Training in progress, step 44000, checkpoint

checkpoint-44000/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-44000/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-44000/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a4954eb5e958316f0a26322389d76b01d49a20f02080202418ce763477464fc1
+size 51007160

checkpoint-44000/optimizer.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:dc36bd5c94e7764132921eefbc5c1d40ae03c81cf1e66007ec1f62883dc944a5
+size 102078202

checkpoint-44000/rng_state.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d7888343c0cb40489e247b37a65446402d8768ecda76516f9a461543c999ed29
+size 14244

checkpoint-44000/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:220eb48d1eab650aa4cfc612d14cb46a30cb04a376bf78cd0902a0f3d8987146
+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": 3.9599227905273438,
+  "best_model_checkpoint": "/home/p318482/babyLM_controlled/models_trained_last/de_clm/de_childes_42/checkpoint-44000",
+  "epoch": 33.04543747653023,
+  "eval_steps": 2000,
+  "global_step": 44000,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 1.5020653398422832,
+      "eval_loss": 7.03363037109375,
+      "eval_runtime": 3.6768,
+      "eval_samples_per_second": 902.97,
+      "eval_steps_per_second": 56.572,
+      "step": 2000
+    },
+    {
+      "epoch": 3.0041306796845664,
+      "grad_norm": 1.6605541706085205,
+      "learning_rate": 1e-05,
+      "loss": 6.9417,
+      "step": 4000
+    },
+    {
+      "epoch": 3.0041306796845664,
+      "eval_loss": 5.811168670654297,
+      "eval_runtime": 3.7246,
+      "eval_samples_per_second": 891.378,
+      "eval_steps_per_second": 55.845,
+      "step": 4000
+    },
+    {
+      "epoch": 4.50619601952685,
+      "eval_loss": 5.458603382110596,
+      "eval_runtime": 3.8082,
+      "eval_samples_per_second": 871.804,
+      "eval_steps_per_second": 54.619,
+      "step": 6000
+    },
+    {
+      "epoch": 6.008261359369133,
+      "grad_norm": 2.27949595451355,
+      "learning_rate": 2e-05,
+      "loss": 5.2077,
+      "step": 8000
+    },
+    {
+      "epoch": 6.008261359369133,
+      "eval_loss": 5.171741008758545,
+      "eval_runtime": 3.8003,
+      "eval_samples_per_second": 873.606,
+      "eval_steps_per_second": 54.732,
+      "step": 8000
+    },
+    {
+      "epoch": 7.510326699211416,
+      "eval_loss": 4.9582695960998535,
+      "eval_runtime": 3.7933,
+      "eval_samples_per_second": 875.227,
+      "eval_steps_per_second": 54.834,
+      "step": 10000
+    },
+    {
+      "epoch": 9.0123920390537,
+      "grad_norm": 2.439126491546631,
+      "learning_rate": 2.9995e-05,
+      "loss": 4.7315,
+      "step": 12000
+    },
+    {
+      "epoch": 9.0123920390537,
+      "eval_loss": 4.800222873687744,
+      "eval_runtime": 3.705,
+      "eval_samples_per_second": 896.095,
+      "eval_steps_per_second": 56.141,
+      "step": 12000
+    },
+    {
+      "epoch": 10.514457378895981,
+      "eval_loss": 4.670550346374512,
+      "eval_runtime": 3.7395,
+      "eval_samples_per_second": 887.823,
+      "eval_steps_per_second": 55.623,
+      "step": 14000
+    },
+    {
+      "epoch": 12.016522718738265,
+      "grad_norm": 2.3431427478790283,
+      "learning_rate": 3.999e-05,
+      "loss": 4.4253,
+      "step": 16000
+    },
+    {
+      "epoch": 12.016522718738265,
+      "eval_loss": 4.555232524871826,
+      "eval_runtime": 3.7089,
+      "eval_samples_per_second": 895.14,
+      "eval_steps_per_second": 56.081,
+      "step": 16000
+    },
+    {
+      "epoch": 13.518588058580548,
+      "eval_loss": 4.4522013664245605,
+      "eval_runtime": 3.4306,
+      "eval_samples_per_second": 967.764,
+      "eval_steps_per_second": 60.631,
+      "step": 18000
+    },
+    {
+      "epoch": 15.020653398422832,
+      "grad_norm": 2.3183419704437256,
+      "learning_rate": 4.9985e-05,
+      "loss": 4.1962,
+      "step": 20000
+    },
+    {
+      "epoch": 15.020653398422832,
+      "eval_loss": 4.373892784118652,
+      "eval_runtime": 3.1907,
+      "eval_samples_per_second": 1040.508,
+      "eval_steps_per_second": 65.188,
+      "step": 20000
+    },
+    {
+      "epoch": 16.522718738265116,
+      "eval_loss": 4.292479515075684,
+      "eval_runtime": 3.6404,
+      "eval_samples_per_second": 911.984,
+      "eval_steps_per_second": 57.136,
+      "step": 22000
+    },
+    {
+      "epoch": 18.0247840781074,
+      "grad_norm": 2.1154417991638184,
+      "learning_rate": 5.9980000000000005e-05,
+      "loss": 4.0105,
+      "step": 24000
+    },
+    {
+      "epoch": 18.0247840781074,
+      "eval_loss": 4.227701663970947,
+      "eval_runtime": 3.6867,
+      "eval_samples_per_second": 900.545,
+      "eval_steps_per_second": 56.42,
+      "step": 24000
+    },
+    {
+      "epoch": 19.52684941794968,
+      "eval_loss": 4.170625686645508,
+      "eval_runtime": 3.5669,
+      "eval_samples_per_second": 930.781,
+      "eval_steps_per_second": 58.314,
+      "step": 26000
+    },
+    {
+      "epoch": 21.028914757791963,
+      "grad_norm": 2.051388740539551,
+      "learning_rate": 6.997500000000001e-05,
+      "loss": 3.857,
+      "step": 28000
+    },
+    {
+      "epoch": 21.028914757791963,
+      "eval_loss": 4.124534606933594,
+      "eval_runtime": 3.7721,
+      "eval_samples_per_second": 880.155,
+      "eval_steps_per_second": 55.142,
+      "step": 28000
+    },
+    {
+      "epoch": 22.53098009763425,
+      "eval_loss": 4.084589004516602,
+      "eval_runtime": 3.7783,
+      "eval_samples_per_second": 878.7,
+      "eval_steps_per_second": 55.051,
+      "step": 30000
+    },
+    {
+      "epoch": 24.03304543747653,
+      "grad_norm": 1.966176986694336,
+      "learning_rate": 7.997e-05,
+      "loss": 3.7298,
+      "step": 32000
+    },
+    {
+      "epoch": 24.03304543747653,
+      "eval_loss": 4.058874130249023,
+      "eval_runtime": 3.6613,
+      "eval_samples_per_second": 906.785,
+      "eval_steps_per_second": 56.811,
+      "step": 32000
+    },
+    {
+      "epoch": 25.535110777318813,
+      "eval_loss": 4.026635646820068,
+      "eval_runtime": 3.722,
+      "eval_samples_per_second": 891.982,
+      "eval_steps_per_second": 55.883,
+      "step": 34000
+    },
+    {
+      "epoch": 27.037176117161096,
+      "grad_norm": 1.8346107006072998,
+      "learning_rate": 8.9965e-05,
+      "loss": 3.6218,
+      "step": 36000
+    },
+    {
+      "epoch": 27.037176117161096,
+      "eval_loss": 4.0028157234191895,
+      "eval_runtime": 3.7484,
+      "eval_samples_per_second": 885.713,
+      "eval_steps_per_second": 55.49,
+      "step": 36000
+    },
+    {
+      "epoch": 28.539241457003378,
+      "eval_loss": 3.986341953277588,
+      "eval_runtime": 3.7709,
+      "eval_samples_per_second": 880.423,
+      "eval_steps_per_second": 55.159,
+      "step": 38000
+    },
+    {
+      "epoch": 30.041306796845664,
+      "grad_norm": 1.8153942823410034,
+      "learning_rate": 9.996000000000001e-05,
+      "loss": 3.5278,
+      "step": 40000
+    },
+    {
+      "epoch": 30.041306796845664,
+      "eval_loss": 3.9731833934783936,
+      "eval_runtime": 3.8054,
+      "eval_samples_per_second": 872.434,
+      "eval_steps_per_second": 54.659,
+      "step": 40000
+    },
+    {
+      "epoch": 31.543372136687946,
+      "eval_loss": 3.971970796585083,
+      "eval_runtime": 3.6892,
+      "eval_samples_per_second": 899.912,
+      "eval_steps_per_second": 56.38,
+      "step": 42000
+    },
+    {
+      "epoch": 33.04543747653023,
+      "grad_norm": 1.8285443782806396,
+      "learning_rate": 9.336333333333334e-05,
+      "loss": 3.4351,
+      "step": 44000
+    },
+    {
+      "epoch": 33.04543747653023,
+      "eval_loss": 3.9599227905273438,
+      "eval_runtime": 3.7507,
+      "eval_samples_per_second": 885.169,
+      "eval_steps_per_second": 55.456,
+      "step": 44000
+    }
+  ],
+  "logging_steps": 4000,
+  "max_steps": 100000,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 76,
+  "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": 1.13747817056256e+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 sha256:77f63c790664816f6ed91c4986d1e19906c6b86d33dddcf3c8935838a9bb688c
+size 5368