Training in progress, step 28000, checkpoint

.gitattributes

@@ -96,3 +96,4 @@ checkpoint-26000/logdata.jsonl filter=lfs diff=lfs merge=lfs -text
 checkpoint-26500/logdata.jsonl filter=lfs diff=lfs merge=lfs -text
 checkpoint-27000/logdata.jsonl filter=lfs diff=lfs merge=lfs -text
 checkpoint-27500/logdata.jsonl filter=lfs diff=lfs merge=lfs -text
+checkpoint-28000/logdata.jsonl filter=lfs diff=lfs merge=lfs -text

checkpoint-28000/config.json

@@ -0,0 +1,32 @@
+{
+  "_name_or_path": "bowphs/pythia-70m-multi",
+  "architectures": [
+    "GPTNeoXForCausalLM"
+  ],
+  "attention_bias": true,
+  "attention_dropout": 0.0,
+  "bos_token_id": 0,
+  "classifier_dropout": 0.1,
+  "eos_token_id": 0,
+  "hidden_act": "gelu",
+  "hidden_dropout": 0.0,
+  "hidden_size": 512,
+  "initializer_range": 0.02,
+  "intermediate_size": 2048,
+  "layer_norm_eps": 1e-05,
+  "max_position_embeddings": 2048,
+  "model_type": "gpt_neox",
+  "num_attention_heads": 8,
+  "num_hidden_layers": 6,
+  "partial_rotary_factor": 0.25,
+  "rope_scaling": null,
+  "rope_theta": 10000,
+  "rotary_emb_base": 10000,
+  "rotary_pct": 0.25,
+  "tie_word_embeddings": false,
+  "torch_dtype": "float32",
+  "transformers_version": "4.48.0.dev0",
+  "use_cache": true,
+  "use_parallel_residual": true,
+  "vocab_size": 50304
+}

checkpoint-28000/generation_config.json

@@ -0,0 +1,6 @@
+{
+  "_from_model_config": true,
+  "bos_token_id": 0,
+  "eos_token_id": 0,
+  "transformers_version": "4.48.0.dev0"
+}

checkpoint-28000/logdata.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bada3bbdafafd73fc364ac6fbca30ad1bc45b3e312b87776b20b72fd0a1ee950
+size 5071621198

checkpoint-28000/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4f99e6227fcd6a4465a24ed20300590da2a729d3967a447453e2afdefa42fecb
+size 281715176

checkpoint-28000/optimizer.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e6dfd183640ef8c1d6d2e1bdab69ae2851e85a717be09369bcf41a665afcf869
+size 563476858

checkpoint-28000/rng_state.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e9624186e1e9b86a4826bc902ee77e6fe9d0bf93ad3560d752f019c58283650e
+size 14244

checkpoint-28000/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3c675878ae4f9290bfd9f59a1a565d5ca16b3fbc30c892c75b366c3d83b6e6fe
+size 1064

checkpoint-28000/special_tokens_map.json

@@ -0,0 +1,23 @@
+{
+  "bos_token": {
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "eos_token": {
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "unk_token": {
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

checkpoint-28000/tokenizer_config.json

@@ -0,0 +1,31 @@
+{
+  "add_bos_token": false,
+  "add_eos_token": false,
+  "add_prefix_space": false,
+  "added_tokens_decoder": {
+    "0": {
+      "content": "<|endoftext|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "1": {
+      "content": "<|padding|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "bos_token": "<|endoftext|>",
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|endoftext|>",
+  "extra_special_tokens": {},
+  "model_max_length": 1000000000000000019884624838656,
+  "pad_token": null,
+  "tokenizer_class": "GPTNeoXTokenizer",
+  "unk_token": "<|endoftext|>"
+}

checkpoint-28000/trainer_state.json

@@ -0,0 +1,392 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 0.9333333333333333,
+  "eval_steps": 2000,
+  "global_step": 28000,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 3.3333333333333335e-05,
+      "eval_accuracy": 0.016437927663734114,
+      "eval_loss": 10.702861785888672,
+      "eval_runtime": 53.5239,
+      "eval_samples_per_second": 93.416,
+      "eval_steps_per_second": 2.933,
+      "step": 1
+    },
+    {
+      "epoch": 6.666666666666667e-05,
+      "eval_accuracy": 0.049564027370478984,
+      "eval_loss": 10.533119201660156,
+      "eval_runtime": 53.3666,
+      "eval_samples_per_second": 93.692,
+      "eval_steps_per_second": 2.942,
+      "step": 2
+    },
+    {
+      "epoch": 0.00013333333333333334,
+      "eval_accuracy": 0.05325317693059629,
+      "eval_loss": 10.302159309387207,
+      "eval_runtime": 53.3789,
+      "eval_samples_per_second": 93.67,
+      "eval_steps_per_second": 2.941,
+      "step": 4
+    },
+    {
+      "epoch": 0.0002666666666666667,
+      "eval_accuracy": 0.053565591397849464,
+      "eval_loss": 10.023477554321289,
+      "eval_runtime": 53.6667,
+      "eval_samples_per_second": 93.168,
+      "eval_steps_per_second": 2.925,
+      "step": 8
+    },
+    {
+      "epoch": 0.0005333333333333334,
+      "eval_accuracy": 0.06345063538611925,
+      "eval_loss": 9.653569221496582,
+      "eval_runtime": 53.2342,
+      "eval_samples_per_second": 93.925,
+      "eval_steps_per_second": 2.949,
+      "step": 16
+    },
+    {
+      "epoch": 0.0010666666666666667,
+      "eval_accuracy": 0.07593020527859237,
+      "eval_loss": 9.028414726257324,
+      "eval_runtime": 53.4229,
+      "eval_samples_per_second": 93.593,
+      "eval_steps_per_second": 2.939,
+      "step": 32
+    },
+    {
+      "epoch": 0.0021333333333333334,
+      "eval_accuracy": 0.08323401759530792,
+      "eval_loss": 8.02489185333252,
+      "eval_runtime": 53.5928,
+      "eval_samples_per_second": 93.296,
+      "eval_steps_per_second": 2.929,
+      "step": 64
+    },
+    {
+      "epoch": 0.004266666666666667,
+      "eval_accuracy": 0.11289071358748778,
+      "eval_loss": 6.917232990264893,
+      "eval_runtime": 53.9308,
+      "eval_samples_per_second": 92.711,
+      "eval_steps_per_second": 2.911,
+      "step": 128
+    },
+    {
+      "epoch": 0.008533333333333334,
+      "eval_accuracy": 0.15581857282502443,
+      "eval_loss": 6.162940502166748,
+      "eval_runtime": 53.873,
+      "eval_samples_per_second": 92.811,
+      "eval_steps_per_second": 2.914,
+      "step": 256
+    },
+    {
+      "epoch": 0.017066666666666667,
+      "eval_accuracy": 0.18169071358748778,
+      "eval_loss": 5.580474376678467,
+      "eval_runtime": 53.993,
+      "eval_samples_per_second": 92.605,
+      "eval_steps_per_second": 2.908,
+      "step": 512
+    },
+    {
+      "epoch": 0.034133333333333335,
+      "eval_accuracy": 0.20284437927663734,
+      "eval_loss": 5.123527526855469,
+      "eval_runtime": 53.733,
+      "eval_samples_per_second": 93.053,
+      "eval_steps_per_second": 2.922,
+      "step": 1024
+    },
+    {
+      "epoch": 0.06666666666666667,
+      "grad_norm": 0.9611704349517822,
+      "learning_rate": 4.666666666666667e-05,
+      "loss": 5.4529,
+      "step": 2000
+    },
+    {
+      "epoch": 0.06666666666666667,
+      "eval_accuracy": 0.2264314760508309,
+      "eval_loss": 4.761301517486572,
+      "eval_runtime": 55.2965,
+      "eval_samples_per_second": 90.422,
+      "eval_steps_per_second": 2.839,
+      "step": 2000
+    },
+    {
+      "epoch": 0.06826666666666667,
+      "eval_accuracy": 0.22808172043010752,
+      "eval_loss": 4.7480597496032715,
+      "eval_runtime": 54.5009,
+      "eval_samples_per_second": 91.742,
+      "eval_steps_per_second": 2.881,
+      "step": 2048
+    },
+    {
+      "epoch": 0.13333333333333333,
+      "grad_norm": 0.9782880544662476,
+      "learning_rate": 4.3333333333333334e-05,
+      "loss": 4.5765,
+      "step": 4000
+    },
+    {
+      "epoch": 0.13333333333333333,
+      "eval_accuracy": 0.2610080156402737,
+      "eval_loss": 4.412333011627197,
+      "eval_runtime": 55.0769,
+      "eval_samples_per_second": 90.782,
+      "eval_steps_per_second": 2.851,
+      "step": 4000
+    },
+    {
+      "epoch": 0.13653333333333334,
+      "eval_accuracy": 0.2625071358748778,
+      "eval_loss": 4.404272556304932,
+      "eval_runtime": 54.9058,
+      "eval_samples_per_second": 91.065,
+      "eval_steps_per_second": 2.859,
+      "step": 4096
+    },
+    {
+      "epoch": 0.2,
+      "grad_norm": 1.0245925188064575,
+      "learning_rate": 4e-05,
+      "loss": 4.3252,
+      "step": 6000
+    },
+    {
+      "epoch": 0.2,
+      "eval_accuracy": 0.2826778103616813,
+      "eval_loss": 4.222066879272461,
+      "eval_runtime": 55.0772,
+      "eval_samples_per_second": 90.782,
+      "eval_steps_per_second": 2.851,
+      "step": 6000
+    },
+    {
+      "epoch": 0.26666666666666666,
+      "grad_norm": 1.9592970609664917,
+      "learning_rate": 3.6666666666666666e-05,
+      "loss": 4.146,
+      "step": 8000
+    },
+    {
+      "epoch": 0.26666666666666666,
+      "eval_accuracy": 0.3098099706744868,
+      "eval_loss": 4.0350446701049805,
+      "eval_runtime": 54.3944,
+      "eval_samples_per_second": 91.921,
+      "eval_steps_per_second": 2.886,
+      "step": 8000
+    },
+    {
+      "epoch": 0.2730666666666667,
+      "eval_accuracy": 0.31292649071358747,
+      "eval_loss": 4.013375759124756,
+      "eval_runtime": 54.918,
+      "eval_samples_per_second": 91.045,
+      "eval_steps_per_second": 2.859,
+      "step": 8192
+    },
+    {
+      "epoch": 0.3333333333333333,
+      "grad_norm": 1.1236926317214966,
+      "learning_rate": 3.3333333333333335e-05,
+      "loss": 3.9652,
+      "step": 10000
+    },
+    {
+      "epoch": 0.3333333333333333,
+      "eval_accuracy": 0.3303505376344086,
+      "eval_loss": 3.885976552963257,
+      "eval_runtime": 54.6579,
+      "eval_samples_per_second": 91.478,
+      "eval_steps_per_second": 2.872,
+      "step": 10000
+    },
+    {
+      "epoch": 0.4,
+      "grad_norm": 1.4833592176437378,
+      "learning_rate": 3e-05,
+      "loss": 3.8441,
+      "step": 12000
+    },
+    {
+      "epoch": 0.4,
+      "eval_accuracy": 0.34179745845552295,
+      "eval_loss": 3.800518274307251,
+      "eval_runtime": 55.3363,
+      "eval_samples_per_second": 90.357,
+      "eval_steps_per_second": 2.837,
+      "step": 12000
+    },
+    {
+      "epoch": 0.4666666666666667,
+      "grad_norm": 1.413355827331543,
+      "learning_rate": 2.6666666666666667e-05,
+      "loss": 3.7739,
+      "step": 14000
+    },
+    {
+      "epoch": 0.4666666666666667,
+      "eval_accuracy": 0.3503352883675464,
+      "eval_loss": 3.7315235137939453,
+      "eval_runtime": 54.617,
+      "eval_samples_per_second": 91.547,
+      "eval_steps_per_second": 2.875,
+      "step": 14000
+    },
+    {
+      "epoch": 0.5333333333333333,
+      "grad_norm": 1.4252855777740479,
+      "learning_rate": 2.3333333333333336e-05,
+      "loss": 3.72,
+      "step": 16000
+    },
+    {
+      "epoch": 0.5333333333333333,
+      "eval_accuracy": 0.35525493646138806,
+      "eval_loss": 3.688023328781128,
+      "eval_runtime": 54.7137,
+      "eval_samples_per_second": 91.385,
+      "eval_steps_per_second": 2.869,
+      "step": 16000
+    },
+    {
+      "epoch": 0.5461333333333334,
+      "eval_accuracy": 0.356411339198436,
+      "eval_loss": 3.6776843070983887,
+      "eval_runtime": 55.1644,
+      "eval_samples_per_second": 90.638,
+      "eval_steps_per_second": 2.846,
+      "step": 16384
+    },
+    {
+      "epoch": 0.6,
+      "grad_norm": 1.4446940422058105,
+      "learning_rate": 2e-05,
+      "loss": 3.6718,
+      "step": 18000
+    },
+    {
+      "epoch": 0.6,
+      "eval_accuracy": 0.3592500488758553,
+      "eval_loss": 3.6533164978027344,
+      "eval_runtime": 53.9881,
+      "eval_samples_per_second": 92.613,
+      "eval_steps_per_second": 2.908,
+      "step": 18000
+    },
+    {
+      "epoch": 0.6666666666666666,
+      "grad_norm": 1.3672009706497192,
+      "learning_rate": 1.6666666666666667e-05,
+      "loss": 3.6527,
+      "step": 20000
+    },
+    {
+      "epoch": 0.6666666666666666,
+      "eval_accuracy": 0.3633403714565005,
+      "eval_loss": 3.6212007999420166,
+      "eval_runtime": 53.6539,
+      "eval_samples_per_second": 93.19,
+      "eval_steps_per_second": 2.926,
+      "step": 20000
+    },
+    {
+      "epoch": 0.7333333333333333,
+      "grad_norm": 1.5204988718032837,
+      "learning_rate": 1.3333333333333333e-05,
+      "loss": 3.6201,
+      "step": 22000
+    },
+    {
+      "epoch": 0.7333333333333333,
+      "eval_accuracy": 0.36595581622678397,
+      "eval_loss": 3.598499298095703,
+      "eval_runtime": 54.0759,
+      "eval_samples_per_second": 92.463,
+      "eval_steps_per_second": 2.903,
+      "step": 22000
+    },
+    {
+      "epoch": 0.8,
+      "grad_norm": 2.7004802227020264,
+      "learning_rate": 1e-05,
+      "loss": 3.593,
+      "step": 24000
+    },
+    {
+      "epoch": 0.8,
+      "eval_accuracy": 0.3679088954056696,
+      "eval_loss": 3.5819320678710938,
+      "eval_runtime": 53.8279,
+      "eval_samples_per_second": 92.889,
+      "eval_steps_per_second": 2.917,
+      "step": 24000
+    },
+    {
+      "epoch": 0.8666666666666667,
+      "grad_norm": 1.3153384923934937,
+      "learning_rate": 6.666666666666667e-06,
+      "loss": 3.5857,
+      "step": 26000
+    },
+    {
+      "epoch": 0.8666666666666667,
+      "eval_accuracy": 0.36973333333333336,
+      "eval_loss": 3.5683183670043945,
+      "eval_runtime": 54.3226,
+      "eval_samples_per_second": 92.043,
+      "eval_steps_per_second": 2.89,
+      "step": 26000
+    },
+    {
+      "epoch": 0.9333333333333333,
+      "grad_norm": 1.3367434740066528,
+      "learning_rate": 3.3333333333333333e-06,
+      "loss": 3.5801,
+      "step": 28000
+    },
+    {
+      "epoch": 0.9333333333333333,
+      "eval_accuracy": 0.37106158357771263,
+      "eval_loss": 3.558154582977295,
+      "eval_runtime": 53.6794,
+      "eval_samples_per_second": 93.146,
+      "eval_steps_per_second": 2.925,
+      "step": 28000
+    }
+  ],
+  "logging_steps": 2000,
+  "max_steps": 30000,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 9223372036854775807,
+  "save_steps": 500,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 2.45914794983424e+17,
+  "train_batch_size": 32,
+  "trial_name": null,
+  "trial_params": null
+}

checkpoint-28000/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7bfb964096776c793ae319346657629d176228270098861d014f2f3c248437f7
+size 5304