Training in progress, step 56000, checkpoint

checkpoint-56000/config.json

@@ -0,0 +1,32 @@
+{
+  "activation_function": "gelu_new",
+  "architectures": [
+    "GPT2LMHeadModel"
+  ],
+  "attn_pdrop": 0.1,
+  "bos_token_id": 0,
+  "embd_pdrop": 0.1,
+  "eos_token_id": 1,
+  "initializer_range": 0.02,
+  "layer_norm_epsilon": 1e-05,
+  "model_type": "gpt2",
+  "n_ctx": 512,
+  "n_embd": 256,
+  "n_head": 8,
+  "n_inner": 2048,
+  "n_layer": 8,
+  "n_positions": 512,
+  "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,
+  "torch_dtype": "float32",
+  "transformers_version": "4.45.2",
+  "use_cache": true,
+  "vocab_size": 8192
+}

checkpoint-56000/generation_config.json

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

checkpoint-56000/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:720f533079fc6130ce14b30b8eafdb5e4076469e1faa3ba38a4e72947d565b0b
+size 51007160

checkpoint-56000/optimizer.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c8d601afe35edd64872bb14c04719f9300b470651e7e555f58f816d084fa31c3
+size 102078202

checkpoint-56000/rng_state.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b6504f403ec0124563e90a8d77beb4d07baefb7092c70b6d168b7c081e93523f
+size 14308

checkpoint-56000/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1f1181b919b10206263db791048c88fc007cb1b4a9fe395db90858e73bef9c62
+size 1000

checkpoint-56000/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-56000/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-56000/trainer_state.json

@@ -0,0 +1,355 @@
+{
+  "best_metric": 4.237761974334717,
+  "best_model_checkpoint": "/home/p318482/babyLM_controlled/models_trained/en_clm/wikipedia_30/checkpoint-56000",
+  "epoch": 29.45043386799895,
+  "eval_steps": 2000,
+  "global_step": 56000,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 1.0518012095713911,
+      "eval_loss": 7.518920421600342,
+      "eval_runtime": 2.2399,
+      "eval_samples_per_second": 1422.808,
+      "eval_steps_per_second": 89.288,
+      "step": 2000
+    },
+    {
+      "epoch": 2.1036024191427822,
+      "grad_norm": 1.2947496175765991,
+      "learning_rate": 1e-05,
+      "loss": 7.6141,
+      "step": 4000
+    },
+    {
+      "epoch": 2.1036024191427822,
+      "eval_loss": 6.568221092224121,
+      "eval_runtime": 2.1236,
+      "eval_samples_per_second": 1500.783,
+      "eval_steps_per_second": 94.182,
+      "step": 4000
+    },
+    {
+      "epoch": 3.155403628714173,
+      "eval_loss": 6.16925573348999,
+      "eval_runtime": 2.1624,
+      "eval_samples_per_second": 1473.825,
+      "eval_steps_per_second": 92.49,
+      "step": 6000
+    },
+    {
+      "epoch": 4.2072048382855645,
+      "grad_norm": 2.330080270767212,
+      "learning_rate": 2e-05,
+      "loss": 6.2396,
+      "step": 8000
+    },
+    {
+      "epoch": 4.2072048382855645,
+      "eval_loss": 5.901575088500977,
+      "eval_runtime": 2.1326,
+      "eval_samples_per_second": 1494.388,
+      "eval_steps_per_second": 93.78,
+      "step": 8000
+    },
+    {
+      "epoch": 5.259006047856955,
+      "eval_loss": 5.67369270324707,
+      "eval_runtime": 2.1364,
+      "eval_samples_per_second": 1491.788,
+      "eval_steps_per_second": 93.617,
+      "step": 10000
+    },
+    {
+      "epoch": 6.310807257428346,
+      "grad_norm": 2.565220594406128,
+      "learning_rate": 2.99925e-05,
+      "loss": 5.7217,
+      "step": 12000
+    },
+    {
+      "epoch": 6.310807257428346,
+      "eval_loss": 5.460005760192871,
+      "eval_runtime": 2.1835,
+      "eval_samples_per_second": 1459.563,
+      "eval_steps_per_second": 91.595,
+      "step": 12000
+    },
+    {
+      "epoch": 7.362608466999737,
+      "eval_loss": 5.272489547729492,
+      "eval_runtime": 2.1833,
+      "eval_samples_per_second": 1459.691,
+      "eval_steps_per_second": 91.603,
+      "step": 14000
+    },
+    {
+      "epoch": 8.414409676571129,
+      "grad_norm": 2.7334115505218506,
+      "learning_rate": 3.99875e-05,
+      "loss": 5.3064,
+      "step": 16000
+    },
+    {
+      "epoch": 8.414409676571129,
+      "eval_loss": 5.1074652671813965,
+      "eval_runtime": 2.2402,
+      "eval_samples_per_second": 1422.652,
+      "eval_steps_per_second": 89.278,
+      "step": 16000
+    },
+    {
+      "epoch": 9.46621088614252,
+      "eval_loss": 4.9716596603393555,
+      "eval_runtime": 2.1736,
+      "eval_samples_per_second": 1466.258,
+      "eval_steps_per_second": 92.015,
+      "step": 18000
+    },
+    {
+      "epoch": 10.51801209571391,
+      "grad_norm": 2.6722376346588135,
+      "learning_rate": 4.9985e-05,
+      "loss": 4.9744,
+      "step": 20000
+    },
+    {
+      "epoch": 10.51801209571391,
+      "eval_loss": 4.864916801452637,
+      "eval_runtime": 2.2031,
+      "eval_samples_per_second": 1446.585,
+      "eval_steps_per_second": 90.78,
+      "step": 20000
+    },
+    {
+      "epoch": 11.569813305285301,
+      "eval_loss": 4.763542652130127,
+      "eval_runtime": 2.4382,
+      "eval_samples_per_second": 1307.131,
+      "eval_steps_per_second": 82.029,
+      "step": 22000
+    },
+    {
+      "epoch": 12.621614514856692,
+      "grad_norm": 2.4539408683776855,
+      "learning_rate": 5.9985e-05,
+      "loss": 4.7273,
+      "step": 24000
+    },
+    {
+      "epoch": 12.621614514856692,
+      "eval_loss": 4.683297634124756,
+      "eval_runtime": 2.452,
+      "eval_samples_per_second": 1299.733,
+      "eval_steps_per_second": 81.565,
+      "step": 24000
+    },
+    {
+      "epoch": 13.673415724428082,
+      "eval_loss": 4.621574401855469,
+      "eval_runtime": 2.1825,
+      "eval_samples_per_second": 1460.226,
+      "eval_steps_per_second": 91.636,
+      "step": 26000
+    },
+    {
+      "epoch": 14.725216933999475,
+      "grad_norm": 2.4913346767425537,
+      "learning_rate": 6.998e-05,
+      "loss": 4.5397,
+      "step": 28000
+    },
+    {
+      "epoch": 14.725216933999475,
+      "eval_loss": 4.562623977661133,
+      "eval_runtime": 2.201,
+      "eval_samples_per_second": 1447.965,
+      "eval_steps_per_second": 90.867,
+      "step": 28000
+    },
+    {
+      "epoch": 15.777018143570865,
+      "eval_loss": 4.506168842315674,
+      "eval_runtime": 2.2491,
+      "eval_samples_per_second": 1416.988,
+      "eval_steps_per_second": 88.923,
+      "step": 30000
+    },
+    {
+      "epoch": 16.828819353142258,
+      "grad_norm": 2.2661523818969727,
+      "learning_rate": 7.99775e-05,
+      "loss": 4.3839,
+      "step": 32000
+    },
+    {
+      "epoch": 16.828819353142258,
+      "eval_loss": 4.464449882507324,
+      "eval_runtime": 2.253,
+      "eval_samples_per_second": 1414.55,
+      "eval_steps_per_second": 88.77,
+      "step": 32000
+    },
+    {
+      "epoch": 17.88062056271365,
+      "eval_loss": 4.426318168640137,
+      "eval_runtime": 2.2688,
+      "eval_samples_per_second": 1404.709,
+      "eval_steps_per_second": 88.152,
+      "step": 34000
+    },
+    {
+      "epoch": 18.93242177228504,
+      "grad_norm": 2.226025104522705,
+      "learning_rate": 8.9975e-05,
+      "loss": 4.2529,
+      "step": 36000
+    },
+    {
+      "epoch": 18.93242177228504,
+      "eval_loss": 4.3944478034973145,
+      "eval_runtime": 2.294,
+      "eval_samples_per_second": 1389.273,
+      "eval_steps_per_second": 87.184,
+      "step": 36000
+    },
+    {
+      "epoch": 19.98422298185643,
+      "eval_loss": 4.365135669708252,
+      "eval_runtime": 2.2176,
+      "eval_samples_per_second": 1437.165,
+      "eval_steps_per_second": 90.189,
+      "step": 38000
+    },
+    {
+      "epoch": 21.03602419142782,
+      "grad_norm": 2.1669206619262695,
+      "learning_rate": 9.997e-05,
+      "loss": 4.1409,
+      "step": 40000
+    },
+    {
+      "epoch": 21.03602419142782,
+      "eval_loss": 4.351284503936768,
+      "eval_runtime": 2.1934,
+      "eval_samples_per_second": 1453.028,
+      "eval_steps_per_second": 91.185,
+      "step": 40000
+    },
+    {
+      "epoch": 22.08782540099921,
+      "eval_loss": 4.321634292602539,
+      "eval_runtime": 2.2273,
+      "eval_samples_per_second": 1430.895,
+      "eval_steps_per_second": 89.796,
+      "step": 42000
+    },
+    {
+      "epoch": 23.139626610570602,
+      "grad_norm": 2.0360896587371826,
+      "learning_rate": 9.335333333333333e-05,
+      "loss": 4.0334,
+      "step": 44000
+    },
+    {
+      "epoch": 23.139626610570602,
+      "eval_loss": 4.300807476043701,
+      "eval_runtime": 2.3226,
+      "eval_samples_per_second": 1372.142,
+      "eval_steps_per_second": 86.109,
+      "step": 44000
+    },
+    {
+      "epoch": 24.191427820141993,
+      "eval_loss": 4.279116153717041,
+      "eval_runtime": 2.2177,
+      "eval_samples_per_second": 1437.082,
+      "eval_steps_per_second": 90.184,
+      "step": 46000
+    },
+    {
+      "epoch": 25.243229029713383,
+      "grad_norm": 2.137535572052002,
+      "learning_rate": 8.668833333333334e-05,
+      "loss": 3.9378,
+      "step": 48000
+    },
+    {
+      "epoch": 25.243229029713383,
+      "eval_loss": 4.2718939781188965,
+      "eval_runtime": 2.2019,
+      "eval_samples_per_second": 1447.39,
+      "eval_steps_per_second": 90.831,
+      "step": 48000
+    },
+    {
+      "epoch": 26.295030239284774,
+      "eval_loss": 4.26237678527832,
+      "eval_runtime": 2.2417,
+      "eval_samples_per_second": 1421.698,
+      "eval_steps_per_second": 89.219,
+      "step": 50000
+    },
+    {
+      "epoch": 27.346831448856165,
+      "grad_norm": 2.1131036281585693,
+      "learning_rate": 8.002333333333334e-05,
+      "loss": 3.8595,
+      "step": 52000
+    },
+    {
+      "epoch": 27.346831448856165,
+      "eval_loss": 4.253220558166504,
+      "eval_runtime": 2.2045,
+      "eval_samples_per_second": 1445.692,
+      "eval_steps_per_second": 90.724,
+      "step": 52000
+    },
+    {
+      "epoch": 28.39863265842756,
+      "eval_loss": 4.243188381195068,
+      "eval_runtime": 2.1654,
+      "eval_samples_per_second": 1471.751,
+      "eval_steps_per_second": 92.36,
+      "step": 54000
+    },
+    {
+      "epoch": 29.45043386799895,
+      "grad_norm": 2.1626858711242676,
+      "learning_rate": 7.335833333333335e-05,
+      "loss": 3.7927,
+      "step": 56000
+    },
+    {
+      "epoch": 29.45043386799895,
+      "eval_loss": 4.237761974334717,
+      "eval_runtime": 2.3384,
+      "eval_samples_per_second": 1362.919,
+      "eval_steps_per_second": 85.53,
+      "step": 56000
+    }
+  ],
+  "logging_steps": 4000,
+  "max_steps": 100000,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 53,
+  "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": 1.4476189099416576e+16,
+  "train_batch_size": 16,
+  "trial_name": null,
+  "trial_params": null
+}

checkpoint-56000/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ab68434ec7156d3a63e4783f824871a48d62c7c2fdcc831fdd5471c5f4aec7fe
+size 5368