mT5ベースのエラー診断モデル

FineTune.sh.o40076479

@@ -0,0 +1,42 @@
+GPU available: True (cuda), used: True
+TPU available: False, using: 0 TPU cores
+IPU available: False, using: 0 IPUs
+HPU available: False, using: 0 HPUs
+Downloading and preparing dataset json/default to /home/ace14459tv/t5maru/cache/json/default-76e405bf2a5f1b35/0.0.0/e347ab1c932092252e717ff3f949105a4dd28b27e842dd53157d2f72e276c2e4...
+
+
+
                                                        
+
                                                                              
+
+
+
                                                        
+
                                                                         
+
+  | Name  | Type            | Params
+------------------------------------------
+0 | model | OptimizedModule | 300 M 
+------------------------------------------
+300 M     Trainable params
+0         Non-trainable params
+300 M     Total params
+1,200.707 Total estimated model params size (MB)
+[2023-06-30 20:54:37,264] torch._inductor.utils: [WARNING] using triton random, expect difference from eager
+Metric val_loss improved. New best score: 1.139
+Metric val_loss improved by 0.743 >= min_delta = 0.0. New best score: 0.397
+Metric val_loss improved by 0.178 >= min_delta = 0.0. New best score: 0.219
+Metric val_loss improved by 0.058 >= min_delta = 0.0. New best score: 0.161
+Metric val_loss improved by 0.027 >= min_delta = 0.0. New best score: 0.134
+Metric val_loss improved by 0.020 >= min_delta = 0.0. New best score: 0.115
+Metric val_loss improved by 0.012 >= min_delta = 0.0. New best score: 0.103
+Metric val_loss improved by 0.005 >= min_delta = 0.0. New best score: 0.098
+Metric val_loss improved by 0.011 >= min_delta = 0.0. New best score: 0.087
+Metric val_loss improved by 0.000 >= min_delta = 0.0. New best score: 0.086
+Metric val_loss improved by 0.004 >= min_delta = 0.0. New best score: 0.083
+Metric val_loss improved by 0.006 >= min_delta = 0.0. New best score: 0.077
+Metric val_loss improved by 0.002 >= min_delta = 0.0. New best score: 0.075
+Monitored metric val_loss did not improve in the last 3 records. Best score: 0.075. Signaling Trainer to stop.
+{"log": "trained", "date": "2023-06-30T20:53:45", "elapsed": "00:05:22", "model": "google/mt5-small", "max_length": 128, "target_max_length": 128, "batch_size": 32, "gradient_accumulation_steps": 1, "train_steps": 2700, "accelerator": "gpu", "devices": "auto", "precision": 32, "strategy": "auto", "gradient_clip_val": 1.0, "compile": true, "solver": "adamw", "lr": 0.0003, "warmup_steps": 1, "training_steps": 100000, "adam_epsilon": 1e-08, "weight_decay": 0.0, "epoch": 17, "step": 1530, "saved": "0630_mT5"}
+😊 testing /home/ace14459tv/t5maru/error_data/0630/error_3_0630_test.jsonl on cuda
+Downloading and preparing dataset generator/default to /home/ace14459tv/t5maru/cache/generator/default-f9a3d4be341e4e78/0.0.0...
+
                                                        
+
                                                                           

config.json

@@ -0,0 +1,31 @@
+{
+  "_name_or_path": "google/mt5-small",
+  "architectures": [
+    "MT5ForConditionalGeneration"
+  ],
+  "d_ff": 1024,
+  "d_kv": 64,
+  "d_model": 512,
+  "decoder_start_token_id": 0,
+  "dense_act_fn": "gelu_new",
+  "dropout_rate": 0.1,
+  "eos_token_id": 1,
+  "feed_forward_proj": "gated-gelu",
+  "initializer_factor": 1.0,
+  "is_encoder_decoder": true,
+  "is_gated_act": true,
+  "layer_norm_epsilon": 1e-06,
+  "model_type": "mt5",
+  "num_decoder_layers": 8,
+  "num_heads": 6,
+  "num_layers": 8,
+  "pad_token_id": 0,
+  "relative_attention_max_distance": 128,
+  "relative_attention_num_buckets": 32,
+  "tie_word_embeddings": false,
+  "tokenizer_class": "T5Tokenizer",
+  "torch_dtype": "float32",
+  "transformers_version": "4.28.1",
+  "use_cache": true,
+  "vocab_size": 250112
+}

generation_config.json

@@ -0,0 +1,7 @@
+{
+  "_from_model_config": true,
+  "decoder_start_token_id": 0,
+  "eos_token_id": 1,
+  "pad_token_id": 0,
+  "transformers_version": "4.28.1"
+}

pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d1338d5d8fe013b9e3c29d74a6d7136c20c259d2fc802699f187ac6d02e48b82
+size 1200769797

special_tokens_map.json

@@ -0,0 +1,5 @@
+{
+  "eos_token": "</s>",
+  "pad_token": "<pad>",
+  "unk_token": "<unk>"
+}

spiece.model

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ef78f86560d809067d12bac6c09f19a462cb3af3f54d2b8acbba26e1433125d6
+size 4309802

t5marulog.jsonl

@@ -0,0 +1 @@
+{"log": "trained", "date": "2023-06-30T20:53:45", "elapsed": "00:05:22", "model": "google/mt5-small", "max_length": 128, "target_max_length": 128, "batch_size": 32, "gradient_accumulation_steps": 1, "train_steps": 2700, "accelerator": "gpu", "devices": "auto", "precision": 32, "strategy": "auto", "gradient_clip_val": 1.0, "compile": true, "solver": "adamw", "lr": 0.0003, "warmup_steps": 1, "training_steps": 100000, "adam_epsilon": 1e-08, "weight_decay": 0.0, "epoch": 17, "step": 1530, "saved": "0630_mT5"}

tokenizer_config.json

@@ -0,0 +1,11 @@
+{
+  "additional_special_tokens": null,
+  "clean_up_tokenization_spaces": true,
+  "eos_token": "</s>",
+  "extra_ids": 0,
+  "model_max_length": 1000000000000000019884624838656,
+  "pad_token": "<pad>",
+  "sp_model_kwargs": {},
+  "tokenizer_class": "T5Tokenizer",
+  "unk_token": "<unk>"
+}