Training in progress, step 500, checkpoint

checkpoint-500/adapter_config.json

@@ -1,7 +1,7 @@
 {
   "alpha_pattern": {},
   "auto_mapping": null,
-  "base_model_name_or_path": "/kaggle/input/vietcuna-for-binary-classification/kalapa-vietcuna-3b",
+  "base_model_name_or_path": "vietcuna-3b-v2/kalapa-vietcuna-3b/",
   "bias": "none",
   "fan_in_fan_out": false,
   "inference_mode": true,
@@ -9,10 +9,10 @@
   "layers_pattern": null,
   "layers_to_transform": null,
   "lora_alpha": 32,
-  "lora_dropout": 0.1,
+  "lora_dropout": 0.05,
   "modules_to_save": null,
   "peft_type": "LORA",
-  "r": 8,
+  "r": 16,
   "rank_pattern": {},
   "revision": null,
   "target_modules": [

checkpoint-500/adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:db222e1a147b6a8ab2a150b3b76f91c052c2d93d984e14053a365a38bf0bfb04
-size 9859800
+oid sha256:6db92a12b77676b0343d62b007088b0a738bece00d0c41341971775bf642b185
+size 19690328

checkpoint-500/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:86b122bbaefe7662c7c2e6142ccdc484e46b407f5bf93721d79ddf02adf7054a
-size 42788
+oid sha256:7aeae66019a19277d7702df58a327e1efde9f30e994d526598b3f31685ac5c27
+size 38087162

checkpoint-500/rng_state.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:7f108122d80153aac758c3467bef120bc47696a5ba28eb813b9d17e3b998c831
+oid sha256:dd8652d2ed9fd8591d94e3ecdf652e2bbb2a82e873960bc84e210d60c0cfaf3f
 size 14244

checkpoint-500/scheduler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:f7ec3f4f7006fb7a1d4b5ed7b1fa2abb9485e6a816f42ae04481c9fa33c29c7b
+oid sha256:3c9aec9a0bb79e6d56e4e5df00ca0d57e668ee525c9d1ed0e14c3fce618d34f1
 size 1064

checkpoint-500/special_tokens_map.json

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

checkpoint-500/tokenizer_config.json

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

checkpoint-500/trainer_state.json

@@ -1,7 +1,7 @@
 {
-  "best_metric": 0.522564172744751,
-  "best_model_checkpoint": "output/checkpoint-50",
-  "epoch": 3.4904013961605584,
+  "best_metric": 0.7192752361297607,
+  "best_model_checkpoint": "output/checkpoint-400",
+  "epoch": 2.5526483726866624,
   "eval_steps": 50,
   "global_step": 500,
   "is_hyper_param_search": false,
@@ -10,166 +10,166 @@
   "log_history": [
     {
       "epoch": 0.01,
-      "learning_rate": 1.0000000000000001e-07,
-      "loss": 0.4196,
+      "learning_rate": 2.777777777777778e-07,
+      "loss": 9.5569,
       "step": 1
     },
     {
-      "epoch": 0.35,
-      "learning_rate": 4.6e-06,
-      "loss": 0.4279,
+      "epoch": 0.26,
+      "learning_rate": 1.25e-05,
+      "loss": 6.958,
       "step": 50
     },
     {
-      "epoch": 0.35,
-      "eval_accuracy": 0.7657142857142857,
-      "eval_loss": 0.522564172744751,
-      "eval_runtime": 29.0326,
-      "eval_samples_per_second": 6.028,
-      "eval_steps_per_second": 1.516,
+      "epoch": 0.26,
+      "eval_accuracy": 0.42857142857142855,
+      "eval_loss": 6.581231117248535,
+      "eval_runtime": 28.2966,
+      "eval_samples_per_second": 6.184,
+      "eval_steps_per_second": 1.555,
       "step": 50
     },
     {
-      "epoch": 0.7,
-      "learning_rate": 8.86829268292683e-06,
-      "loss": 0.4089,
+      "epoch": 0.51,
+      "learning_rate": 2.6388888888888892e-05,
+      "loss": 1.6141,
       "step": 100
     },
     {
-      "epoch": 0.7,
-      "eval_accuracy": 0.7428571428571429,
-      "eval_loss": 0.5401451587677002,
-      "eval_runtime": 29.0003,
-      "eval_samples_per_second": 6.034,
-      "eval_steps_per_second": 1.517,
+      "epoch": 0.51,
+      "eval_accuracy": 0.5028571428571429,
+      "eval_loss": 1.6640697717666626,
+      "eval_runtime": 28.1139,
+      "eval_samples_per_second": 6.225,
+      "eval_steps_per_second": 1.565,
       "step": 100
     },
     {
-      "epoch": 1.05,
-      "learning_rate": 7.792682926829268e-06,
-      "loss": 0.4018,
+      "epoch": 0.77,
+      "learning_rate": 4.027777777777778e-05,
+      "loss": 0.8533,
       "step": 150
     },
     {
-      "epoch": 1.05,
-      "eval_accuracy": 0.7428571428571429,
-      "eval_loss": 0.5425500273704529,
-      "eval_runtime": 28.9813,
-      "eval_samples_per_second": 6.038,
-      "eval_steps_per_second": 1.518,
+      "epoch": 0.77,
+      "eval_accuracy": 0.6342857142857142,
+      "eval_loss": 1.2631328105926514,
+      "eval_runtime": 26.205,
+      "eval_samples_per_second": 6.678,
+      "eval_steps_per_second": 1.679,
       "step": 150
     },
     {
-      "epoch": 1.4,
-      "learning_rate": 6.695121951219512e-06,
-      "loss": 0.41,
+      "epoch": 1.02,
+      "learning_rate": 4.908536585365854e-05,
+      "loss": 0.4474,
       "step": 200
     },
     {
-      "epoch": 1.4,
-      "eval_accuracy": 0.7542857142857143,
-      "eval_loss": 0.5254732966423035,
-      "eval_runtime": 28.8187,
-      "eval_samples_per_second": 6.072,
-      "eval_steps_per_second": 1.527,
+      "epoch": 1.02,
+      "eval_accuracy": 0.7142857142857143,
+      "eval_loss": 0.9036799073219299,
+      "eval_runtime": 26.208,
+      "eval_samples_per_second": 6.677,
+      "eval_steps_per_second": 1.679,
       "step": 200
     },
     {
-      "epoch": 1.75,
-      "learning_rate": 5.597560975609756e-06,
-      "loss": 0.3927,
+      "epoch": 1.28,
+      "learning_rate": 4.603658536585366e-05,
+      "loss": 0.2996,
       "step": 250
     },
     {
-      "epoch": 1.75,
-      "eval_accuracy": 0.7485714285714286,
-      "eval_loss": 0.5406104326248169,
-      "eval_runtime": 27.5585,
-      "eval_samples_per_second": 6.35,
-      "eval_steps_per_second": 1.597,
+      "epoch": 1.28,
+      "eval_accuracy": 0.6742857142857143,
+      "eval_loss": 1.1182247400283813,
+      "eval_runtime": 26.2071,
+      "eval_samples_per_second": 6.678,
+      "eval_steps_per_second": 1.679,
       "step": 250
     },
     {
-      "epoch": 2.09,
-      "learning_rate": 4.5e-06,
-      "loss": 0.428,
+      "epoch": 1.53,
+      "learning_rate": 4.298780487804878e-05,
+      "loss": 0.2403,
       "step": 300
     },
     {
-      "epoch": 2.09,
-      "eval_accuracy": 0.76,
-      "eval_loss": 0.5347301363945007,
-      "eval_runtime": 28.9003,
-      "eval_samples_per_second": 6.055,
-      "eval_steps_per_second": 1.522,
+      "epoch": 1.53,
+      "eval_accuracy": 0.72,
+      "eval_loss": 1.076162338256836,
+      "eval_runtime": 26.2142,
+      "eval_samples_per_second": 6.676,
+      "eval_steps_per_second": 1.678,
       "step": 300
     },
     {
-      "epoch": 2.44,
-      "learning_rate": 3.402439024390244e-06,
-      "loss": 0.403,
+      "epoch": 1.79,
+      "learning_rate": 3.9939024390243905e-05,
+      "loss": 0.1982,
       "step": 350
     },
     {
-      "epoch": 2.44,
-      "eval_accuracy": 0.7428571428571429,
-      "eval_loss": 0.5431011319160461,
-      "eval_runtime": 26.5045,
-      "eval_samples_per_second": 6.603,
-      "eval_steps_per_second": 1.66,
+      "epoch": 1.79,
+      "eval_accuracy": 0.6571428571428571,
+      "eval_loss": 1.4400347471237183,
+      "eval_runtime": 26.2187,
+      "eval_samples_per_second": 6.675,
+      "eval_steps_per_second": 1.678,
       "step": 350
     },
     {
-      "epoch": 2.79,
-      "learning_rate": 2.304878048780488e-06,
-      "loss": 0.4024,
+      "epoch": 2.04,
+      "learning_rate": 3.6890243902439025e-05,
+      "loss": 0.1926,
       "step": 400
     },
     {
-      "epoch": 2.79,
-      "eval_accuracy": 0.7485714285714286,
-      "eval_loss": 0.5433735847473145,
-      "eval_runtime": 26.4908,
-      "eval_samples_per_second": 6.606,
-      "eval_steps_per_second": 1.661,
+      "epoch": 2.04,
+      "eval_accuracy": 0.7942857142857143,
+      "eval_loss": 0.7192752361297607,
+      "eval_runtime": 26.2201,
+      "eval_samples_per_second": 6.674,
+      "eval_steps_per_second": 1.678,
       "step": 400
     },
     {
-      "epoch": 3.14,
-      "learning_rate": 1.2073170731707318e-06,
-      "loss": 0.4208,
+      "epoch": 2.3,
+      "learning_rate": 3.384146341463415e-05,
+      "loss": 0.1258,
       "step": 450
     },
     {
-      "epoch": 3.14,
-      "eval_accuracy": 0.76,
-      "eval_loss": 0.5310953259468079,
-      "eval_runtime": 26.5064,
-      "eval_samples_per_second": 6.602,
-      "eval_steps_per_second": 1.66,
+      "epoch": 2.3,
+      "eval_accuracy": 0.7657142857142857,
+      "eval_loss": 1.0054305791854858,
+      "eval_runtime": 26.2093,
+      "eval_samples_per_second": 6.677,
+      "eval_steps_per_second": 1.679,
       "step": 450
     },
     {
-      "epoch": 3.49,
-      "learning_rate": 1.0975609756097562e-07,
-      "loss": 0.4032,
+      "epoch": 2.55,
+      "learning_rate": 3.079268292682927e-05,
+      "loss": 0.0857,
       "step": 500
     },
     {
-      "epoch": 3.49,
-      "eval_accuracy": 0.76,
-      "eval_loss": 0.5342433452606201,
-      "eval_runtime": 26.4949,
-      "eval_samples_per_second": 6.605,
-      "eval_steps_per_second": 1.661,
+      "epoch": 2.55,
+      "eval_accuracy": 0.7828571428571428,
+      "eval_loss": 1.1228740215301514,
+      "eval_runtime": 26.1929,
+      "eval_samples_per_second": 6.681,
+      "eval_steps_per_second": 1.68,
       "step": 500
     }
   ],
   "logging_steps": 50,
-  "max_steps": 500,
-  "num_train_epochs": 4,
+  "max_steps": 1000,
+  "num_train_epochs": 6,
   "save_steps": 50,
-  "total_flos": 1.8791934766301184e+17,
+  "total_flos": 1.863409147892736e+17,
   "trial_name": null,
   "trial_params": null
 }

checkpoint-500/training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:b1b0a94354c0f3dfee9c93e225dbd48ad6f856f98977ef38ff3415910328e64a
+oid sha256:5a0ef36d7c31265f68a3edc8ae53ce9272dcc90a79f261e29a2a46e040c77a33
 size 4600