Training in progress, epoch 1

all_results.json

@@ -1,26 +1,26 @@
 {
-    "epoch": 20.0,
-    "eval_gen_len": 27.2356,
-    "eval_loss": 2.2613940238952637,
-    "eval_rouge-1": 72.7529,
-    "eval_rouge-2": 64.3941,
-    "eval_rouge-l": 71.5976,
-    "eval_runtime": 129.9628,
+    "epoch": 8.0,
+    "eval_gen_len": 19.3489,
+    "eval_loss": 2.904005527496338,
+    "eval_rouge-1": 39.3585,
+    "eval_rouge-2": 22.1712,
+    "eval_rouge-l": 38.6987,
+    "eval_runtime": 102.7662,
     "eval_samples": 450,
-    "eval_samples_per_second": 3.463,
-    "eval_steps_per_second": 0.439,
-    "predict_gen_len": 26.4604,
-    "predict_loss": 2.305396795272827,
-    "predict_rouge-1": 73.3028,
-    "predict_rouge-2": 65.6876,
-    "predict_rouge-l": 72.5101,
-    "predict_runtime": 115.003,
+    "eval_samples_per_second": 4.379,
+    "eval_steps_per_second": 0.555,
+    "predict_gen_len": 19.5492,
+    "predict_loss": 2.9291486740112305,
+    "predict_rouge-1": 38.0677,
+    "predict_rouge-2": 21.3309,
+    "predict_rouge-l": 37.1069,
+    "predict_runtime": 98.6844,
     "predict_samples": 417,
-    "predict_samples_per_second": 3.626,
-    "predict_steps_per_second": 0.461,
-    "train_loss": 2.047671499282029,
-    "train_runtime": 9522.6649,
+    "predict_samples_per_second": 4.226,
+    "predict_steps_per_second": 0.537,
+    "train_loss": 2.820291715247609,
+    "train_runtime": 3245.0074,
     "train_samples": 5133,
-    "train_samples_per_second": 26.951,
-    "train_steps_per_second": 3.371
+    "train_samples_per_second": 79.091,
+    "train_steps_per_second": 9.892
 }

eval_results.json

@@ -1,12 +1,12 @@
 {
-    "epoch": 20.0,
-    "eval_gen_len": 27.2356,
-    "eval_loss": 2.2613940238952637,
-    "eval_rouge-1": 72.7529,
-    "eval_rouge-2": 64.3941,
-    "eval_rouge-l": 71.5976,
-    "eval_runtime": 129.9628,
+    "epoch": 8.0,
+    "eval_gen_len": 19.3489,
+    "eval_loss": 2.904005527496338,
+    "eval_rouge-1": 39.3585,
+    "eval_rouge-2": 22.1712,
+    "eval_rouge-l": 38.6987,
+    "eval_runtime": 102.7662,
     "eval_samples": 450,
-    "eval_samples_per_second": 3.463,
-    "eval_steps_per_second": 0.439
+    "eval_samples_per_second": 4.379,
+    "eval_steps_per_second": 0.555
 }

predict_results.json

@@ -1,11 +1,11 @@
 {
-    "predict_gen_len": 26.4604,
-    "predict_loss": 2.305396795272827,
-    "predict_rouge-1": 73.3028,
-    "predict_rouge-2": 65.6876,
-    "predict_rouge-l": 72.5101,
-    "predict_runtime": 115.003,
+    "predict_gen_len": 19.5492,
+    "predict_loss": 2.9291486740112305,
+    "predict_rouge-1": 38.0677,
+    "predict_rouge-2": 21.3309,
+    "predict_rouge-l": 37.1069,
+    "predict_runtime": 98.6844,
     "predict_samples": 417,
-    "predict_samples_per_second": 3.626,
-    "predict_steps_per_second": 0.461
+    "predict_samples_per_second": 4.226,
+    "predict_steps_per_second": 0.537
 }

train_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 20.0,
-    "train_loss": 2.047671499282029,
-    "train_runtime": 9522.6649,
+    "epoch": 8.0,
+    "train_loss": 2.820291715247609,
+    "train_runtime": 3245.0074,
     "train_samples": 5133,
-    "train_samples_per_second": 26.951,
-    "train_steps_per_second": 3.371
+    "train_samples_per_second": 79.091,
+    "train_steps_per_second": 9.892
 }

trainer_state.json

@@ -1,280 +1,130 @@
 {
-  "best_metric": 2.2613940238952637,
-  "best_model_checkpoint": "PEFT/adapters-lib/output/mam/dataset-5100/test/checkpoint-9630",
-  "epoch": 20.0,
+  "best_metric": 2.904005527496338,
+  "best_model_checkpoint": "PEFT/adapters-lib/output/mam/dataset-5100/test/checkpoint-1926",
+  "epoch": 8.0,
   "eval_steps": 500,
-  "global_step": 12840,
+  "global_step": 5136,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
       "epoch": 1.0,
-      "eval_gen_len": 27.1578,
-      "eval_loss": 3.0984482765197754,
-      "eval_rouge-1": 35.9221,
-      "eval_rouge-2": 20.801,
-      "eval_rouge-l": 33.0593,
-      "eval_runtime": 132.8327,
-      "eval_samples_per_second": 3.388,
-      "eval_steps_per_second": 0.429,
+      "eval_gen_len": 26.2222,
+      "eval_loss": 3.184800863265991,
+      "eval_rouge-1": 33.9748,
+      "eval_rouge-2": 18.3088,
+      "eval_rouge-l": 33.4231,
+      "eval_runtime": 130.5235,
+      "eval_samples_per_second": 3.448,
+      "eval_steps_per_second": 0.437,
       "step": 642
     },
     {
       "epoch": 2.0,
-      "eval_gen_len": 23.6511,
-      "eval_loss": 2.8489246368408203,
-      "eval_rouge-1": 39.2966,
-      "eval_rouge-2": 22.654,
-      "eval_rouge-l": 37.1498,
-      "eval_runtime": 117.4318,
-      "eval_samples_per_second": 3.832,
-      "eval_steps_per_second": 0.485,
+      "eval_gen_len": 20.3911,
+      "eval_loss": 3.0025367736816406,
+      "eval_rouge-1": 37.1532,
+      "eval_rouge-2": 19.9004,
+      "eval_rouge-l": 36.5443,
+      "eval_runtime": 108.9318,
+      "eval_samples_per_second": 4.131,
+      "eval_steps_per_second": 0.523,
       "step": 1284
     },
     {
       "epoch": 3.0,
-      "eval_gen_len": 26.7267,
-      "eval_loss": 2.5933380126953125,
-      "eval_rouge-1": 47.1185,
-      "eval_rouge-2": 30.2869,
-      "eval_rouge-l": 44.1491,
-      "eval_runtime": 135.6918,
-      "eval_samples_per_second": 3.316,
-      "eval_steps_per_second": 0.42,
+      "eval_gen_len": 19.3489,
+      "eval_loss": 2.904005527496338,
+      "eval_rouge-1": 39.3585,
+      "eval_rouge-2": 22.1712,
+      "eval_rouge-l": 38.6987,
+      "eval_runtime": 101.7517,
+      "eval_samples_per_second": 4.423,
+      "eval_steps_per_second": 0.56,
       "step": 1926
     },
     {
       "epoch": 4.0,
-      "eval_gen_len": 25.22,
-      "eval_loss": 2.476445198059082,
-      "eval_rouge-1": 50.0244,
-      "eval_rouge-2": 34.9212,
-      "eval_rouge-l": 48.057,
-      "eval_runtime": 123.2916,
-      "eval_samples_per_second": 3.65,
-      "eval_steps_per_second": 0.462,
+      "eval_gen_len": 19.2956,
+      "eval_loss": 2.9436676502227783,
+      "eval_rouge-1": 39.0962,
+      "eval_rouge-2": 21.9264,
+      "eval_rouge-l": 38.3542,
+      "eval_runtime": 102.5418,
+      "eval_samples_per_second": 4.388,
+      "eval_steps_per_second": 0.556,
       "step": 2568
     },
     {
       "epoch": 5.0,
-      "eval_gen_len": 26.6533,
-      "eval_loss": 2.412160873413086,
-      "eval_rouge-1": 54.5479,
-      "eval_rouge-2": 40.7397,
-      "eval_rouge-l": 52.5031,
-      "eval_runtime": 131.3008,
-      "eval_samples_per_second": 3.427,
-      "eval_steps_per_second": 0.434,
+      "eval_gen_len": 17.7667,
+      "eval_loss": 2.945244312286377,
+      "eval_rouge-1": 38.377,
+      "eval_rouge-2": 21.3498,
+      "eval_rouge-l": 37.5445,
+      "eval_runtime": 96.4996,
+      "eval_samples_per_second": 4.663,
+      "eval_steps_per_second": 0.591,
       "step": 3210
     },
     {
       "epoch": 6.0,
-      "eval_gen_len": 26.2667,
-      "eval_loss": 2.342738151550293,
-      "eval_rouge-1": 57.6653,
-      "eval_rouge-2": 44.2728,
-      "eval_rouge-l": 55.9363,
-      "eval_runtime": 127.5752,
-      "eval_samples_per_second": 3.527,
-      "eval_steps_per_second": 0.447,
+      "eval_gen_len": 19.7756,
+      "eval_loss": 2.973095178604126,
+      "eval_rouge-1": 37.8249,
+      "eval_rouge-2": 21.0556,
+      "eval_rouge-l": 37.3231,
+      "eval_runtime": 102.758,
+      "eval_samples_per_second": 4.379,
+      "eval_steps_per_second": 0.555,
       "step": 3852
     },
     {
       "epoch": 7.0,
-      "eval_gen_len": 26.1711,
-      "eval_loss": 2.3344807624816895,
-      "eval_rouge-1": 62.4422,
-      "eval_rouge-2": 50.3255,
-      "eval_rouge-l": 60.7065,
-      "eval_runtime": 127.1935,
-      "eval_samples_per_second": 3.538,
-      "eval_steps_per_second": 0.448,
+      "eval_gen_len": 18.5222,
+      "eval_loss": 3.0167415142059326,
+      "eval_rouge-1": 38.291,
+      "eval_rouge-2": 20.6795,
+      "eval_rouge-l": 37.3757,
+      "eval_runtime": 97.7101,
+      "eval_samples_per_second": 4.605,
+      "eval_steps_per_second": 0.583,
       "step": 4494
     },
     {
       "epoch": 7.79,
       "learning_rate": 8.715112540192926e-05,
-      "loss": 2.5283,
+      "loss": 2.8376,
       "step": 5000
     },
     {
       "epoch": 8.0,
-      "eval_gen_len": 26.3044,
-      "eval_loss": 2.3082587718963623,
-      "eval_rouge-1": 65.6385,
-      "eval_rouge-2": 55.386,
-      "eval_rouge-l": 64.2792,
-      "eval_runtime": 129.227,
-      "eval_samples_per_second": 3.482,
-      "eval_steps_per_second": 0.441,
+      "eval_gen_len": 18.0733,
+      "eval_loss": 3.062363624572754,
+      "eval_rouge-1": 39.3253,
+      "eval_rouge-2": 21.9098,
+      "eval_rouge-l": 38.5872,
+      "eval_runtime": 96.2268,
+      "eval_samples_per_second": 4.676,
+      "eval_steps_per_second": 0.592,
       "step": 5136
     },
     {
-      "epoch": 9.0,
-      "eval_gen_len": 26.0844,
-      "eval_loss": 2.288771390914917,
-      "eval_rouge-1": 69.2728,
-      "eval_rouge-2": 59.5819,
-      "eval_rouge-l": 68.0772,
-      "eval_runtime": 127.02,
-      "eval_samples_per_second": 3.543,
-      "eval_steps_per_second": 0.449,
-      "step": 5778
-    },
-    {
-      "epoch": 10.0,
-      "eval_gen_len": 26.5111,
-      "eval_loss": 2.2961621284484863,
-      "eval_rouge-1": 69.5351,
-      "eval_rouge-2": 60.4728,
-      "eval_rouge-l": 68.5563,
-      "eval_runtime": 128.3845,
-      "eval_samples_per_second": 3.505,
-      "eval_steps_per_second": 0.444,
-      "step": 6420
-    },
-    {
-      "epoch": 11.0,
-      "eval_gen_len": 26.4467,
-      "eval_loss": 2.294877052307129,
-      "eval_rouge-1": 71.1387,
-      "eval_rouge-2": 62.2653,
-      "eval_rouge-l": 69.9201,
-      "eval_runtime": 125.5037,
-      "eval_samples_per_second": 3.586,
-      "eval_steps_per_second": 0.454,
-      "step": 7062
-    },
-    {
-      "epoch": 12.0,
-      "eval_gen_len": 26.3911,
-      "eval_loss": 2.270343065261841,
-      "eval_rouge-1": 72.6498,
-      "eval_rouge-2": 63.8234,
-      "eval_rouge-l": 71.505,
-      "eval_runtime": 128.4901,
-      "eval_samples_per_second": 3.502,
-      "eval_steps_per_second": 0.444,
-      "step": 7704
-    },
-    {
-      "epoch": 13.0,
-      "eval_gen_len": 26.3556,
-      "eval_loss": 2.2668042182922363,
-      "eval_rouge-1": 73.2494,
-      "eval_rouge-2": 65.1934,
-      "eval_rouge-l": 72.2586,
-      "eval_runtime": 125.8233,
-      "eval_samples_per_second": 3.576,
-      "eval_steps_per_second": 0.453,
-      "step": 8346
-    },
-    {
-      "epoch": 14.0,
-      "eval_gen_len": 26.5333,
-      "eval_loss": 2.267202854156494,
-      "eval_rouge-1": 73.2228,
-      "eval_rouge-2": 64.7993,
-      "eval_rouge-l": 71.9902,
-      "eval_runtime": 128.1385,
-      "eval_samples_per_second": 3.512,
-      "eval_steps_per_second": 0.445,
-      "step": 8988
-    },
-    {
-      "epoch": 15.0,
-      "eval_gen_len": 27.2356,
-      "eval_loss": 2.2613940238952637,
-      "eval_rouge-1": 72.7529,
-      "eval_rouge-2": 64.3941,
-      "eval_rouge-l": 71.5976,
-      "eval_runtime": 129.55,
-      "eval_samples_per_second": 3.474,
-      "eval_steps_per_second": 0.44,
-      "step": 9630
-    },
-    {
-      "epoch": 15.58,
-      "learning_rate": 7.109003215434083e-05,
-      "loss": 1.7733,
-      "step": 10000
-    },
-    {
-      "epoch": 16.0,
-      "eval_gen_len": 26.5311,
-      "eval_loss": 2.2660067081451416,
-      "eval_rouge-1": 73.5002,
-      "eval_rouge-2": 65.0894,
-      "eval_rouge-l": 72.2767,
-      "eval_runtime": 129.1203,
-      "eval_samples_per_second": 3.485,
-      "eval_steps_per_second": 0.441,
-      "step": 10272
-    },
-    {
-      "epoch": 17.0,
-      "eval_gen_len": 26.4089,
-      "eval_loss": 2.2624781131744385,
-      "eval_rouge-1": 73.9016,
-      "eval_rouge-2": 66.1357,
-      "eval_rouge-l": 72.7275,
-      "eval_runtime": 125.6051,
-      "eval_samples_per_second": 3.583,
-      "eval_steps_per_second": 0.454,
-      "step": 10914
-    },
-    {
-      "epoch": 18.0,
-      "eval_gen_len": 26.0111,
-      "eval_loss": 2.2674496173858643,
-      "eval_rouge-1": 74.1365,
-      "eval_rouge-2": 66.7011,
-      "eval_rouge-l": 73.0304,
-      "eval_runtime": 123.8827,
-      "eval_samples_per_second": 3.632,
-      "eval_steps_per_second": 0.46,
-      "step": 11556
-    },
-    {
-      "epoch": 19.0,
-      "eval_gen_len": 26.7356,
-      "eval_loss": 2.264456033706665,
-      "eval_rouge-1": 74.1459,
-      "eval_rouge-2": 66.8989,
-      "eval_rouge-l": 73.1443,
-      "eval_runtime": 128.4694,
-      "eval_samples_per_second": 3.503,
-      "eval_steps_per_second": 0.444,
-      "step": 12198
-    },
-    {
-      "epoch": 20.0,
-      "eval_gen_len": 26.5311,
-      "eval_loss": 2.2706127166748047,
-      "eval_rouge-1": 74.3728,
-      "eval_rouge-2": 66.4207,
-      "eval_rouge-l": 73.2288,
-      "eval_runtime": 125.0151,
-      "eval_samples_per_second": 3.6,
-      "eval_steps_per_second": 0.456,
-      "step": 12840
-    },
-    {
-      "epoch": 20.0,
-      "step": 12840,
-      "total_flos": 3.46468539506688e+16,
-      "train_loss": 2.047671499282029,
-      "train_runtime": 9522.6649,
-      "train_samples_per_second": 26.951,
-      "train_steps_per_second": 3.371
+      "epoch": 8.0,
+      "step": 5136,
+      "total_flos": 1.346091953939712e+16,
+      "train_loss": 2.820291715247609,
+      "train_runtime": 3245.0074,
+      "train_samples_per_second": 79.091,
+      "train_steps_per_second": 9.892
     }
   ],
   "logging_steps": 5000,
   "max_steps": 32100,
   "num_train_epochs": 50,
   "save_steps": 500,
-  "total_flos": 3.46468539506688e+16,
+  "total_flos": 1.346091953939712e+16,
   "trial_name": null,
   "trial_params": null
 }

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:d444c556d98d61f5a310f8cd77b7aba754ec17362f58ab54e999eb36a8984a33
+oid sha256:c7d4d434932a5c4020dee8ce14e514f63ee54df0b17c30eea5c82c25f05a1da4
 size 4792