Training in progress, step 850, checkpoint

checkpoint-850/README.md

@@ -216,4 +216,4 @@ The following `bitsandbytes` quantization config was used during training:
 ### Framework versions
 
 
-- PEFT 0.6.0.dev0
+- PEFT 0.6.0

checkpoint-850/adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7b52ab4c4c927565a41f1d99513f6f953d146216b1089eb3182bf5688c5055d0
+size 9859800

checkpoint-850/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:544058eca646103eae4e5ed05297f05f52b50acfaa50b9c5b7b238e9d40708e4
+oid sha256:7b0625b6b6f7960d7e548611c5607962ea4bf5db365bb69e80cbf8b95bef3263
 size 42788

checkpoint-850/rng_state.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:6784940ad6cee59b35e412cf9a48d85bce003e4a5dc400826feb80b5059bf5a8
+oid sha256:93d4bff78b1b6aec43331984a63906e760531a251df861582d0db0df3c419aec
 size 14244

checkpoint-850/scheduler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:4f0f86f9dc4b0cc0536f7c9ff96dc960f4104628b6689268041ab40eb9a1e50b
+oid sha256:6be0f41d7109a2cb27d5156d66c2980a90a39bbf7ca5d16ad9948d236bd48a4c
 size 1064

checkpoint-850/special_tokens_map.json

@@ -2,6 +2,6 @@
   "bos_token": "<s>",
   "eos_token": "</s>",
   "pad_token": "<pad>",
-  "sep_token": "[SEP]",
+  "sep_token": "<s>",
   "unk_token": "<unk>"
 }

checkpoint-850/tokenizer_config.json

@@ -39,7 +39,7 @@
   "eos_token": "</s>",
   "model_max_length": 1000000000000000019884624838656,
   "pad_token": "<pad>",
-  "sep_token": "[SEP]",
+  "sep_token": "<s>",
   "tokenizer_class": "BloomTokenizer",
   "unk_token": "<unk>"
 }

checkpoint-850/trainer_state.json

@@ -1,7 +1,7 @@
 {
-  "best_metric": 0.6430063247680664,
-  "best_model_checkpoint": "output/checkpoint-50",
-  "epoch": 0.6814310051107325,
+  "best_metric": 0.5551439523696899,
+  "best_model_checkpoint": "output/checkpoint-400",
+  "epoch": 0.6818409706206758,
   "eval_steps": 50,
   "global_step": 850,
   "is_hyper_param_search": false,
@@ -11,270 +11,270 @@
     {
       "epoch": 0.0,
       "learning_rate": 4.444444444444445e-07,
-      "loss": 0.5542,
+      "loss": 1.0919,
       "step": 1
     },
     {
       "epoch": 0.04,
       "learning_rate": 2e-05,
-      "loss": 0.554,
+      "loss": 0.7661,
       "step": 50
     },
     {
       "epoch": 0.04,
-      "eval_accuracy": 0.6626865671641791,
-      "eval_loss": 0.6430063247680664,
-      "eval_runtime": 61.7692,
-      "eval_samples_per_second": 5.423,
-      "eval_steps_per_second": 1.36,
+      "eval_accuracy": 0.7142857142857143,
+      "eval_loss": 0.6954500079154968,
+      "eval_runtime": 28.4403,
+      "eval_samples_per_second": 6.153,
+      "eval_steps_per_second": 1.547,
       "step": 50
     },
     {
       "epoch": 0.08,
       "learning_rate": 4.222222222222222e-05,
-      "loss": 0.5957,
+      "loss": 0.6119,
       "step": 100
     },
     {
       "epoch": 0.08,
-      "eval_accuracy": 0.6686567164179105,
-      "eval_loss": 0.6705919504165649,
-      "eval_runtime": 60.4388,
-      "eval_samples_per_second": 5.543,
-      "eval_steps_per_second": 1.39,
+      "eval_accuracy": 0.7314285714285714,
+      "eval_loss": 0.5819395780563354,
+      "eval_runtime": 28.3595,
+      "eval_samples_per_second": 6.171,
+      "eval_steps_per_second": 1.552,
       "step": 100
     },
     {
       "epoch": 0.12,
       "learning_rate": 6.444444444444446e-05,
-      "loss": 0.5668,
+      "loss": 0.5869,
       "step": 150
     },
     {
       "epoch": 0.12,
-      "eval_accuracy": 0.6805970149253732,
-      "eval_loss": 0.7224913835525513,
-      "eval_runtime": 56.4882,
-      "eval_samples_per_second": 5.93,
-      "eval_steps_per_second": 1.487,
+      "eval_accuracy": 0.7371428571428571,
+      "eval_loss": 0.6004045605659485,
+      "eval_runtime": 28.3043,
+      "eval_samples_per_second": 6.183,
+      "eval_steps_per_second": 1.555,
       "step": 150
     },
     {
       "epoch": 0.16,
       "learning_rate": 8.666666666666667e-05,
-      "loss": 0.5626,
+      "loss": 0.588,
       "step": 200
     },
     {
       "epoch": 0.16,
-      "eval_accuracy": 0.6746268656716418,
-      "eval_loss": 0.7115103006362915,
-      "eval_runtime": 56.506,
-      "eval_samples_per_second": 5.929,
-      "eval_steps_per_second": 1.487,
+      "eval_accuracy": 0.76,
+      "eval_loss": 0.5582989454269409,
+      "eval_runtime": 28.3002,
+      "eval_samples_per_second": 6.184,
+      "eval_steps_per_second": 1.555,
       "step": 200
     },
     {
       "epoch": 0.2,
-      "learning_rate": 9.80430528375734e-05,
-      "loss": 0.5604,
+      "learning_rate": 9.80411361410382e-05,
+      "loss": 0.5701,
       "step": 250
     },
     {
       "epoch": 0.2,
-      "eval_accuracy": 0.6746268656716418,
-      "eval_loss": 0.700947105884552,
-      "eval_runtime": 56.4927,
-      "eval_samples_per_second": 5.93,
-      "eval_steps_per_second": 1.487,
+      "eval_accuracy": 0.7142857142857143,
+      "eval_loss": 0.6456161141395569,
+      "eval_runtime": 28.2527,
+      "eval_samples_per_second": 6.194,
+      "eval_steps_per_second": 1.557,
       "step": 250
     },
     {
       "epoch": 0.24,
-      "learning_rate": 9.315068493150684e-05,
-      "loss": 0.5675,
+      "learning_rate": 9.31439764936337e-05,
+      "loss": 0.5652,
       "step": 300
     },
     {
       "epoch": 0.24,
-      "eval_accuracy": 0.6597014925373135,
-      "eval_loss": 0.662155032157898,
-      "eval_runtime": 56.4903,
-      "eval_samples_per_second": 5.93,
-      "eval_steps_per_second": 1.487,
+      "eval_accuracy": 0.72,
+      "eval_loss": 0.6802676916122437,
+      "eval_runtime": 28.3389,
+      "eval_samples_per_second": 6.175,
+      "eval_steps_per_second": 1.553,
       "step": 300
     },
     {
       "epoch": 0.28,
-      "learning_rate": 8.825831702544032e-05,
-      "loss": 0.5725,
+      "learning_rate": 8.82468168462292e-05,
+      "loss": 0.5936,
       "step": 350
     },
     {
       "epoch": 0.28,
-      "eval_accuracy": 0.6567164179104478,
-      "eval_loss": 0.6746364831924438,
-      "eval_runtime": 56.482,
-      "eval_samples_per_second": 5.931,
-      "eval_steps_per_second": 1.487,
+      "eval_accuracy": 0.7542857142857143,
+      "eval_loss": 0.5668751001358032,
+      "eval_runtime": 28.2611,
+      "eval_samples_per_second": 6.192,
+      "eval_steps_per_second": 1.557,
       "step": 350
     },
     {
       "epoch": 0.32,
-      "learning_rate": 8.336594911937378e-05,
-      "loss": 0.5884,
+      "learning_rate": 8.334965719882468e-05,
+      "loss": 0.5961,
       "step": 400
     },
     {
       "epoch": 0.32,
-      "eval_accuracy": 0.6656716417910448,
-      "eval_loss": 0.7047849893569946,
-      "eval_runtime": 56.4873,
-      "eval_samples_per_second": 5.931,
-      "eval_steps_per_second": 1.487,
+      "eval_accuracy": 0.76,
+      "eval_loss": 0.5551439523696899,
+      "eval_runtime": 26.2906,
+      "eval_samples_per_second": 6.656,
+      "eval_steps_per_second": 1.674,
       "step": 400
     },
     {
       "epoch": 0.36,
-      "learning_rate": 7.847358121330725e-05,
-      "loss": 0.5567,
+      "learning_rate": 7.845249755142018e-05,
+      "loss": 0.5842,
       "step": 450
     },
     {
       "epoch": 0.36,
-      "eval_accuracy": 0.6477611940298508,
-      "eval_loss": 0.6832321286201477,
-      "eval_runtime": 56.4814,
-      "eval_samples_per_second": 5.931,
-      "eval_steps_per_second": 1.487,
+      "eval_accuracy": 0.7314285714285714,
+      "eval_loss": 0.6041642427444458,
+      "eval_runtime": 26.2867,
+      "eval_samples_per_second": 6.657,
+      "eval_steps_per_second": 1.674,
       "step": 450
     },
     {
       "epoch": 0.4,
-      "learning_rate": 7.35812133072407e-05,
-      "loss": 0.5723,
+      "learning_rate": 7.355533790401568e-05,
+      "loss": 0.5749,
       "step": 500
     },
     {
       "epoch": 0.4,
-      "eval_accuracy": 0.6955223880597015,
-      "eval_loss": 0.7437570691108704,
-      "eval_runtime": 56.4677,
-      "eval_samples_per_second": 5.933,
-      "eval_steps_per_second": 1.488,
+      "eval_accuracy": 0.7257142857142858,
+      "eval_loss": 0.6361325979232788,
+      "eval_runtime": 28.3165,
+      "eval_samples_per_second": 6.18,
+      "eval_steps_per_second": 1.554,
       "step": 500
     },
     {
       "epoch": 0.44,
-      "learning_rate": 6.868884540117417e-05,
-      "loss": 0.5908,
+      "learning_rate": 6.865817825661118e-05,
+      "loss": 0.6174,
       "step": 550
     },
     {
       "epoch": 0.44,
-      "eval_accuracy": 0.6238805970149254,
-      "eval_loss": 0.6612433791160583,
-      "eval_runtime": 56.4799,
-      "eval_samples_per_second": 5.931,
-      "eval_steps_per_second": 1.487,
+      "eval_accuracy": 0.7485714285714286,
+      "eval_loss": 0.582287073135376,
+      "eval_runtime": 28.3412,
+      "eval_samples_per_second": 6.175,
+      "eval_steps_per_second": 1.553,
       "step": 550
     },
     {
       "epoch": 0.48,
-      "learning_rate": 6.379647749510763e-05,
-      "loss": 0.5681,
+      "learning_rate": 6.376101860920667e-05,
+      "loss": 0.5752,
       "step": 600
     },
     {
       "epoch": 0.48,
-      "eval_accuracy": 0.6507462686567164,
-      "eval_loss": 0.6833732724189758,
-      "eval_runtime": 56.4689,
-      "eval_samples_per_second": 5.932,
-      "eval_steps_per_second": 1.488,
+      "eval_accuracy": 0.7314285714285714,
+      "eval_loss": 0.6466975212097168,
+      "eval_runtime": 28.3593,
+      "eval_samples_per_second": 6.171,
+      "eval_steps_per_second": 1.552,
       "step": 600
     },
     {
       "epoch": 0.52,
-      "learning_rate": 5.89041095890411e-05,
-      "loss": 0.5782,
+      "learning_rate": 5.886385896180215e-05,
+      "loss": 0.5679,
       "step": 650
     },
     {
       "epoch": 0.52,
-      "eval_accuracy": 0.6507462686567164,
-      "eval_loss": 0.67350834608078,
-      "eval_runtime": 56.4637,
-      "eval_samples_per_second": 5.933,
-      "eval_steps_per_second": 1.488,
+      "eval_accuracy": 0.72,
+      "eval_loss": 0.5773769617080688,
+      "eval_runtime": 28.2118,
+      "eval_samples_per_second": 6.203,
+      "eval_steps_per_second": 1.56,
       "step": 650
     },
     {
       "epoch": 0.56,
-      "learning_rate": 5.401174168297456e-05,
-      "loss": 0.569,
+      "learning_rate": 5.396669931439765e-05,
+      "loss": 0.5793,
       "step": 700
     },
     {
       "epoch": 0.56,
-      "eval_accuracy": 0.6686567164179105,
-      "eval_loss": 0.7085319757461548,
-      "eval_runtime": 56.4627,
-      "eval_samples_per_second": 5.933,
-      "eval_steps_per_second": 1.488,
+      "eval_accuracy": 0.6914285714285714,
+      "eval_loss": 0.5711085796356201,
+      "eval_runtime": 28.2951,
+      "eval_samples_per_second": 6.185,
+      "eval_steps_per_second": 1.555,
       "step": 700
     },
     {
       "epoch": 0.6,
-      "learning_rate": 4.911937377690802e-05,
-      "loss": 0.5661,
+      "learning_rate": 4.906953966699315e-05,
+      "loss": 0.5897,
       "step": 750
     },
     {
       "epoch": 0.6,
-      "eval_accuracy": 0.6835820895522388,
-      "eval_loss": 0.7564846277236938,
-      "eval_runtime": 56.4648,
-      "eval_samples_per_second": 5.933,
-      "eval_steps_per_second": 1.488,
+      "eval_accuracy": 0.7028571428571428,
+      "eval_loss": 0.5825155377388,
+      "eval_runtime": 28.19,
+      "eval_samples_per_second": 6.208,
+      "eval_steps_per_second": 1.561,
       "step": 750
     },
     {
       "epoch": 0.64,
-      "learning_rate": 4.422700587084149e-05,
-      "loss": 0.5675,
+      "learning_rate": 4.4172380019588636e-05,
+      "loss": 0.5782,
       "step": 800
     },
     {
       "epoch": 0.64,
-      "eval_accuracy": 0.6417910447761194,
-      "eval_loss": 0.6759604215621948,
-      "eval_runtime": 56.4652,
-      "eval_samples_per_second": 5.933,
-      "eval_steps_per_second": 1.488,
+      "eval_accuracy": 0.7257142857142858,
+      "eval_loss": 0.5976516008377075,
+      "eval_runtime": 28.2521,
+      "eval_samples_per_second": 6.194,
+      "eval_steps_per_second": 1.557,
       "step": 800
     },
     {
       "epoch": 0.68,
-      "learning_rate": 3.933463796477495e-05,
-      "loss": 0.5611,
+      "learning_rate": 3.927522037218414e-05,
+      "loss": 0.593,
       "step": 850
     },
     {
       "epoch": 0.68,
-      "eval_accuracy": 0.6477611940298508,
-      "eval_loss": 0.6798116564750671,
-      "eval_runtime": 56.4714,
-      "eval_samples_per_second": 5.932,
-      "eval_steps_per_second": 1.487,
+      "eval_accuracy": 0.7257142857142858,
+      "eval_loss": 0.6304438710212708,
+      "eval_runtime": 28.304,
+      "eval_samples_per_second": 6.183,
+      "eval_steps_per_second": 1.555,
       "step": 850
     }
   ],
   "logging_steps": 50,
-  "max_steps": 1247,
+  "max_steps": 1246,
   "num_train_epochs": 1,
   "save_steps": 50,
-  "total_flos": 3.173042205045965e+17,
+  "total_flos": 3.150688487167181e+17,
   "trial_name": null,
   "trial_params": null
 }

checkpoint-850/training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:87bd30bd4b21b4a0c66f1fd545858ef6e2d33b5984a8d292e79011acc3affd6b
-size 4472
+oid sha256:b25ecbcf37659c2f593e1603dab54c3eaea43862c7f192a7e6df4c4dcc86d20d
+size 4600