Training in progress, epoch 3, checkpoint

checkpoint-744/added_tokens.json

@@ -0,0 +1,3 @@
+{
+  "<|endoftext|>": 50256
+}

checkpoint-744/config.json

@@ -0,0 +1,42 @@
+{
+  "_name_or_path": "gpt2-xl",
+  "activation_function": "gelu_new",
+  "architectures": [
+    "GPT2LMHeadModel"
+  ],
+  "attn_pdrop": 0.1,
+  "bos_token_id": 50256,
+  "do_sample": true,
+  "embd_pdrop": 0.1,
+  "eos_token_id": 50256,
+  "initializer_range": 0.02,
+  "layer_norm_epsilon": 1e-05,
+  "max_length": 50,
+  "model_type": "gpt2",
+  "n_ctx": 1024,
+  "n_embd": 1600,
+  "n_head": 25,
+  "n_inner": null,
+  "n_layer": 48,
+  "n_positions": 1024,
+  "output_past": true,
+  "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,
+  "task_specific_params": {
+    "text-generation": {
+      "do_sample": true,
+      "max_length": 50
+    }
+  },
+  "torch_dtype": "float32",
+  "transformers_version": "4.34.0",
+  "use_cache": true,
+  "vocab_size": 50257
+}

checkpoint-744/generation_config.json

@@ -0,0 +1,8 @@
+{
+  "_from_model_config": true,
+  "bos_token_id": 50256,
+  "do_sample": true,
+  "eos_token_id": 50256,
+  "max_length": 50,
+  "transformers_version": "4.34.0"
+}

checkpoint-744/optimizer.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c55d0647f09c313fe291fc482f4dfbbb94e7e6761f42a3c1ded1d4df7b1f0634
+size 12461385454

checkpoint-744/pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:836174061df39bc54fd4c106b19f4fdf178b5c9d5c4e204755a9349ba9e5a667
+size 6230637102

checkpoint-744/rng_state.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b234dd541f6edf7b1bd46744cde570c69bbe7919aee598201ec2ea09761cef8b
+size 14244

checkpoint-744/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e3add6fdcd0986b651450acf085d2dfbffaa2881b5adc1e78d28db76ffa12daa
+size 1064

checkpoint-744/special_tokens_map.json

@@ -0,0 +1,6 @@
+{
+  "bos_token": "<|endoftext|>",
+  "eos_token": "<|endoftext|>",
+  "pad_token": "<|endoftext|>",
+  "unk_token": "<|endoftext|>"
+}

checkpoint-744/tokenizer_config.json

@@ -0,0 +1,20 @@
+{
+  "add_prefix_space": false,
+  "added_tokens_decoder": {
+    "50256": {
+      "content": "<|endoftext|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "additional_special_tokens": [],
+  "bos_token": "<|endoftext|>",
+  "clean_up_tokenization_spaces": true,
+  "eos_token": "<|endoftext|>",
+  "model_max_length": 1024,
+  "tokenizer_class": "GPT2Tokenizer",
+  "unk_token": "<|endoftext|>"
+}

checkpoint-744/trainer_state.json

@@ -0,0 +1,160 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 3.0,
+  "eval_steps": 248,
+  "global_step": 744,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.2,
+      "learning_rate": 3e-05,
+      "loss": 2.8428,
+      "step": 50
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 3e-05,
+      "loss": 2.1589,
+      "step": 100
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 3e-05,
+      "loss": 2.0527,
+      "step": 150
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 3e-05,
+      "loss": 1.9823,
+      "step": 200
+    },
+    {
+      "epoch": 1.0,
+      "eval_accuracy": 0.6398457059679767,
+      "eval_loss": 1.6149228811264038,
+      "eval_runtime": 17.4504,
+      "eval_samples_per_second": 28.653,
+      "eval_steps_per_second": 1.834,
+      "step": 248
+    },
+    {
+      "epoch": 1.0,
+      "eval_exact_match": 9.0,
+      "eval_f1": 13.45666666666666,
+      "eval_qa_bleu": 7.063432441196736,
+      "eval_qa_exact_match": 0.09,
+      "eval_recite_bleu": 15.344658855358656,
+      "eval_recite_exact_match": 0.0,
+      "step": 248
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 3e-05,
+      "loss": 1.9153,
+      "step": 250
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 3e-05,
+      "loss": 1.49,
+      "step": 300
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 3e-05,
+      "loss": 1.4406,
+      "step": 350
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 3e-05,
+      "loss": 1.4161,
+      "step": 400
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 3e-05,
+      "loss": 1.3868,
+      "step": 450
+    },
+    {
+      "epoch": 2.0,
+      "eval_accuracy": 0.6700320232896652,
+      "eval_loss": 1.1928881406784058,
+      "eval_runtime": 15.4486,
+      "eval_samples_per_second": 32.365,
+      "eval_steps_per_second": 2.071,
+      "step": 496
+    },
+    {
+      "epoch": 2.0,
+      "eval_exact_match": 9.2,
+      "eval_f1": 14.735098500533276,
+      "eval_qa_bleu": 7.255412763224127,
+      "eval_qa_exact_match": 0.088,
+      "eval_recite_bleu": 17.979965007898493,
+      "eval_recite_exact_match": 0.0,
+      "step": 496
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 3e-05,
+      "loss": 1.309,
+      "step": 500
+    },
+    {
+      "epoch": 2.22,
+      "learning_rate": 3e-05,
+      "loss": 0.9401,
+      "step": 550
+    },
+    {
+      "epoch": 2.42,
+      "learning_rate": 3e-05,
+      "loss": 0.9429,
+      "step": 600
+    },
+    {
+      "epoch": 2.62,
+      "learning_rate": 3e-05,
+      "loss": 0.9115,
+      "step": 650
+    },
+    {
+      "epoch": 2.82,
+      "learning_rate": 3e-05,
+      "loss": 0.9102,
+      "step": 700
+    },
+    {
+      "epoch": 3.0,
+      "eval_accuracy": 0.6992489082969432,
+      "eval_loss": 0.8513368368148804,
+      "eval_runtime": 15.7421,
+      "eval_samples_per_second": 31.762,
+      "eval_steps_per_second": 2.033,
+      "step": 744
+    },
+    {
+      "epoch": 3.0,
+      "eval_exact_match": 10.2,
+      "eval_f1": 15.888450046685344,
+      "eval_qa_bleu": 7.2379640900667646,
+      "eval_qa_exact_match": 0.098,
+      "eval_recite_bleu": 20.876580135938546,
+      "eval_recite_exact_match": 0.0,
+      "step": 744
+    }
+  ],
+  "logging_steps": 50,
+  "max_steps": 4960,
+  "num_train_epochs": 20,
+  "save_steps": 500,
+  "total_flos": 3.4656338231808e+16,
+  "trial_name": null,
+  "trial_params": null
+}

checkpoint-744/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:acdd81e53b27f902dc4c1896d0c5253609dec7800ac5e5d60f5a323620c45c4a
+size 4728