Model save

README.md

@@ -1,27 +1,27 @@
 ---
 base_model: meta-llama/Meta-Llama-3-8B
 datasets:
-- llama-duo/synth_summarize_dataset_dedup
+- generator
 library_name: peft
 license: llama3
 tags:
-- alignment-handbook
 - trl
 - sft
+- alignment-handbook
 - generated_from_trainer
 model-index:
-- name: llama3.1-8b-gpt4o_100k_summarize-kasa
+- name: llama3.1-8b-gpt4o_100k_summarize-k
   results: []
 ---
 
 <!-- This model card has been generated automatically according to the information the Trainer had access to. You
 should probably proofread and complete it, then remove this comment. -->
 
-# llama3.1-8b-gpt4o_100k_summarize-kasa
+# llama3.1-8b-gpt4o_100k_summarize-k
 
-This model is a fine-tuned version of [meta-llama/Meta-Llama-3-8B](https://huggingface.co/meta-llama/Meta-Llama-3-8B) on the llama-duo/synth_summarize_dataset_dedup dataset.
+This model is a fine-tuned version of [meta-llama/Meta-Llama-3-8B](https://huggingface.co/meta-llama/Meta-Llama-3-8B) on the generator dataset.
 It achieves the following results on the evaluation set:
-- Loss: 2.1303
+- Loss: 2.0198
 
 ## Model description
 
@@ -58,7 +58,7 @@ The following hyperparameters were used during training:
 
 | Training Loss | Epoch | Step | Validation Loss |
 |:-------------:|:-----:|:----:|:---------------:|
-| 0.9593        | 1.0   | 195  | 2.1303          |
+| 0.8483        | 1.0   | 195  | 2.0198          |
 
 
 ### Framework versions

all_results.json

@@ -6,9 +6,9 @@
     "eval_samples_per_second": 8.549,
     "eval_steps_per_second": 0.855,
     "total_flos": 5.7571825768464384e+17,
-    "train_loss": 1.1034836133321126,
-    "train_runtime": 2146.0273,
+    "train_loss": 0.9919692577459873,
+    "train_runtime": 2099.4843,
     "train_samples": 115376,
-    "train_samples_per_second": 5.814,
-    "train_steps_per_second": 0.091
+    "train_samples_per_second": 5.943,
+    "train_steps_per_second": 0.093
 }

train_results.json

@@ -1,9 +1,9 @@
 {
     "epoch": 1.0,
     "total_flos": 5.7571825768464384e+17,
-    "train_loss": 1.1034836133321126,
-    "train_runtime": 2146.0273,
+    "train_loss": 0.9919692577459873,
+    "train_runtime": 2099.4843,
     "train_samples": 115376,
-    "train_samples_per_second": 5.814,
-    "train_steps_per_second": 0.091
+    "train_samples_per_second": 5.943,
+    "train_steps_per_second": 0.093
 }

trainer_state.json

@@ -10,300 +10,300 @@
   "log_history": [
     {
       "epoch": 0.005128205128205128,
-      "grad_norm": 0.7783320546150208,
+      "grad_norm": 0.7784392237663269,
       "learning_rate": 1.4999999999999999e-05,
-      "loss": 2.3664,
+      "loss": 2.2555,
       "step": 1
     },
     {
       "epoch": 0.02564102564102564,
-      "grad_norm": 0.969870388507843,
+      "grad_norm": 0.9956199526786804,
       "learning_rate": 7.5e-05,
-      "loss": 2.3576,
+      "loss": 2.2474,
       "step": 5
     },
     {
       "epoch": 0.05128205128205128,
-      "grad_norm": 1.2408868074417114,
+      "grad_norm": 1.2538105249404907,
       "learning_rate": 0.00015,
-      "loss": 2.2471,
+      "loss": 2.1364,
       "step": 10
     },
     {
       "epoch": 0.07692307692307693,
-      "grad_norm": 1.4603853225708008,
+      "grad_norm": 1.435035228729248,
       "learning_rate": 0.000225,
-      "loss": 1.8023,
+      "loss": 1.6881,
       "step": 15
     },
     {
       "epoch": 0.10256410256410256,
-      "grad_norm": 2.0677075386047363,
+      "grad_norm": 1.9407316446304321,
       "learning_rate": 0.0003,
-      "loss": 1.3707,
+      "loss": 1.2552,
       "step": 20
     },
     {
       "epoch": 0.1282051282051282,
-      "grad_norm": 0.6989911794662476,
+      "grad_norm": 0.6618439555168152,
       "learning_rate": 0.00029939614409928584,
-      "loss": 1.2607,
+      "loss": 1.1476,
       "step": 25
     },
     {
       "epoch": 0.15384615384615385,
-      "grad_norm": 0.4506840705871582,
+      "grad_norm": 0.4018121063709259,
       "learning_rate": 0.00029758943828979444,
-      "loss": 1.1864,
+      "loss": 1.0741,
       "step": 30
     },
     {
       "epoch": 0.1794871794871795,
-      "grad_norm": 0.322531133890152,
+      "grad_norm": 0.3278277814388275,
       "learning_rate": 0.00029459442910437797,
-      "loss": 1.1286,
+      "loss": 1.0159,
       "step": 35
     },
     {
       "epoch": 0.20512820512820512,
-      "grad_norm": 0.2993707060813904,
+      "grad_norm": 0.2918931245803833,
       "learning_rate": 0.00029043523059596053,
-      "loss": 1.0941,
+      "loss": 0.9818,
       "step": 40
     },
     {
       "epoch": 0.23076923076923078,
-      "grad_norm": 0.220694437623024,
+      "grad_norm": 0.21536950767040253,
       "learning_rate": 0.0002851453301853628,
-      "loss": 1.0717,
+      "loss": 0.96,
       "step": 45
     },
     {
       "epoch": 0.2564102564102564,
-      "grad_norm": 0.2211257368326187,
+      "grad_norm": 0.187772735953331,
       "learning_rate": 0.0002787673190402799,
-      "loss": 1.0575,
+      "loss": 0.9458,
       "step": 50
     },
     {
       "epoch": 0.28205128205128205,
-      "grad_norm": 0.2125224620103836,
+      "grad_norm": 0.2083010971546173,
       "learning_rate": 0.0002713525491562421,
-      "loss": 1.0366,
+      "loss": 0.9253,
       "step": 55
     },
     {
       "epoch": 0.3076923076923077,
-      "grad_norm": 0.22398926317691803,
+      "grad_norm": 0.2108624428510666,
       "learning_rate": 0.00026296071990054165,
-      "loss": 1.0334,
+      "loss": 0.9218,
       "step": 60
     },
     {
       "epoch": 0.3333333333333333,
-      "grad_norm": 0.1814454346895218,
+      "grad_norm": 0.18083609640598297,
       "learning_rate": 0.0002536593973480297,
-      "loss": 1.0198,
+      "loss": 0.9083,
       "step": 65
     },
     {
       "epoch": 0.358974358974359,
-      "grad_norm": 0.18179959058761597,
+      "grad_norm": 0.178299680352211,
       "learning_rate": 0.00024352347027881003,
-      "loss": 1.0152,
+      "loss": 0.9036,
       "step": 70
     },
     {
       "epoch": 0.38461538461538464,
-      "grad_norm": 0.18999409675598145,
+      "grad_norm": 0.19863168895244598,
       "learning_rate": 0.00023263454721781537,
-      "loss": 1.0149,
+      "loss": 0.9033,
       "step": 75
     },
     {
       "epoch": 0.41025641025641024,
-      "grad_norm": 0.21783891320228577,
+      "grad_norm": 0.22889919579029083,
       "learning_rate": 0.0002210802993709498,
-      "loss": 1.0023,
+      "loss": 0.8912,
       "step": 80
     },
     {
       "epoch": 0.4358974358974359,
-      "grad_norm": 0.19977697730064392,
+      "grad_norm": 0.20287582278251648,
       "learning_rate": 0.00020895375474808852,
-      "loss": 1.0016,
+      "loss": 0.8899,
       "step": 85
     },
     {
       "epoch": 0.46153846153846156,
-      "grad_norm": 0.21022814512252808,
+      "grad_norm": 0.21162137389183044,
       "learning_rate": 0.0001963525491562421,
-      "loss": 0.9924,
+      "loss": 0.8813,
       "step": 90
     },
     {
       "epoch": 0.48717948717948717,
-      "grad_norm": 0.21291780471801758,
+      "grad_norm": 0.21930070221424103,
       "learning_rate": 0.00018337814009344714,
-      "loss": 0.9906,
+      "loss": 0.8795,
       "step": 95
     },
     {
       "epoch": 0.5128205128205128,
-      "grad_norm": 0.2277144193649292,
+      "grad_norm": 0.23387882113456726,
       "learning_rate": 0.00017013498987264832,
-      "loss": 0.9901,
+      "loss": 0.8794,
       "step": 100
     },
     {
       "epoch": 0.5384615384615384,
-      "grad_norm": 0.20027609169483185,
+      "grad_norm": 0.19868768751621246,
       "learning_rate": 0.00015672972455257723,
-      "loss": 0.9862,
+      "loss": 0.8752,
       "step": 105
     },
     {
       "epoch": 0.5641025641025641,
-      "grad_norm": 0.22872835397720337,
+      "grad_norm": 0.21123158931732178,
       "learning_rate": 0.0001432702754474228,
-      "loss": 0.9639,
+      "loss": 0.8524,
       "step": 110
     },
     {
       "epoch": 0.5897435897435898,
-      "grad_norm": 0.2403298318386078,
+      "grad_norm": 0.25728657841682434,
       "learning_rate": 0.0001298650101273517,
-      "loss": 0.9843,
+      "loss": 0.8734,
       "step": 115
     },
     {
       "epoch": 0.6153846153846154,
-      "grad_norm": 0.21371488273143768,
+      "grad_norm": 0.21154211461544037,
       "learning_rate": 0.00011662185990655284,
-      "loss": 0.9712,
+      "loss": 0.8597,
       "step": 120
     },
     {
       "epoch": 0.6410256410256411,
-      "grad_norm": 0.21134255826473236,
+      "grad_norm": 0.21001510322093964,
       "learning_rate": 0.0001036474508437579,
-      "loss": 0.9712,
+      "loss": 0.8606,
       "step": 125
     },
     {
       "epoch": 0.6666666666666666,
-      "grad_norm": 0.21926414966583252,
+      "grad_norm": 0.2222093790769577,
       "learning_rate": 9.104624525191145e-05,
-      "loss": 0.9669,
+      "loss": 0.8563,
       "step": 130
     },
     {
       "epoch": 0.6923076923076923,
-      "grad_norm": 0.19241148233413696,
+      "grad_norm": 0.19596728682518005,
       "learning_rate": 7.89197006290502e-05,
-      "loss": 0.966,
+      "loss": 0.855,
       "step": 135
     },
     {
       "epoch": 0.717948717948718,
-      "grad_norm": 0.20280851423740387,
+      "grad_norm": 0.20535971224308014,
       "learning_rate": 6.736545278218463e-05,
-      "loss": 0.9702,
+      "loss": 0.8592,
       "step": 140
     },
     {
       "epoch": 0.7435897435897436,
-      "grad_norm": 0.1985516995191574,
+      "grad_norm": 0.20298361778259277,
       "learning_rate": 5.6476529721189974e-05,
-      "loss": 0.9658,
+      "loss": 0.8549,
       "step": 145
     },
     {
       "epoch": 0.7692307692307693,
-      "grad_norm": 0.19389475882053375,
+      "grad_norm": 0.19454172253608704,
       "learning_rate": 4.63406026519703e-05,
-      "loss": 0.9687,
+      "loss": 0.8578,
       "step": 150
     },
     {
       "epoch": 0.7948717948717948,
-      "grad_norm": 0.1933661550283432,
+      "grad_norm": 0.19015519320964813,
       "learning_rate": 3.7039280099458366e-05,
-      "loss": 0.9695,
+      "loss": 0.8579,
       "step": 155
     },
     {
       "epoch": 0.8205128205128205,
-      "grad_norm": 0.20416520535945892,
+      "grad_norm": 0.20558574795722961,
       "learning_rate": 2.8647450843757897e-05,
-      "loss": 0.9623,
+      "loss": 0.8508,
       "step": 160
     },
     {
       "epoch": 0.8461538461538461,
-      "grad_norm": 0.18719200789928436,
+      "grad_norm": 0.19159014523029327,
       "learning_rate": 2.1232680959720082e-05,
-      "loss": 0.9527,
+      "loss": 0.8415,
       "step": 165
     },
     {
       "epoch": 0.8717948717948718,
-      "grad_norm": 0.18083591759204865,
+      "grad_norm": 0.17908993363380432,
       "learning_rate": 1.4854669814637143e-05,
-      "loss": 0.9624,
+      "loss": 0.8514,
       "step": 170
     },
     {
       "epoch": 0.8974358974358975,
-      "grad_norm": 0.1832273304462433,
+      "grad_norm": 0.18227402865886688,
       "learning_rate": 9.564769404039419e-06,
-      "loss": 0.9616,
+      "loss": 0.8503,
       "step": 175
     },
     {
       "epoch": 0.9230769230769231,
-      "grad_norm": 0.2028147280216217,
+      "grad_norm": 0.2028588354587555,
       "learning_rate": 5.405570895622013e-06,
-      "loss": 0.9562,
+      "loss": 0.8449,
       "step": 180
     },
     {
       "epoch": 0.9487179487179487,
-      "grad_norm": 0.19322623312473297,
+      "grad_norm": 0.19046302139759064,
       "learning_rate": 2.4105617102055496e-06,
-      "loss": 0.957,
+      "loss": 0.8456,
       "step": 185
     },
     {
       "epoch": 0.9743589743589743,
-      "grad_norm": 0.1807137131690979,
+      "grad_norm": 0.18208374083042145,
       "learning_rate": 6.038559007141397e-07,
-      "loss": 0.9652,
+      "loss": 0.854,
       "step": 190
     },
     {
       "epoch": 1.0,
-      "grad_norm": 0.2156570702791214,
+      "grad_norm": 0.21627038717269897,
       "learning_rate": 0.0,
-      "loss": 0.9593,
+      "loss": 0.8483,
       "step": 195
     },
     {
       "epoch": 1.0,
-      "eval_loss": 2.1302647590637207,
-      "eval_runtime": 1.1576,
-      "eval_samples_per_second": 8.639,
-      "eval_steps_per_second": 0.864,
+      "eval_loss": 2.0198426246643066,
+      "eval_runtime": 1.151,
+      "eval_samples_per_second": 8.688,
+      "eval_steps_per_second": 0.869,
       "step": 195
     },
     {
       "epoch": 1.0,
       "step": 195,
       "total_flos": 5.7571825768464384e+17,
-      "train_loss": 1.1034836133321126,
-      "train_runtime": 2146.0273,
-      "train_samples_per_second": 5.814,
-      "train_steps_per_second": 0.091
+      "train_loss": 0.9919692577459873,
+      "train_runtime": 2099.4843,
+      "train_samples_per_second": 5.943,
+      "train_steps_per_second": 0.093
     }
   ],
   "logging_steps": 5,