Training in progress, step 44000, checkpoint

checkpoint-44000/config.json

@@ -0,0 +1,28 @@
+{
+  "architectures": [
+    "RobertaForMaskedLM"
+  ],
+  "attention_probs_dropout_prob": 0.1,
+  "bos_token_id": 0,
+  "classifier_dropout": null,
+  "eos_token_id": 1,
+  "hidden_act": "gelu",
+  "hidden_dropout_prob": 0.1,
+  "hidden_size": 256,
+  "initializer_range": 0.02,
+  "intermediate_size": 2048,
+  "layer_norm_eps": 1e-05,
+  "mask_token_id": 4,
+  "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-44000/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b58861fb6cb5df1ff6b1be7e7e47b92a208bd5bc0cdf317f631c3f7e155cc3e2
+size 59702184

checkpoint-44000/optimizer.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4f2b0e226ed4e813d53909bb3acac74504b2d6b0d17d7b6ac6ad0d25e19a9448
+size 119488058

checkpoint-44000/rng_state.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:700716dadfcb255a5bb9e0b8cee6ddec4d2e9aefc26f377da6e328784e0d23a3
+size 14308

checkpoint-44000/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4e53fb67ec7b3857f368c463cd4917682c4e26bcf666e32939458e19db2f6e5d
+size 1000

checkpoint-44000/special_tokens_map.json

@@ -0,0 +1,34 @@
+{
+  "additional_special_tokens": [
+    "<|endoftext|>"
+  ],
+  "bos_token": {
+    "content": "<s>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "eos_token": {
+    "content": "</s>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "mask_token": {
+    "content": "<mask>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": "</s>",
+  "unk_token": {
+    "content": "<unk>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

checkpoint-44000/tokenizer_config.json

@@ -0,0 +1,63 @@
+{
+  "added_tokens_decoder": {
+    "0": {
+      "content": "<s>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "1": {
+      "content": "</s>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "2": {
+      "content": "<pad>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "3": {
+      "content": "<unk>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "4": {
+      "content": "<mask>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "5": {
+      "content": "<|endoftext|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "additional_special_tokens": [
+    "<|endoftext|>"
+  ],
+  "bos_token": "<s>",
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "</s>",
+  "mask_token": "<mask>",
+  "model_max_length": 128,
+  "pad_token": "</s>",
+  "tokenizer_class": "PreTrainedTokenizerFast",
+  "unk_token": "<unk>"
+}

checkpoint-44000/trainer_state.json

@@ -0,0 +1,286 @@
+{
+  "best_metric": 3.9708411693573,
+  "best_model_checkpoint": "/home/p318482/babyLM_controlled/models_trained/de_mlm/wikipedia_13/checkpoint-44000",
+  "epoch": 32.24624404543789,
+  "eval_steps": 2000,
+  "global_step": 44000,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 1.4657383657017222,
+      "eval_loss": 8.087221145629883,
+      "eval_runtime": 2.215,
+      "eval_samples_per_second": 1035.232,
+      "eval_steps_per_second": 65.012,
+      "step": 2000
+    },
+    {
+      "epoch": 2.9314767314034444,
+      "grad_norm": 0.9454871416091919,
+      "learning_rate": 1e-05,
+      "loss": 8.1326,
+      "step": 4000
+    },
+    {
+      "epoch": 2.9314767314034444,
+      "eval_loss": 7.397463798522949,
+      "eval_runtime": 2.2812,
+      "eval_samples_per_second": 1005.158,
+      "eval_steps_per_second": 63.124,
+      "step": 4000
+    },
+    {
+      "epoch": 4.397215097105167,
+      "eval_loss": 7.2718353271484375,
+      "eval_runtime": 2.3459,
+      "eval_samples_per_second": 977.446,
+      "eval_steps_per_second": 61.383,
+      "step": 6000
+    },
+    {
+      "epoch": 5.862953462806889,
+      "grad_norm": 1.4085547924041748,
+      "learning_rate": 2e-05,
+      "loss": 7.2846,
+      "step": 8000
+    },
+    {
+      "epoch": 5.862953462806889,
+      "eval_loss": 7.183506965637207,
+      "eval_runtime": 2.2787,
+      "eval_samples_per_second": 1006.271,
+      "eval_steps_per_second": 63.194,
+      "step": 8000
+    },
+    {
+      "epoch": 7.328691828508611,
+      "eval_loss": 7.09923791885376,
+      "eval_runtime": 2.1833,
+      "eval_samples_per_second": 1050.26,
+      "eval_steps_per_second": 65.956,
+      "step": 10000
+    },
+    {
+      "epoch": 8.794430194210333,
+      "grad_norm": 1.4775667190551758,
+      "learning_rate": 2.9995e-05,
+      "loss": 7.1078,
+      "step": 12000
+    },
+    {
+      "epoch": 8.794430194210333,
+      "eval_loss": 7.033052444458008,
+      "eval_runtime": 2.2436,
+      "eval_samples_per_second": 1022.003,
+      "eval_steps_per_second": 64.182,
+      "step": 12000
+    },
+    {
+      "epoch": 10.260168559912056,
+      "eval_loss": 6.949419975280762,
+      "eval_runtime": 2.3034,
+      "eval_samples_per_second": 995.501,
+      "eval_steps_per_second": 62.517,
+      "step": 14000
+    },
+    {
+      "epoch": 11.725906925613778,
+      "grad_norm": 2.364133358001709,
+      "learning_rate": 3.999e-05,
+      "loss": 6.942,
+      "step": 16000
+    },
+    {
+      "epoch": 11.725906925613778,
+      "eval_loss": 6.889880180358887,
+      "eval_runtime": 2.3275,
+      "eval_samples_per_second": 985.162,
+      "eval_steps_per_second": 61.868,
+      "step": 16000
+    },
+    {
+      "epoch": 13.191645291315501,
+      "eval_loss": 6.782171726226807,
+      "eval_runtime": 2.2331,
+      "eval_samples_per_second": 1026.812,
+      "eval_steps_per_second": 64.484,
+      "step": 18000
+    },
+    {
+      "epoch": 14.657383657017222,
+      "grad_norm": 2.5763492584228516,
+      "learning_rate": 4.9985e-05,
+      "loss": 6.7676,
+      "step": 20000
+    },
+    {
+      "epoch": 14.657383657017222,
+      "eval_loss": 6.718458652496338,
+      "eval_runtime": 2.2092,
+      "eval_samples_per_second": 1037.91,
+      "eval_steps_per_second": 65.181,
+      "step": 20000
+    },
+    {
+      "epoch": 16.123122022718945,
+      "eval_loss": 6.653624057769775,
+      "eval_runtime": 2.33,
+      "eval_samples_per_second": 984.135,
+      "eval_steps_per_second": 61.804,
+      "step": 22000
+    },
+    {
+      "epoch": 17.588860388420667,
+      "grad_norm": 4.001818656921387,
+      "learning_rate": 5.9980000000000005e-05,
+      "loss": 6.5959,
+      "step": 24000
+    },
+    {
+      "epoch": 17.588860388420667,
+      "eval_loss": 6.543092727661133,
+      "eval_runtime": 2.3092,
+      "eval_samples_per_second": 992.981,
+      "eval_steps_per_second": 62.359,
+      "step": 24000
+    },
+    {
+      "epoch": 19.054598754122388,
+      "eval_loss": 6.392500400543213,
+      "eval_runtime": 2.4925,
+      "eval_samples_per_second": 919.956,
+      "eval_steps_per_second": 57.773,
+      "step": 26000
+    },
+    {
+      "epoch": 20.520337119824113,
+      "grad_norm": 4.359928607940674,
+      "learning_rate": 6.99775e-05,
+      "loss": 6.3624,
+      "step": 28000
+    },
+    {
+      "epoch": 20.520337119824113,
+      "eval_loss": 6.211884021759033,
+      "eval_runtime": 2.3499,
+      "eval_samples_per_second": 975.797,
+      "eval_steps_per_second": 61.28,
+      "step": 28000
+    },
+    {
+      "epoch": 21.986075485525834,
+      "eval_loss": 5.952591896057129,
+      "eval_runtime": 2.3365,
+      "eval_samples_per_second": 981.37,
+      "eval_steps_per_second": 61.63,
+      "step": 30000
+    },
+    {
+      "epoch": 23.451813851227556,
+      "grad_norm": 5.47481632232666,
+      "learning_rate": 7.997250000000001e-05,
+      "loss": 5.9309,
+      "step": 32000
+    },
+    {
+      "epoch": 23.451813851227556,
+      "eval_loss": 5.4161834716796875,
+      "eval_runtime": 2.2165,
+      "eval_samples_per_second": 1034.515,
+      "eval_steps_per_second": 64.967,
+      "step": 32000
+    },
+    {
+      "epoch": 24.917552216929277,
+      "eval_loss": 5.0255045890808105,
+      "eval_runtime": 2.3492,
+      "eval_samples_per_second": 976.095,
+      "eval_steps_per_second": 61.299,
+      "step": 34000
+    },
+    {
+      "epoch": 26.383290582631002,
+      "grad_norm": 5.85042667388916,
+      "learning_rate": 8.996750000000001e-05,
+      "loss": 5.0575,
+      "step": 36000
+    },
+    {
+      "epoch": 26.383290582631002,
+      "eval_loss": 4.76804256439209,
+      "eval_runtime": 2.2482,
+      "eval_samples_per_second": 1019.914,
+      "eval_steps_per_second": 64.05,
+      "step": 36000
+    },
+    {
+      "epoch": 27.849028948332723,
+      "eval_loss": 4.502030372619629,
+      "eval_runtime": 2.3066,
+      "eval_samples_per_second": 994.084,
+      "eval_steps_per_second": 62.428,
+      "step": 38000
+    },
+    {
+      "epoch": 29.314767314034444,
+      "grad_norm": 4.6283698081970215,
+      "learning_rate": 9.99625e-05,
+      "loss": 4.5282,
+      "step": 40000
+    },
+    {
+      "epoch": 29.314767314034444,
+      "eval_loss": 4.321413993835449,
+      "eval_runtime": 2.2564,
+      "eval_samples_per_second": 1016.223,
+      "eval_steps_per_second": 63.819,
+      "step": 40000
+    },
+    {
+      "epoch": 30.780505679736166,
+      "eval_loss": 4.131194114685059,
+      "eval_runtime": 2.27,
+      "eval_samples_per_second": 1010.131,
+      "eval_steps_per_second": 63.436,
+      "step": 42000
+    },
+    {
+      "epoch": 32.24624404543789,
+      "grad_norm": 4.427154064178467,
+      "learning_rate": 9.336166666666667e-05,
+      "loss": 4.1335,
+      "step": 44000
+    },
+    {
+      "epoch": 32.24624404543789,
+      "eval_loss": 3.9708411693573,
+      "eval_runtime": 2.1314,
+      "eval_samples_per_second": 1075.829,
+      "eval_steps_per_second": 67.562,
+      "step": 44000
+    }
+  ],
+  "logging_steps": 4000,
+  "max_steps": 100000,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 74,
+  "save_steps": 4000,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 1.3723342297104384e+16,
+  "train_batch_size": 16,
+  "trial_name": null,
+  "trial_params": null
+}

checkpoint-44000/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:18d5335e549995a666047ad0aa350511260e15302e12c49cfc12029ab6ff3c4a
+size 5368