Training in progress, step 28000, checkpoint

checkpoint-28000/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-28000/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-28000/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6f5333ff2e023c6ac4c27b611ed819b72a871ca4d0101c35323bcb922baa184e
+size 51007160

checkpoint-28000/optimizer.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:dad6265f696ad6e419ccacf450d5f273bf3f2c17108301a954e7dcf73fb05514
+size 102078202

checkpoint-28000/rng_state.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2d244deb5d4375f202d66736e641cc08c9ebe6d6e36386aa989114691b973b8e
+size 14308

checkpoint-28000/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e76aa2aaca10756a154d6ed5dccc354d970fd46a0d8af2654f80e4661eb3f74a
+size 1000

checkpoint-28000/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-28000/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-28000/trainer_state.json

@@ -0,0 +1,194 @@
+{
+  "best_metric": 4.0739641189575195,
+  "best_model_checkpoint": "/home/p318482/babyLM_controlled/models_trained/en_clm/childes_30/checkpoint-28000",
+  "epoch": 18.586126783936276,
+  "eval_steps": 2000,
+  "global_step": 28000,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 1.3275804845668768,
+      "eval_loss": 6.8483405113220215,
+      "eval_runtime": 1.2535,
+      "eval_samples_per_second": 1396.096,
+      "eval_steps_per_second": 87.755,
+      "step": 2000
+    },
+    {
+      "epoch": 2.6551609691337537,
+      "grad_norm": 1.5963196754455566,
+      "learning_rate": 1e-05,
+      "loss": 6.7292,
+      "step": 4000
+    },
+    {
+      "epoch": 2.6551609691337537,
+      "eval_loss": 5.239430904388428,
+      "eval_runtime": 1.2364,
+      "eval_samples_per_second": 1415.409,
+      "eval_steps_per_second": 88.969,
+      "step": 4000
+    },
+    {
+      "epoch": 3.9827414537006307,
+      "eval_loss": 4.829074859619141,
+      "eval_runtime": 1.2415,
+      "eval_samples_per_second": 1409.56,
+      "eval_steps_per_second": 88.601,
+      "step": 6000
+    },
+    {
+      "epoch": 5.310321938267507,
+      "grad_norm": 2.6859147548675537,
+      "learning_rate": 1.9997500000000003e-05,
+      "loss": 4.5966,
+      "step": 8000
+    },
+    {
+      "epoch": 5.310321938267507,
+      "eval_loss": 4.593608856201172,
+      "eval_runtime": 1.2358,
+      "eval_samples_per_second": 1416.133,
+      "eval_steps_per_second": 89.014,
+      "step": 8000
+    },
+    {
+      "epoch": 6.637902422834385,
+      "eval_loss": 4.449240684509277,
+      "eval_runtime": 1.3247,
+      "eval_samples_per_second": 1321.089,
+      "eval_steps_per_second": 83.04,
+      "step": 10000
+    },
+    {
+      "epoch": 7.965482907401261,
+      "grad_norm": 2.4272868633270264,
+      "learning_rate": 2.9995e-05,
+      "loss": 4.1336,
+      "step": 12000
+    },
+    {
+      "epoch": 7.965482907401261,
+      "eval_loss": 4.348520278930664,
+      "eval_runtime": 1.2394,
+      "eval_samples_per_second": 1411.988,
+      "eval_steps_per_second": 88.754,
+      "step": 12000
+    },
+    {
+      "epoch": 9.293063391968138,
+      "eval_loss": 4.277740955352783,
+      "eval_runtime": 1.2759,
+      "eval_samples_per_second": 1371.628,
+      "eval_steps_per_second": 86.217,
+      "step": 14000
+    },
+    {
+      "epoch": 10.620643876535015,
+      "grad_norm": 2.4516773223876953,
+      "learning_rate": 3.999e-05,
+      "loss": 3.8576,
+      "step": 16000
+    },
+    {
+      "epoch": 10.620643876535015,
+      "eval_loss": 4.223555564880371,
+      "eval_runtime": 1.2226,
+      "eval_samples_per_second": 1431.375,
+      "eval_steps_per_second": 89.972,
+      "step": 16000
+    },
+    {
+      "epoch": 11.948224361101891,
+      "eval_loss": 4.18381929397583,
+      "eval_runtime": 1.2371,
+      "eval_samples_per_second": 1414.6,
+      "eval_steps_per_second": 88.918,
+      "step": 18000
+    },
+    {
+      "epoch": 13.275804845668768,
+      "grad_norm": 2.2412021160125732,
+      "learning_rate": 4.9985e-05,
+      "loss": 3.6557,
+      "step": 20000
+    },
+    {
+      "epoch": 13.275804845668768,
+      "eval_loss": 4.157003879547119,
+      "eval_runtime": 1.238,
+      "eval_samples_per_second": 1413.601,
+      "eval_steps_per_second": 88.855,
+      "step": 20000
+    },
+    {
+      "epoch": 14.603385330235646,
+      "eval_loss": 4.130462646484375,
+      "eval_runtime": 1.2695,
+      "eval_samples_per_second": 1378.51,
+      "eval_steps_per_second": 86.649,
+      "step": 22000
+    },
+    {
+      "epoch": 15.930965814802523,
+      "grad_norm": 2.039700508117676,
+      "learning_rate": 5.9980000000000005e-05,
+      "loss": 3.4956,
+      "step": 24000
+    },
+    {
+      "epoch": 15.930965814802523,
+      "eval_loss": 4.100128173828125,
+      "eval_runtime": 1.2566,
+      "eval_samples_per_second": 1392.686,
+      "eval_steps_per_second": 87.54,
+      "step": 24000
+    },
+    {
+      "epoch": 17.2585462993694,
+      "eval_loss": 4.096436500549316,
+      "eval_runtime": 1.2377,
+      "eval_samples_per_second": 1413.914,
+      "eval_steps_per_second": 88.875,
+      "step": 26000
+    },
+    {
+      "epoch": 18.586126783936276,
+      "grad_norm": 1.9197760820388794,
+      "learning_rate": 6.997500000000001e-05,
+      "loss": 3.361,
+      "step": 28000
+    },
+    {
+      "epoch": 18.586126783936276,
+      "eval_loss": 4.0739641189575195,
+      "eval_runtime": 1.2403,
+      "eval_samples_per_second": 1410.995,
+      "eval_steps_per_second": 88.691,
+      "step": 28000
+    }
+  ],
+  "logging_steps": 4000,
+  "max_steps": 100000,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 67,
+  "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": 7238981572116480.0,
+  "train_batch_size": 16,
+  "trial_name": null,
+  "trial_params": null
+}

checkpoint-28000/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:85a361fa7873534a4ba4b592e19d32a6f23bc55630c1df486f010c443c87a5fa
+size 5368