norm": 0.6941242814064026,
       "learning_rate": 7.760793399827937e-07,
-      "loss": 1.6459,
+      "loss": 1.6378,
       "step": 135
     },
     {
       "epoch": 1.0,
-      "grad_norm": 0.7374704480171204,
+      "grad_norm": 0.7584343552589417,
       "learning_rate": 0.0,
-      "loss": 1.6561,
+      "loss": 1.6467,
       "step": 140
     },
     {
       "epoch": 1.0,
-      "eval_loss": 1.6581236124038696,
-      "eval_runtime": 3.8242,
-      "eval_samples_per_second": 46.545,
-      "eval_steps_per_second": 0.784,
+      "eval_loss": 1.6531758308410645,
+      "eval_runtime": 1.3594,
+      "eval_samples_per_second": 130.936,
+      "eval_steps_per_second": 2.207,
       "step": 140
     },
     {
       "epoch": 1.0,
       "step": 140,
       "total_flos": 4.268849030789857e+17,
-      "train_loss": 5.955788305827549,
-      "train_runtime": 1733.7968,
-      "train_samples_per_second": 10.326,
-      "train_steps_per_second": 0.081
+      "train_loss": 5.926357351030622,
+      "train_runtime": 363.5109,
+      "train_samples_per_second": 49.253,
+      "train_steps_per_second": 0.385
     }
   ],
   "logging_steps": 5,