Training in progress, step 650, checkpoint

checkpoint-650/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-650/adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:98026cfbf59aca63cacde04ce0e65b037cc6038bedf649deb244f18453bf25e4
+size 9859800

checkpoint-650/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:71c3b1bd572d49227acc3d03118a098e4ffe804993869774b897b875e3d633dc
+oid sha256:9ce5af8d936731d026dd4b1d099d3afa1c3c3fd63294529bf43a6a8711cca5d6
 size 42788

checkpoint-650/rng_state.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:c56079d48a243cf026c6d3d5c378236d3c59649d65fd49aa8ae53890ee109464
+oid sha256:cf38f0d117bdbe49b60f913177ef967fab472747cc09d39dde8ac43699e24c9d
 size 14244

checkpoint-650/scheduler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:0085ea5b951ca032bafeceed85ef10530bbbbc633f304cdd4cb4369d24375219
+oid sha256:519de33a8b24b33a9ac94de0012c572f6ab10543df30cd715b717a976ff8f2e5
 size 1064

checkpoint-650/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-650/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-650/trainer_state.json

@@ -1,7 +1,7 @@
 {
-  "best_metric": 0.665068507194519,
-  "best_model_checkpoint": "output/checkpoint-50",
-  "epoch": 2.659846547314578,
+  "best_metric": 0.5551439523696899,
+  "best_model_checkpoint": "output/checkpoint-400",
+  "epoch": 0.5214078010628698,
   "eval_steps": 50,
   "global_step": 650,
   "is_hyper_param_search": false,
@@ -10,211 +10,211 @@
   "log_history": [
     {
       "epoch": 0.0,
-      "learning_rate": 7.575757575757576e-07,
-      "loss": 0.4968,
+      "learning_rate": 4.444444444444445e-07,
+      "loss": 1.0919,
       "step": 1
     },
     {
-      "epoch": 0.2,
-      "learning_rate": 3.484848484848485e-05,
-      "loss": 0.5254,
+      "epoch": 0.04,
+      "learning_rate": 2e-05,
+      "loss": 0.7661,
       "step": 50
     },
     {
-      "epoch": 0.2,
-      "eval_accuracy": 0.6507462686567164,
-      "eval_loss": 0.665068507194519,
-      "eval_runtime": 61.5592,
-      "eval_samples_per_second": 5.442,
-      "eval_steps_per_second": 1.365,
+      "epoch": 0.04,
+      "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": 0.41,
-      "learning_rate": 7.272727272727273e-05,
-      "loss": 0.4971,
+      "epoch": 0.08,
+      "learning_rate": 4.222222222222222e-05,
+      "loss": 0.6119,
       "step": 100
     },
     {
-      "epoch": 0.41,
-      "eval_accuracy": 0.6656716417910448,
-      "eval_loss": 0.8002150058746338,
-      "eval_runtime": 61.211,
-      "eval_samples_per_second": 5.473,
-      "eval_steps_per_second": 1.372,
+      "epoch": 0.08,
+      "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": 0.61,
-      "learning_rate": 9.783333333333334e-05,
-      "loss": 0.5039,
+      "epoch": 0.12,
+      "learning_rate": 6.444444444444446e-05,
+      "loss": 0.5869,
       "step": 150
     },
     {
-      "epoch": 0.61,
-      "eval_accuracy": 0.6865671641791045,
-      "eval_loss": 0.7404947280883789,
-      "eval_runtime": 61.4487,
-      "eval_samples_per_second": 5.452,
-      "eval_steps_per_second": 1.367,
+      "epoch": 0.12,
+      "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": 0.82,
-      "learning_rate": 8.950000000000001e-05,
-      "loss": 0.4944,
+      "epoch": 0.16,
+      "learning_rate": 8.666666666666667e-05,
+      "loss": 0.588,
       "step": 200
     },
     {
-      "epoch": 0.82,
-      "eval_accuracy": 0.6238805970149254,
-      "eval_loss": 0.9986834526062012,
-      "eval_runtime": 60.5135,
-      "eval_samples_per_second": 5.536,
-      "eval_steps_per_second": 1.388,
+      "epoch": 0.16,
+      "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": 1.02,
-      "learning_rate": 8.116666666666667e-05,
-      "loss": 0.508,
+      "epoch": 0.2,
+      "learning_rate": 9.80411361410382e-05,
+      "loss": 0.5701,
       "step": 250
     },
     {
-      "epoch": 1.02,
-      "eval_accuracy": 0.6268656716417911,
-      "eval_loss": 0.9990125894546509,
-      "eval_runtime": 56.4381,
-      "eval_samples_per_second": 5.936,
-      "eval_steps_per_second": 1.488,
+      "epoch": 0.2,
+      "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": 1.23,
-      "learning_rate": 7.283333333333335e-05,
-      "loss": 0.4904,
+      "epoch": 0.24,
+      "learning_rate": 9.31439764936337e-05,
+      "loss": 0.5652,
       "step": 300
     },
     {
-      "epoch": 1.23,
-      "eval_accuracy": 0.6119402985074627,
-      "eval_loss": 1.0483125448226929,
-      "eval_runtime": 56.4117,
-      "eval_samples_per_second": 5.938,
-      "eval_steps_per_second": 1.489,
+      "epoch": 0.24,
+      "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": 1.43,
-      "learning_rate": 6.450000000000001e-05,
-      "loss": 0.4694,
+      "epoch": 0.28,
+      "learning_rate": 8.82468168462292e-05,
+      "loss": 0.5936,
       "step": 350
     },
     {
-      "epoch": 1.43,
-      "eval_accuracy": 0.6417910447761194,
-      "eval_loss": 0.9735069274902344,
-      "eval_runtime": 56.422,
-      "eval_samples_per_second": 5.937,
-      "eval_steps_per_second": 1.489,
+      "epoch": 0.28,
+      "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": 1.64,
-      "learning_rate": 5.6166666666666665e-05,
-      "loss": 0.4937,
+      "epoch": 0.32,
+      "learning_rate": 8.334965719882468e-05,
+      "loss": 0.5961,
       "step": 400
     },
     {
-      "epoch": 1.64,
-      "eval_accuracy": 0.5880597014925373,
-      "eval_loss": 1.1042370796203613,
-      "eval_runtime": 56.4248,
-      "eval_samples_per_second": 5.937,
-      "eval_steps_per_second": 1.489,
+      "epoch": 0.32,
+      "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": 1.84,
-      "learning_rate": 4.7833333333333335e-05,
-      "loss": 0.4877,
+      "epoch": 0.36,
+      "learning_rate": 7.845249755142018e-05,
+      "loss": 0.5842,
       "step": 450
     },
     {
-      "epoch": 1.84,
-      "eval_accuracy": 0.6477611940298508,
-      "eval_loss": 0.9591583609580994,
-      "eval_runtime": 56.4028,
-      "eval_samples_per_second": 5.939,
-      "eval_steps_per_second": 1.489,
+      "epoch": 0.36,
+      "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": 2.05,
-      "learning_rate": 3.9500000000000005e-05,
-      "loss": 0.482,
+      "epoch": 0.4,
+      "learning_rate": 7.355533790401568e-05,
+      "loss": 0.5749,
       "step": 500
     },
     {
-      "epoch": 2.05,
-      "eval_accuracy": 0.6716417910447762,
-      "eval_loss": 0.89241623878479,
-      "eval_runtime": 56.408,
-      "eval_samples_per_second": 5.939,
-      "eval_steps_per_second": 1.489,
+      "epoch": 0.4,
+      "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": 2.25,
-      "learning_rate": 3.116666666666667e-05,
-      "loss": 0.4688,
+      "epoch": 0.44,
+      "learning_rate": 6.865817825661118e-05,
+      "loss": 0.6174,
       "step": 550
     },
     {
-      "epoch": 2.25,
-      "eval_accuracy": 0.6447761194029851,
-      "eval_loss": 0.9692543148994446,
-      "eval_runtime": 56.4147,
-      "eval_samples_per_second": 5.938,
-      "eval_steps_per_second": 1.489,
+      "epoch": 0.44,
+      "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": 2.46,
-      "learning_rate": 2.2833333333333334e-05,
-      "loss": 0.4818,
+      "epoch": 0.48,
+      "learning_rate": 6.376101860920667e-05,
+      "loss": 0.5752,
       "step": 600
     },
     {
-      "epoch": 2.46,
-      "eval_accuracy": 0.6417910447761194,
-      "eval_loss": 0.981917679309845,
-      "eval_runtime": 56.4148,
-      "eval_samples_per_second": 5.938,
-      "eval_steps_per_second": 1.489,
+      "epoch": 0.48,
+      "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": 2.66,
-      "learning_rate": 1.45e-05,
-      "loss": 0.4663,
+      "epoch": 0.52,
+      "learning_rate": 5.886385896180215e-05,
+      "loss": 0.5679,
       "step": 650
     },
     {
-      "epoch": 2.66,
-      "eval_accuracy": 0.6417910447761194,
-      "eval_loss": 0.9769092202186584,
-      "eval_runtime": 56.4075,
-      "eval_samples_per_second": 5.939,
-      "eval_steps_per_second": 1.489,
+      "epoch": 0.52,
+      "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
     }
   ],
   "logging_steps": 50,
-  "max_steps": 732,
-  "num_train_epochs": 3,
+  "max_steps": 1246,
+  "num_train_epochs": 1,
   "save_steps": 50,
-  "total_flos": 2.7691286123624448e+17,
+  "total_flos": 2.4083913788694528e+17,
   "trial_name": null,
   "trial_params": null
 }

checkpoint-650/training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:9feea516a8a208c393c324d84d047c0ab5036bbdf3fb1f3eea6a2563e9c79a6b
-size 4472
+oid sha256:b25ecbcf37659c2f593e1603dab54c3eaea43862c7f192a7e6df4c4dcc86d20d
+size 4600