Training in progress, step 58000, checkpoint

checkpoint-58000/config.json

@@ -0,0 +1,39 @@
+{
+  "_name_or_path": "gpt2",
+  "activation_function": "gelu_new",
+  "architectures": [
+    "GPT2LMHeadModel"
+  ],
+  "attn_pdrop": 0.1,
+  "bos_token_id": 5,
+  "embd_pdrop": 0.1,
+  "eos_token_id": 4,
+  "initializer_range": 0.02,
+  "layer_norm_epsilon": 1e-05,
+  "model_type": "gpt2",
+  "n_ctx": 128,
+  "n_embd": 128,
+  "n_head": 4,
+  "n_inner": null,
+  "n_layer": 4,
+  "n_positions": 128,
+  "reorder_and_upcast_attn": false,
+  "resid_pdrop": 0.1,
+  "scale_attn_by_inverse_layer_idx": false,
+  "scale_attn_weights": true,
+  "summary_activation": null,
+  "summary_first_dropout": 0.1,
+  "summary_proj_to_labels": true,
+  "summary_type": "cls_index",
+  "summary_use_proj": true,
+  "task_specific_params": {
+    "text-generation": {
+      "do_sample": true,
+      "max_length": 50
+    }
+  },
+  "torch_dtype": "float32",
+  "transformers_version": "4.45.2",
+  "use_cache": true,
+  "vocab_size": 8192
+}

checkpoint-58000/generation_config.json

@@ -0,0 +1,6 @@
+{
+  "_from_model_config": true,
+  "bos_token_id": 5,
+  "eos_token_id": 4,
+  "transformers_version": "4.45.2"
+}

checkpoint-58000/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ec17ab192175603914859365cde78aa6f4419a87c537d7faf0dc91e670df2d2a
+size 7438256

checkpoint-58000/optimizer.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:49121e83370977aec8c6dfc200e4da523596161d1495777f9405507e0c5948f8
+size 14908602

checkpoint-58000/rng_state.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fc389a2b8ab6ea55db8e084309f05ce0e8f297067a478bed277d844656375a68
+size 13990

checkpoint-58000/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:77ee956d507eb33601245f0cbd3303997b6e999df3836aeed519409765381e06
+size 1192

checkpoint-58000/special_tokens_map.json

@@ -0,0 +1,17 @@
+{
+  "bos_token": {
+    "content": "[BOS]",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "eos_token": {
+    "content": "[EOS]",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": "[EOS]"
+}

checkpoint-58000/tokenizer_config.json

@@ -0,0 +1,58 @@
+{
+  "added_tokens_decoder": {
+    "0": {
+      "content": "[UNK]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "1": {
+      "content": "[CLS]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "2": {
+      "content": "[SEP]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "3": {
+      "content": "[PAD]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "4": {
+      "content": "[EOS]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "5": {
+      "content": "[BOS]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "bos_token": "[BOS]",
+  "clean_up_tokenization_spaces": true,
+  "eos_token": "[EOS]",
+  "model_max_length": 1000000000000000019884624838656,
+  "pad_token": "[EOS]",
+  "tokenizer_class": "PreTrainedTokenizerFast"
+}

checkpoint-58000/trainer_state.json

@@ -0,0 +1,477 @@
+{
+  "best_metric": 3.4978885650634766,
+  "best_model_checkpoint": "/Users/frapadovani/Desktop/babyLM_controlled/models_trained/reduce_on_plateau_convergence/age_random/checkpoint-58000",
+  "epoch": 0.30690916018012393,
+  "eval_steps": 2000,
+  "global_step": 58000,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.010583074488969791,
+      "grad_norm": 1.3282362222671509,
+      "learning_rate": 0.0001,
+      "loss": 4.8574,
+      "step": 2000
+    },
+    {
+      "epoch": 0.010583074488969791,
+      "eval_loss": 4.1984477043151855,
+      "eval_runtime": 2.3492,
+      "eval_samples_per_second": 744.941,
+      "eval_steps_per_second": 46.825,
+      "step": 2000
+    },
+    {
+      "epoch": 0.021166148977939583,
+      "grad_norm": 1.406479835510254,
+      "learning_rate": 0.0001,
+      "loss": 4.145,
+      "step": 4000
+    },
+    {
+      "epoch": 0.021166148977939583,
+      "eval_loss": 3.9832916259765625,
+      "eval_runtime": 2.1928,
+      "eval_samples_per_second": 798.069,
+      "eval_steps_per_second": 50.164,
+      "step": 4000
+    },
+    {
+      "epoch": 0.03174922346690937,
+      "grad_norm": 1.4979146718978882,
+      "learning_rate": 0.0001,
+      "loss": 3.9948,
+      "step": 6000
+    },
+    {
+      "epoch": 0.03174922346690937,
+      "eval_loss": 3.872448444366455,
+      "eval_runtime": 2.0908,
+      "eval_samples_per_second": 837.0,
+      "eval_steps_per_second": 52.611,
+      "step": 6000
+    },
+    {
+      "epoch": 0.042332297955879165,
+      "grad_norm": 1.4652293920516968,
+      "learning_rate": 0.0001,
+      "loss": 3.9047,
+      "step": 8000
+    },
+    {
+      "epoch": 0.042332297955879165,
+      "eval_loss": 3.8042104244232178,
+      "eval_runtime": 2.0939,
+      "eval_samples_per_second": 835.759,
+      "eval_steps_per_second": 52.533,
+      "step": 8000
+    },
+    {
+      "epoch": 0.05291537244484895,
+      "grad_norm": 1.523988127708435,
+      "learning_rate": 0.0001,
+      "loss": 3.8428,
+      "step": 10000
+    },
+    {
+      "epoch": 0.05291537244484895,
+      "eval_loss": 3.751901149749756,
+      "eval_runtime": 2.1539,
+      "eval_samples_per_second": 812.471,
+      "eval_steps_per_second": 51.07,
+      "step": 10000
+    },
+    {
+      "epoch": 0.06349844693381874,
+      "grad_norm": 1.6126410961151123,
+      "learning_rate": 0.0001,
+      "loss": 3.7955,
+      "step": 12000
+    },
+    {
+      "epoch": 0.06349844693381874,
+      "eval_loss": 3.71409010887146,
+      "eval_runtime": 2.0662,
+      "eval_samples_per_second": 846.973,
+      "eval_steps_per_second": 53.238,
+      "step": 12000
+    },
+    {
+      "epoch": 0.07408152142278854,
+      "grad_norm": 1.656988263130188,
+      "learning_rate": 0.0001,
+      "loss": 3.7557,
+      "step": 14000
+    },
+    {
+      "epoch": 0.07408152142278854,
+      "eval_loss": 3.6855504512786865,
+      "eval_runtime": 2.0523,
+      "eval_samples_per_second": 852.711,
+      "eval_steps_per_second": 53.599,
+      "step": 14000
+    },
+    {
+      "epoch": 0.08466459591175833,
+      "grad_norm": 1.6087511777877808,
+      "learning_rate": 0.0001,
+      "loss": 3.725,
+      "step": 16000
+    },
+    {
+      "epoch": 0.08466459591175833,
+      "eval_loss": 3.658583164215088,
+      "eval_runtime": 2.1352,
+      "eval_samples_per_second": 819.607,
+      "eval_steps_per_second": 51.518,
+      "step": 16000
+    },
+    {
+      "epoch": 0.09524767040072811,
+      "grad_norm": 1.6472262144088745,
+      "learning_rate": 0.0001,
+      "loss": 3.6947,
+      "step": 18000
+    },
+    {
+      "epoch": 0.09524767040072811,
+      "eval_loss": 3.638516664505005,
+      "eval_runtime": 95.757,
+      "eval_samples_per_second": 18.275,
+      "eval_steps_per_second": 1.149,
+      "step": 18000
+    },
+    {
+      "epoch": 0.1058307448896979,
+      "grad_norm": 1.6628919839859009,
+      "learning_rate": 0.0001,
+      "loss": 3.6726,
+      "step": 20000
+    },
+    {
+      "epoch": 0.1058307448896979,
+      "eval_loss": 3.622284412384033,
+      "eval_runtime": 2.1518,
+      "eval_samples_per_second": 813.27,
+      "eval_steps_per_second": 51.12,
+      "step": 20000
+    },
+    {
+      "epoch": 0.1164138193786677,
+      "grad_norm": 1.5287901163101196,
+      "learning_rate": 0.0001,
+      "loss": 3.6556,
+      "step": 22000
+    },
+    {
+      "epoch": 0.1164138193786677,
+      "eval_loss": 3.6024463176727295,
+      "eval_runtime": 2.0124,
+      "eval_samples_per_second": 869.594,
+      "eval_steps_per_second": 54.66,
+      "step": 22000
+    },
+    {
+      "epoch": 0.1269968938676375,
+      "grad_norm": 1.6352204084396362,
+      "learning_rate": 0.0001,
+      "loss": 3.6309,
+      "step": 24000
+    },
+    {
+      "epoch": 0.1269968938676375,
+      "eval_loss": 3.5923054218292236,
+      "eval_runtime": 2.0331,
+      "eval_samples_per_second": 860.735,
+      "eval_steps_per_second": 54.103,
+      "step": 24000
+    },
+    {
+      "epoch": 0.13757996835660727,
+      "grad_norm": 1.5545786619186401,
+      "learning_rate": 0.0001,
+      "loss": 3.619,
+      "step": 26000
+    },
+    {
+      "epoch": 0.13757996835660727,
+      "eval_loss": 3.5808773040771484,
+      "eval_runtime": 2.1131,
+      "eval_samples_per_second": 828.175,
+      "eval_steps_per_second": 52.057,
+      "step": 26000
+    },
+    {
+      "epoch": 0.14816304284557708,
+      "grad_norm": 1.6340019702911377,
+      "learning_rate": 0.0001,
+      "loss": 3.6045,
+      "step": 28000
+    },
+    {
+      "epoch": 0.14816304284557708,
+      "eval_loss": 3.5688397884368896,
+      "eval_runtime": 2.0238,
+      "eval_samples_per_second": 864.697,
+      "eval_steps_per_second": 54.352,
+      "step": 28000
+    },
+    {
+      "epoch": 0.15874611733454685,
+      "grad_norm": 1.6332552433013916,
+      "learning_rate": 0.0001,
+      "loss": 3.5854,
+      "step": 30000
+    },
+    {
+      "epoch": 0.15874611733454685,
+      "eval_loss": 3.5667507648468018,
+      "eval_runtime": 1.9535,
+      "eval_samples_per_second": 895.836,
+      "eval_steps_per_second": 56.31,
+      "step": 30000
+    },
+    {
+      "epoch": 0.16932919182351666,
+      "grad_norm": 1.6340084075927734,
+      "learning_rate": 0.0001,
+      "loss": 3.5793,
+      "step": 32000
+    },
+    {
+      "epoch": 0.16932919182351666,
+      "eval_loss": 3.557473659515381,
+      "eval_runtime": 2.0124,
+      "eval_samples_per_second": 869.611,
+      "eval_steps_per_second": 54.661,
+      "step": 32000
+    },
+    {
+      "epoch": 0.17991226631248644,
+      "grad_norm": 1.5192989110946655,
+      "learning_rate": 0.0001,
+      "loss": 3.5685,
+      "step": 34000
+    },
+    {
+      "epoch": 0.17991226631248644,
+      "eval_loss": 3.545267343521118,
+      "eval_runtime": 1.995,
+      "eval_samples_per_second": 877.2,
+      "eval_steps_per_second": 55.138,
+      "step": 34000
+    },
+    {
+      "epoch": 0.19049534080145622,
+      "grad_norm": 1.5996226072311401,
+      "learning_rate": 0.0001,
+      "loss": 3.556,
+      "step": 36000
+    },
+    {
+      "epoch": 0.19049534080145622,
+      "eval_loss": 3.540522336959839,
+      "eval_runtime": 1.9783,
+      "eval_samples_per_second": 884.601,
+      "eval_steps_per_second": 55.603,
+      "step": 36000
+    },
+    {
+      "epoch": 0.20107841529042603,
+      "grad_norm": 1.5821503400802612,
+      "learning_rate": 0.0001,
+      "loss": 3.5475,
+      "step": 38000
+    },
+    {
+      "epoch": 0.20107841529042603,
+      "eval_loss": 3.533203125,
+      "eval_runtime": 1.9335,
+      "eval_samples_per_second": 905.095,
+      "eval_steps_per_second": 56.892,
+      "step": 38000
+    },
+    {
+      "epoch": 0.2116614897793958,
+      "grad_norm": 1.63215970993042,
+      "learning_rate": 0.0001,
+      "loss": 3.5358,
+      "step": 40000
+    },
+    {
+      "epoch": 0.2116614897793958,
+      "eval_loss": 3.5292651653289795,
+      "eval_runtime": 2.0497,
+      "eval_samples_per_second": 853.794,
+      "eval_steps_per_second": 53.667,
+      "step": 40000
+    },
+    {
+      "epoch": 0.2222445642683656,
+      "grad_norm": 1.6347813606262207,
+      "learning_rate": 0.0001,
+      "loss": 3.5305,
+      "step": 42000
+    },
+    {
+      "epoch": 0.2222445642683656,
+      "eval_loss": 3.5237081050872803,
+      "eval_runtime": 2.0264,
+      "eval_samples_per_second": 863.619,
+      "eval_steps_per_second": 54.285,
+      "step": 42000
+    },
+    {
+      "epoch": 0.2328276387573354,
+      "grad_norm": 1.617659091949463,
+      "learning_rate": 0.0001,
+      "loss": 3.5257,
+      "step": 44000
+    },
+    {
+      "epoch": 0.2328276387573354,
+      "eval_loss": 3.518353223800659,
+      "eval_runtime": 2.0166,
+      "eval_samples_per_second": 867.799,
+      "eval_steps_per_second": 54.547,
+      "step": 44000
+    },
+    {
+      "epoch": 0.24341071324630517,
+      "grad_norm": 1.5734567642211914,
+      "learning_rate": 0.0001,
+      "loss": 3.5122,
+      "step": 46000
+    },
+    {
+      "epoch": 0.24341071324630517,
+      "eval_loss": 3.516266345977783,
+      "eval_runtime": 2.0078,
+      "eval_samples_per_second": 871.618,
+      "eval_steps_per_second": 54.787,
+      "step": 46000
+    },
+    {
+      "epoch": 0.253993787735275,
+      "grad_norm": 1.559778094291687,
+      "learning_rate": 0.0001,
+      "loss": 3.5099,
+      "step": 48000
+    },
+    {
+      "epoch": 0.253993787735275,
+      "eval_loss": 3.509239435195923,
+      "eval_runtime": 2.0647,
+      "eval_samples_per_second": 847.6,
+      "eval_steps_per_second": 53.278,
+      "step": 48000
+    },
+    {
+      "epoch": 0.26457686222424476,
+      "grad_norm": 1.4903326034545898,
+      "learning_rate": 0.0001,
+      "loss": 3.503,
+      "step": 50000
+    },
+    {
+      "epoch": 0.26457686222424476,
+      "eval_loss": 3.509716510772705,
+      "eval_runtime": 2.0953,
+      "eval_samples_per_second": 835.185,
+      "eval_steps_per_second": 52.497,
+      "step": 50000
+    },
+    {
+      "epoch": 0.27515993671321454,
+      "grad_norm": 1.5261775255203247,
+      "learning_rate": 0.0001,
+      "loss": 3.494,
+      "step": 52000
+    },
+    {
+      "epoch": 0.27515993671321454,
+      "eval_loss": 3.505798578262329,
+      "eval_runtime": 2.0237,
+      "eval_samples_per_second": 864.765,
+      "eval_steps_per_second": 54.357,
+      "step": 52000
+    },
+    {
+      "epoch": 0.28574301120218437,
+      "grad_norm": 1.5844075679779053,
+      "learning_rate": 0.0001,
+      "loss": 3.4923,
+      "step": 54000
+    },
+    {
+      "epoch": 0.28574301120218437,
+      "eval_loss": 3.5029475688934326,
+      "eval_runtime": 2.0259,
+      "eval_samples_per_second": 863.817,
+      "eval_steps_per_second": 54.297,
+      "step": 54000
+    },
+    {
+      "epoch": 0.29632608569115415,
+      "grad_norm": 1.5451163053512573,
+      "learning_rate": 0.0001,
+      "loss": 3.4839,
+      "step": 56000
+    },
+    {
+      "epoch": 0.29632608569115415,
+      "eval_loss": 3.4993185997009277,
+      "eval_runtime": 2.0088,
+      "eval_samples_per_second": 871.183,
+      "eval_steps_per_second": 54.76,
+      "step": 56000
+    },
+    {
+      "epoch": 0.30690916018012393,
+      "grad_norm": 1.5733637809753418,
+      "learning_rate": 0.0001,
+      "loss": 3.4794,
+      "step": 58000
+    },
+    {
+      "epoch": 0.30690916018012393,
+      "eval_loss": 3.4978885650634766,
+      "eval_runtime": 1.9995,
+      "eval_samples_per_second": 875.202,
+      "eval_steps_per_second": 55.013,
+      "step": 58000
+    }
+  ],
+  "logging_steps": 2000,
+  "max_steps": 188981,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 1,
+  "save_steps": 2000,
+  "stateful_callbacks": {
+    "EarlyStoppingCallback": {
+      "args": {
+        "early_stopping_patience": 3,
+        "early_stopping_threshold": 0.001
+      },
+      "attributes": {
+        "early_stopping_patience_counter": 0
+      }
+    },
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 565419366014976.0,
+  "train_batch_size": 16,
+  "trial_name": null,
+  "trial_params": null
+}

checkpoint-58000/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0c0a2b1811aeefb927284bf3c5c8de0a89d0ad3fd57a0b61184713eb6f012eba
+size 5432