Model save

README.md

@@ -2,13 +2,12 @@
 license: apache-2.0
 library_name: peft
 tags:
-- alignment-handbook
-- generated_from_trainer
 - trl
 - sft
+- generated_from_trainer
 base_model: mistralai/Mistral-7B-v0.1
 datasets:
-- robust-control/PKU-alignment-better-safer-2
+- generator
 model-index:
 - name: zephyr-7b-sft-qlora-alignment-10000
   results: []
@@ -19,9 +18,9 @@ should probably proofread and complete it, then remove this comment. -->
 
 # zephyr-7b-sft-qlora-alignment-10000
 
-This model is a fine-tuned version of [mistralai/Mistral-7B-v0.1](https://huggingface.co/mistralai/Mistral-7B-v0.1) on the robust-control/PKU-alignment-better-safer-2 dataset.
+This model is a fine-tuned version of [mistralai/Mistral-7B-v0.1](https://huggingface.co/mistralai/Mistral-7B-v0.1) on the generator dataset.
 It achieves the following results on the evaluation set:
-- Loss: 1.2067
+- Loss: 1.2123
 
 ## Model description
 
@@ -56,7 +55,7 @@ The following hyperparameters were used during training:
 
 | Training Loss | Epoch | Step | Validation Loss |
 |:-------------:|:-----:|:----:|:---------------:|
-| 1.1641        | 1.0   | 171  | 1.2067          |
+| 1.1007        | 1.0   | 274  | 1.2123          |
 
 
 ### Framework versions

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:c0a0e67a4cbe385556f2419aa87172cbd0f015851253480033dc03d20fad0e95
+oid sha256:b298414960be2dd3919b6f17ef2429c9937f9cd8739855ec9e6408fa449c49d0
 size 83946192

all_results.json

@@ -1,13 +1,13 @@
 {
     "epoch": 1.0,
-    "eval_loss": 1.2067391872406006,
-    "eval_runtime": 235.4446,
+    "eval_loss": 1.2122626304626465,
+    "eval_runtime": 236.2386,
     "eval_samples": 9907,
-    "eval_samples_per_second": 2.68,
-    "eval_steps_per_second": 0.671,
-    "train_loss": 1.2207316202029848,
-    "train_runtime": 1132.3581,
-    "train_samples": 10853,
-    "train_samples_per_second": 0.604,
-    "train_steps_per_second": 0.151
+    "eval_samples_per_second": 2.671,
+    "eval_steps_per_second": 0.669,
+    "train_loss": 1.1792847770844064,
+    "train_runtime": 1678.8973,
+    "train_samples": 17365,
+    "train_samples_per_second": 0.653,
+    "train_steps_per_second": 0.163
 }

eval_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 1.0,
-    "eval_loss": 1.2067391872406006,
-    "eval_runtime": 235.4446,
+    "eval_loss": 1.2122626304626465,
+    "eval_runtime": 236.2386,
     "eval_samples": 9907,
-    "eval_samples_per_second": 2.68,
-    "eval_steps_per_second": 0.671
+    "eval_samples_per_second": 2.671,
+    "eval_steps_per_second": 0.669
 }

runs/May11_15-37-10_compute-permanent-node-269/events.out.tfevents.1715441861.compute-permanent-node-269.196531.0

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:a11074319838279bc08e731f5169277ab772f24d1e4a43848f68b7287cfd0514
-size 11256
+oid sha256:59e36ea9b0b56d19d244b6cc429ae3148683eba2e36e66b7abb87553e0f1f8d7
+size 14079

runs/May11_15-37-10_compute-permanent-node-269/events.out.tfevents.1715443776.compute-permanent-node-269.196531.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b9601799c51fda3f03a7ba1b71d4421c05b80f0d9767301d5cb9ef25556bf76b
+size 359

train_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 1.0,
-    "train_loss": 1.2207316202029848,
-    "train_runtime": 1132.3581,
-    "train_samples": 10853,
-    "train_samples_per_second": 0.604,
-    "train_steps_per_second": 0.151
+    "train_loss": 1.1792847770844064,
+    "train_runtime": 1678.8973,
+    "train_samples": 17365,
+    "train_samples_per_second": 0.653,
+    "train_steps_per_second": 0.163
 }

trainer_state.json

@@ -1,247 +1,367 @@
 {
   "best_metric": null,
   "best_model_checkpoint": null,
-  "epoch": 1.0,
+  "epoch": 0.9981785063752276,
   "eval_steps": 500,
-  "global_step": 171,
+  "global_step": 274,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
-      "epoch": 0.01,
-      "learning_rate": 1.1111111111111112e-05,
-      "loss": 1.8643,
+      "epoch": 0.0,
+      "learning_rate": 7.142857142857143e-06,
+      "loss": 1.7833,
       "step": 1
     },
     {
-      "epoch": 0.03,
-      "learning_rate": 5.555555555555556e-05,
-      "loss": 1.7663,
+      "epoch": 0.02,
+      "learning_rate": 3.571428571428572e-05,
+      "loss": 1.7803,
       "step": 5
     },
     {
-      "epoch": 0.06,
-      "learning_rate": 0.00011111111111111112,
-      "loss": 1.4851,
+      "epoch": 0.04,
+      "learning_rate": 7.142857142857143e-05,
+      "loss": 1.5541,
       "step": 10
     },
     {
-      "epoch": 0.09,
-      "learning_rate": 0.0001666666666666667,
-      "loss": 1.3257,
+      "epoch": 0.05,
+      "learning_rate": 0.00010714285714285715,
+      "loss": 1.3651,
       "step": 15
     },
     {
-      "epoch": 0.12,
-      "learning_rate": 0.0001999156886888064,
-      "loss": 1.2837,
+      "epoch": 0.07,
+      "learning_rate": 0.00014285714285714287,
+      "loss": 1.2684,
       "step": 20
     },
     {
-      "epoch": 0.15,
-      "learning_rate": 0.00019896881839082556,
-      "loss": 1.2272,
+      "epoch": 0.09,
+      "learning_rate": 0.0001785714285714286,
+      "loss": 1.2601,
       "step": 25
     },
     {
-      "epoch": 0.18,
-      "learning_rate": 0.00019697969360350098,
-      "loss": 1.2551,
+      "epoch": 0.11,
+      "learning_rate": 0.00019996738360808565,
+      "loss": 1.2498,
       "step": 30
     },
     {
-      "epoch": 0.2,
-      "learning_rate": 0.00019396926207859084,
-      "loss": 1.2263,
+      "epoch": 0.13,
+      "learning_rate": 0.00019960069350344548,
+      "loss": 1.2452,
       "step": 35
     },
     {
-      "epoch": 0.23,
-      "learning_rate": 0.00018996922709216455,
-      "loss": 1.2222,
+      "epoch": 0.15,
+      "learning_rate": 0.00019882804237803488,
+      "loss": 1.2473,
       "step": 40
     },
     {
-      "epoch": 0.26,
-      "learning_rate": 0.00018502171357296144,
-      "loss": 1.242,
+      "epoch": 0.16,
+      "learning_rate": 0.00019765257946935944,
+      "loss": 1.1842,
       "step": 45
     },
     {
-      "epoch": 0.29,
-      "learning_rate": 0.00017917882447886582,
-      "loss": 1.214,
+      "epoch": 0.18,
+      "learning_rate": 0.00019607909582962477,
+      "loss": 1.2224,
       "step": 50
     },
     {
-      "epoch": 0.32,
-      "learning_rate": 0.00017250209209335927,
-      "loss": 1.2186,
+      "epoch": 0.2,
+      "learning_rate": 0.00019411400479795617,
+      "loss": 1.2516,
       "step": 55
     },
     {
-      "epoch": 0.35,
-      "learning_rate": 0.0001650618300204242,
-      "loss": 1.2202,
+      "epoch": 0.22,
+      "learning_rate": 0.0001917653158603628,
+      "loss": 1.2017,
       "step": 60
     },
     {
-      "epoch": 0.38,
-      "learning_rate": 0.00015693639270213136,
-      "loss": 1.1952,
+      "epoch": 0.24,
+      "learning_rate": 0.00018904260200399006,
+      "loss": 1.1943,
       "step": 65
     },
     {
-      "epoch": 0.41,
-      "learning_rate": 0.0001482113502570349,
-      "loss": 1.2162,
+      "epoch": 0.26,
+      "learning_rate": 0.00018595696069872013,
+      "loss": 1.213,
       "step": 70
     },
     {
-      "epoch": 0.44,
-      "learning_rate": 0.00013897858732926793,
-      "loss": 1.1951,
+      "epoch": 0.27,
+      "learning_rate": 0.00018252096866515558,
+      "loss": 1.1741,
       "step": 75
     },
     {
-      "epoch": 0.47,
-      "learning_rate": 0.00012933533543848461,
-      "loss": 1.1806,
+      "epoch": 0.29,
+      "learning_rate": 0.00017874863061334657,
+      "loss": 1.1809,
       "step": 80
     },
     {
-      "epoch": 0.5,
-      "learning_rate": 0.00011938314902110701,
-      "loss": 1.2158,
+      "epoch": 0.31,
+      "learning_rate": 0.00017465532216119625,
+      "loss": 1.1773,
       "step": 85
     },
     {
-      "epoch": 0.53,
-      "learning_rate": 0.00010922683594633021,
-      "loss": 1.1702,
+      "epoch": 0.33,
+      "learning_rate": 0.00017025772716520323,
+      "loss": 1.1956,
       "step": 90
     },
     {
-      "epoch": 0.56,
-      "learning_rate": 9.897335376977102e-05,
-      "loss": 1.186,
+      "epoch": 0.35,
+      "learning_rate": 0.00016557376971897266,
+      "loss": 1.1867,
       "step": 95
     },
     {
-      "epoch": 0.58,
-      "learning_rate": 8.87306833484679e-05,
-      "loss": 1.1802,
+      "epoch": 0.36,
+      "learning_rate": 0.0001606225410966638,
+      "loss": 1.182,
       "step": 100
     },
     {
-      "epoch": 0.61,
-      "learning_rate": 7.860669167935028e-05,
-      "loss": 1.1626,
+      "epoch": 0.38,
+      "learning_rate": 0.0001554242219391425,
+      "loss": 1.1625,
       "step": 105
     },
     {
-      "epoch": 0.64,
-      "learning_rate": 6.870799593678459e-05,
-      "loss": 1.177,
+      "epoch": 0.4,
+      "learning_rate": 0.00015000000000000001,
+      "loss": 1.1626,
       "step": 110
     },
     {
-      "epoch": 0.67,
-      "learning_rate": 5.913884067217685e-05,
-      "loss": 1.1495,
+      "epoch": 0.42,
+      "learning_rate": 0.00014437198378669598,
+      "loss": 1.1716,
       "step": 115
     },
     {
-      "epoch": 0.7,
-      "learning_rate": 5.000000000000002e-05,
-      "loss": 1.1842,
+      "epoch": 0.44,
+      "learning_rate": 0.0001385631124488136,
+      "loss": 1.1372,
       "step": 120
     },
     {
-      "epoch": 0.73,
-      "learning_rate": 4.1387716331478565e-05,
-      "loss": 1.1736,
+      "epoch": 0.46,
+      "learning_rate": 0.00013259706228071285,
+      "loss": 1.161,
       "step": 125
     },
     {
-      "epoch": 0.76,
-      "learning_rate": 3.339268683227499e-05,
-      "loss": 1.1638,
+      "epoch": 0.47,
+      "learning_rate": 0.0001264981502196662,
+      "loss": 1.1578,
       "step": 130
     },
     {
-      "epoch": 0.79,
-      "learning_rate": 2.6099108277934103e-05,
-      "loss": 1.1465,
+      "epoch": 0.49,
+      "learning_rate": 0.00012029123473280668,
+      "loss": 1.1553,
       "step": 135
     },
     {
-      "epoch": 0.82,
-      "learning_rate": 1.9583790365845822e-05,
-      "loss": 1.1706,
+      "epoch": 0.51,
+      "learning_rate": 0.00011400161449686293,
+      "loss": 1.1518,
       "step": 140
     },
     {
-      "epoch": 0.85,
-      "learning_rate": 1.3915346821563235e-05,
-      "loss": 1.152,
+      "epoch": 0.53,
+      "learning_rate": 0.0001076549252836496,
+      "loss": 1.15,
       "step": 145
     },
     {
-      "epoch": 0.88,
-      "learning_rate": 9.153472818047625e-06,
-      "loss": 1.1454,
+      "epoch": 0.55,
+      "learning_rate": 0.00010127703547159739,
+      "loss": 1.1362,
       "step": 150
     },
     {
-      "epoch": 0.91,
-      "learning_rate": 5.348316317440549e-06,
-      "loss": 1.1639,
+      "epoch": 0.56,
+      "learning_rate": 9.489394060920496e-05,
+      "loss": 1.1724,
       "step": 155
     },
     {
-      "epoch": 0.94,
-      "learning_rate": 2.539949955849985e-06,
-      "loss": 1.1614,
+      "epoch": 0.58,
+      "learning_rate": 8.853165746015997e-05,
+      "loss": 1.1116,
       "step": 160
     },
     {
-      "epoch": 0.96,
-      "learning_rate": 7.579490328064265e-07,
-      "loss": 1.1605,
+      "epoch": 0.6,
+      "learning_rate": 8.221611796198985e-05,
+      "loss": 1.1193,
       "step": 165
     },
     {
-      "epoch": 0.99,
-      "learning_rate": 2.108004964086474e-08,
-      "loss": 1.1641,
+      "epoch": 0.62,
+      "learning_rate": 7.597306353045393e-05,
+      "loss": 1.132,
       "step": 170
     },
+    {
+      "epoch": 0.64,
+      "learning_rate": 6.982794014048077e-05,
+      "loss": 1.1159,
+      "step": 175
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 6.380579461128819e-05,
+      "loss": 1.1067,
+      "step": 180
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 5.793117251841659e-05,
+      "loss": 1.1426,
+      "step": 185
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 5.222801814877369e-05,
+      "loss": 1.1398,
+      "step": 190
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.671957690646345e-05,
+      "loss": 1.1299,
+      "step": 195
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 4.142830056718052e-05,
+      "loss": 1.1109,
+      "step": 200
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 3.637575576734404e-05,
+      "loss": 1.0975,
+      "step": 205
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 3.158253610095697e-05,
+      "loss": 1.1381,
+      "step": 210
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 2.706817818247551e-05,
+      "loss": 1.1048,
+      "step": 215
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 2.2851082017805703e-05,
+      "loss": 1.145,
+      "step": 220
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.8948436007986546e-05,
+      "loss": 1.104,
+      "step": 225
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.5376146891235598e-05,
+      "loss": 1.0886,
+      "step": 230
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.214877490890578e-05,
+      "loss": 1.1076,
+      "step": 235
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 9.279474459608805e-06,
+      "loss": 1.127,
+      "step": 240
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 6.779940483393032e-06,
+      "loss": 1.1085,
+      "step": 245
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.660360794506946e-06,
+      "loss": 1.1209,
+      "step": 250
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 2.929374557035036e-06,
+      "loss": 1.0841,
+      "step": 255
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.5940370726542863e-06,
+      "loss": 1.0937,
+      "step": 260
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 6.597910240324967e-07,
+      "loss": 1.1317,
+      "step": 265
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 1.3044429107700318e-07,
+      "loss": 1.1007,
+      "step": 270
+    },
     {
       "epoch": 1.0,
-      "eval_loss": 1.2067391872406006,
-      "eval_runtime": 235.4134,
-      "eval_samples_per_second": 2.68,
-      "eval_steps_per_second": 0.671,
-      "step": 171
+      "eval_loss": 1.2122626304626465,
+      "eval_runtime": 236.2559,
+      "eval_samples_per_second": 2.671,
+      "eval_steps_per_second": 0.669,
+      "step": 274
     },
     {
       "epoch": 1.0,
-      "step": 171,
-      "total_flos": 6.011757435184742e+16,
-      "train_loss": 1.2207316202029848,
-      "train_runtime": 1132.3581,
-      "train_samples_per_second": 0.604,
-      "train_steps_per_second": 0.151
+      "step": 274,
+      "total_flos": 9.632874487611392e+16,
+      "train_loss": 1.1792847770844064,
+      "train_runtime": 1678.8973,
+      "train_samples_per_second": 0.653,
+      "train_steps_per_second": 0.163
     }
   ],
   "logging_steps": 5,
-  "max_steps": 171,
+  "max_steps": 274,
   "num_input_tokens_seen": 0,
   "num_train_epochs": 1,
   "save_steps": 100,
-  "total_flos": 6.011757435184742e+16,
+  "total_flos": 9.632874487611392e+16,
   "train_batch_size": 2,
   "trial_name": null,
   "trial_params": null