Training in progress, step 20000, checkpoint

checkpoint-20000/config.json

@@ -0,0 +1,28 @@
+{
+  "architectures": [
+    "RobertaForMaskedLM"
+  ],
+  "attention_probs_dropout_prob": 0.1,
+  "bos_token_id": 3,
+  "classifier_dropout": null,
+  "eos_token_id": 4,
+  "hidden_act": "gelu",
+  "hidden_dropout_prob": 0.1,
+  "hidden_size": 256,
+  "initializer_range": 0.02,
+  "intermediate_size": 1024,
+  "layer_norm_eps": 1e-05,
+  "mask_token_id": 6,
+  "max_position_embeddings": 512,
+  "model_type": "roberta",
+  "num_attention_heads": 8,
+  "num_hidden_layers": 8,
+  "pad_token_id": 1,
+  "position_embedding_type": "absolute",
+  "tie_word_embeddings": false,
+  "torch_dtype": "float32",
+  "transformers_version": "4.45.2",
+  "type_vocab_size": 2,
+  "use_cache": true,
+  "vocab_size": 8192
+}

checkpoint-20000/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bd0b237e29d150a0a966b3f7df95d1ec5819f124e66ec822de43aa9b8584ecda
+size 42892200

checkpoint-20000/optimizer.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1e5618794eaf3821ecd10f8fa0e73cb2261c80a100483f9256d98f13d6e4522e
+size 85868090

checkpoint-20000/rng_state.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0947f629569205c94659a9f175a03e9b7cea0e3d18976a425ff71fd79ef97fa0
+size 14308

checkpoint-20000/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9b829d85dc7da6821d0e712c6412224f53dac0b2594c55ed2b4bfc9fa0b522b0
+size 1000

checkpoint-20000/special_tokens_map.json

@@ -0,0 +1,6 @@
+{
+  "bos_token": "[PAD]",
+  "eos_token": "[EOS]",
+  "mask_token": "[MASK]",
+  "pad_token": "[CLS]"
+}

checkpoint-20000/tokenizer_config.json

@@ -0,0 +1,67 @@
+{
+  "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
+    },
+    "6": {
+      "content": "[MASK]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "bos_token": "[PAD]",
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "[EOS]",
+  "mask_token": "[MASK]",
+  "model_max_length": 128,
+  "pad_token": "[CLS]",
+  "tokenizer_class": "PreTrainedTokenizerFast"
+}

checkpoint-20000/trainer_state.json

@@ -0,0 +1,157 @@
+{
+  "best_metric": 3.362109899520874,
+  "best_model_checkpoint": "/home/p318482/babyLM_controlled/models_trained_recent/mlm_unmasking/en_mlm_new/childes_mlm_unmasking_sent_13/checkpoint-20000",
+  "epoch": 0.7406995907634761,
+  "eval_steps": 2000,
+  "global_step": 20000,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.07406995907634761,
+      "eval_loss": 5.4637370109558105,
+      "eval_runtime": 24.2102,
+      "eval_samples_per_second": 1237.041,
+      "eval_steps_per_second": 77.323,
+      "step": 2000
+    },
+    {
+      "epoch": 0.14813991815269523,
+      "grad_norm": 4.785717487335205,
+      "learning_rate": 4e-05,
+      "loss": 6.1476,
+      "step": 4000
+    },
+    {
+      "epoch": 0.14813991815269523,
+      "eval_loss": 4.718623161315918,
+      "eval_runtime": 24.2123,
+      "eval_samples_per_second": 1236.934,
+      "eval_steps_per_second": 77.316,
+      "step": 4000
+    },
+    {
+      "epoch": 0.22220987722904284,
+      "eval_loss": 4.352034091949463,
+      "eval_runtime": 24.1788,
+      "eval_samples_per_second": 1238.646,
+      "eval_steps_per_second": 77.423,
+      "step": 6000
+    },
+    {
+      "epoch": 0.29627983630539045,
+      "grad_norm": 7.057693004608154,
+      "learning_rate": 8e-05,
+      "loss": 4.4471,
+      "step": 8000
+    },
+    {
+      "epoch": 0.29627983630539045,
+      "eval_loss": 3.99312686920166,
+      "eval_runtime": 24.1035,
+      "eval_samples_per_second": 1242.518,
+      "eval_steps_per_second": 77.665,
+      "step": 8000
+    },
+    {
+      "epoch": 0.37034979538173807,
+      "eval_loss": 3.763000965118408,
+      "eval_runtime": 24.2844,
+      "eval_samples_per_second": 1233.262,
+      "eval_steps_per_second": 77.087,
+      "step": 10000
+    },
+    {
+      "epoch": 0.4444197544580857,
+      "grad_norm": 8.325577735900879,
+      "learning_rate": 0.00012,
+      "loss": 3.8939,
+      "step": 12000
+    },
+    {
+      "epoch": 0.4444197544580857,
+      "eval_loss": 3.6284842491149902,
+      "eval_runtime": 24.245,
+      "eval_samples_per_second": 1235.265,
+      "eval_steps_per_second": 77.212,
+      "step": 12000
+    },
+    {
+      "epoch": 0.5184897135344333,
+      "eval_loss": 3.4887945652008057,
+      "eval_runtime": 24.2373,
+      "eval_samples_per_second": 1235.655,
+      "eval_steps_per_second": 77.236,
+      "step": 14000
+    },
+    {
+      "epoch": 0.5925596726107809,
+      "grad_norm": 10.0059232711792,
+      "learning_rate": 0.00016,
+      "loss": 3.6013,
+      "step": 16000
+    },
+    {
+      "epoch": 0.5925596726107809,
+      "eval_loss": 3.443763256072998,
+      "eval_runtime": 24.5656,
+      "eval_samples_per_second": 1219.141,
+      "eval_steps_per_second": 76.204,
+      "step": 16000
+    },
+    {
+      "epoch": 0.6666296316871285,
+      "eval_loss": 3.3501055240631104,
+      "eval_runtime": 24.1675,
+      "eval_samples_per_second": 1239.227,
+      "eval_steps_per_second": 77.459,
+      "step": 18000
+    },
+    {
+      "epoch": 0.7406995907634761,
+      "grad_norm": 5.149896621704102,
+      "learning_rate": 0.0002,
+      "loss": 3.4806,
+      "step": 20000
+    },
+    {
+      "epoch": 0.7406995907634761,
+      "eval_loss": 3.362109899520874,
+      "eval_runtime": 24.3668,
+      "eval_samples_per_second": 1229.089,
+      "eval_steps_per_second": 76.826,
+      "step": 20000
+    }
+  ],
+  "logging_steps": 4000,
+  "max_steps": 400000,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 15,
+  "save_steps": 4000,
+  "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": 4173122764800000.0,
+  "train_batch_size": 16,
+  "trial_name": null,
+  "trial_params": null
+}

checkpoint-20000/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9dd67d3cacfd6861c6144cee0155c16fcc5b69dd58f8d3c81877d5095c915a2b
+size 5432