Training in progress, epoch 1

emissions.csv

@@ -6,4 +6,7 @@ timestamp,project_name,run_id,duration,emissions,emissions_rate,cpu_power,gpu_po
 2024-03-20T11:09:11,codecarbon,8c5cd6be-b949-4d3e-aeee-44d3635c6e6a,202.64030289649963,0.000394073813688,1.944696134259563e-06,42.5,121.5938351011336,96.0,0.002392233763304,0.006508886873771,0.0054028290430704,0.0143039496801455,Norway,NOR,viken,,,Linux-5.14.0-362.8.1.el9_3.x86_64-x86_64-with-glibc2.34,3.11.3,2.3.4,24,AMD EPYC 7642 48-Core Processor,1,1 x NVIDIA A100-PCIE-40GB,10.8034,59.6573,256,machine,N,1.0
 2024-03-20T11:10:18,codecarbon,8c5cd6be-b949-4d3e-aeee-44d3635c6e6a,269.8455903530121,0.0005163006217623,1.913318728265911e-06,42.5,99.17360382771624,96.0,0.003185620978309,0.008360091132511,0.0071947823270161,0.0187404944378362,Norway,NOR,viken,,,Linux-5.14.0-362.8.1.el9_3.x86_64-x86_64-with-glibc2.34,3.11.3,2.3.4,24,AMD EPYC 7642 48-Core Processor,1,1 x NVIDIA A100-PCIE-40GB,10.8034,59.6573,256,machine,N,1.0
 2024-03-20T11:14:00,codecarbon,8c5cd6be-b949-4d3e-aeee-44d3635c6e6a,492.169837474823,0.0009453419254476,1.920763633744622e-06,42.5,113.67679084499753,96.0,0.005810271401041,0.0153801814708014,0.013123228632609,0.0343136815044515,Norway,NOR,viken,,,Linux-5.14.0-362.8.1.el9_3.x86_64-x86_64-with-glibc2.34,3.11.3,2.3.4,24,AMD EPYC 7642 48-Core Processor,1,1 x NVIDIA A100-PCIE-40GB,10.8034,59.6573,256,machine,N,1.0
-2024-03-20T11:15:05,codecarbon,8c5cd6be-b949-4d3e-aeee-44d3635c6e6a,556.7771844863892,0.001067438719091254,1.917173959051385e-06,42.5,108.4736124826506,96.0,0.006572990199923516,0.017326653305753226,0.014845863539377849,0.03874550704505459,Norway,NOR,viken,,,Linux-5.14.0-362.8.1.el9_3.x86_64-x86_64-with-glibc2.34,3.11.3,2.3.4,24,AMD EPYC 7642 48-Core Processor,1,1 x NVIDIA A100-PCIE-40GB,10.8034,59.6573,256,machine,N,1.0
+2024-03-20T11:15:05,codecarbon,8c5cd6be-b949-4d3e-aeee-44d3635c6e6a,556.7771844863892,0.0010674387190912,1.917173959051385e-06,42.5,108.4736124826506,96.0,0.0065729901999235,0.0173266533057532,0.0148458635393778,0.0387455070450545,Norway,NOR,viken,,,Linux-5.14.0-362.8.1.el9_3.x86_64-x86_64-with-glibc2.34,3.11.3,2.3.4,24,AMD EPYC 7642 48-Core Processor,1,1 x NVIDIA A100-PCIE-40GB,10.8034,59.6573,256,machine,N,1.0
+2024-03-20T11:15:34,codecarbon,8c5cd6be-b949-4d3e-aeee-44d3635c6e6a,585.5882697105408,0.0011148245747178,1.903768624444728e-06,42.5,76.4417344149013,96.0,0.0069131124711698,0.0179383418506606,0.0156140474828084,0.040465501804639,Norway,NOR,viken,,,Linux-5.14.0-362.8.1.el9_3.x86_64-x86_64-with-glibc2.34,3.11.3,2.3.4,24,AMD EPYC 7642 48-Core Processor,1,1 x NVIDIA A100-PCIE-40GB,10.8034,59.6573,256,machine,N,1.0
+2024-03-20T11:16:20,codecarbon,8c5cd6be-b949-4d3e-aeee-44d3635c6e6a,631.8965435028076,0.0012026320726012,1.903210398864824e-06,42.5,109.304377937163,96.0,0.0074597954246732,0.0193441649197634,0.0168487464650472,0.0436527068094839,Norway,NOR,viken,,,Linux-5.14.0-362.8.1.el9_3.x86_64-x86_64-with-glibc2.34,3.11.3,2.3.4,24,AMD EPYC 7642 48-Core Processor,1,1 x NVIDIA A100-PCIE-40GB,10.8034,59.6573,256,machine,N,1.0
+2024-03-20T11:16:38,codecarbon,8c5cd6be-b949-4d3e-aeee-44d3635c6e6a,650.12664103508,0.001239097760201365,1.9059329090537992e-06,42.5,122.96952663448235,96.0,0.007675000440412098,0.01996663625107331,0.01733468854268392,0.04497632523416932,Norway,NOR,viken,,,Linux-5.14.0-362.8.1.el9_3.x86_64-x86_64-with-glibc2.34,3.11.3,2.3.4,24,AMD EPYC 7642 48-Core Processor,1,1 x NVIDIA A100-PCIE-40GB,10.8034,59.6573,256,machine,N,1.0

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:243a6355c9ccdf996b8c52460c68b2382106a7e923e7402167b0e7a83fdce44c
+oid sha256:602c8c455abf7fad327e2d64c5bedf2ddfbae8013f9edc3c3826cd3c67ff6ed0
 size 267832560

run-5/checkpoint-268/config.json

@@ -0,0 +1,25 @@
+{
+  "_name_or_path": "distilbert-base-uncased",
+  "activation": "gelu",
+  "architectures": [
+    "DistilBertForSequenceClassification"
+  ],
+  "attention_dropout": 0.1,
+  "dim": 768,
+  "dropout": 0.1,
+  "hidden_dim": 3072,
+  "initializer_range": 0.02,
+  "max_position_embeddings": 512,
+  "model_type": "distilbert",
+  "n_heads": 12,
+  "n_layers": 6,
+  "pad_token_id": 0,
+  "problem_type": "single_label_classification",
+  "qa_dropout": 0.1,
+  "seq_classif_dropout": 0.2,
+  "sinusoidal_pos_embds": false,
+  "tie_weights_": true,
+  "torch_dtype": "float32",
+  "transformers_version": "4.38.2",
+  "vocab_size": 30522
+}

run-5/checkpoint-268/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:444d248ddceb88fc80e192b2b328afd1c3023c5873926a0970c9d815cc25f761
+size 267832560

run-5/checkpoint-268/optimizer.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9cb0908dd4d17587419d1beccf4a01d03625214d4e448ace57e54a79b93aada1
+size 535727290

run-5/checkpoint-268/rng_state.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:910018c3ed131ef749bb949c95c7724cab02051ac2e5905521739bc37755dc10
+size 14244

run-5/checkpoint-268/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ecfd0b9369b10d457f301b0210d6d3049dd5e24299911ce7a874f59bcbf1cd78
+size 1064

run-5/checkpoint-268/special_tokens_map.json

@@ -0,0 +1,7 @@
+{
+  "cls_token": "[CLS]",
+  "mask_token": "[MASK]",
+  "pad_token": "[PAD]",
+  "sep_token": "[SEP]",
+  "unk_token": "[UNK]"
+}

run-5/checkpoint-268/tokenizer_config.json

@@ -0,0 +1,55 @@
+{
+  "added_tokens_decoder": {
+    "0": {
+      "content": "[PAD]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "100": {
+      "content": "[UNK]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "101": {
+      "content": "[CLS]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "102": {
+      "content": "[SEP]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "103": {
+      "content": "[MASK]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "clean_up_tokenization_spaces": true,
+  "cls_token": "[CLS]",
+  "do_lower_case": true,
+  "mask_token": "[MASK]",
+  "model_max_length": 512,
+  "pad_token": "[PAD]",
+  "sep_token": "[SEP]",
+  "strip_accents": null,
+  "tokenize_chinese_chars": true,
+  "tokenizer_class": "DistilBertTokenizer",
+  "unk_token": "[UNK]"
+}

run-5/checkpoint-268/trainer_state.json

@@ -0,0 +1,44 @@
+{
+  "best_metric": 0.4231731541032522,
+  "best_model_checkpoint": "distilbert-base-uncased-finetuned-cola/run-5/checkpoint-268",
+  "epoch": 2.0,
+  "eval_steps": 500,
+  "global_step": 268,
+  "is_hyper_param_search": true,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 1.0,
+      "eval_loss": 0.5264953970909119,
+      "eval_matthews_correlation": 0.3429695650358,
+      "eval_runtime": 0.34,
+      "eval_samples_per_second": 3068.067,
+      "eval_steps_per_second": 194.144,
+      "step": 134
+    },
+    {
+      "epoch": 2.0,
+      "eval_loss": 0.5003471374511719,
+      "eval_matthews_correlation": 0.4231731541032522,
+      "eval_runtime": 0.3582,
+      "eval_samples_per_second": 2911.918,
+      "eval_steps_per_second": 184.263,
+      "step": 268
+    }
+  ],
+  "logging_steps": 500,
+  "max_steps": 536,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 4,
+  "save_steps": 500,
+  "total_flos": 0,
+  "train_batch_size": 64,
+  "trial_name": null,
+  "trial_params": {
+    "learning_rate": 1.1390296624220138e-05,
+    "num_train_epochs": 4,
+    "per_device_train_batch_size": 64,
+    "seed": 11
+  }
+}

run-5/checkpoint-268/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7b0a20b7e1ced1e066ae0be151da2892022f2d3caa0764af95b56a89f7778c56
+size 4984

run-5/checkpoint-402/config.json

@@ -0,0 +1,25 @@
+{
+  "_name_or_path": "distilbert-base-uncased",
+  "activation": "gelu",
+  "architectures": [
+    "DistilBertForSequenceClassification"
+  ],
+  "attention_dropout": 0.1,
+  "dim": 768,
+  "dropout": 0.1,
+  "hidden_dim": 3072,
+  "initializer_range": 0.02,
+  "max_position_embeddings": 512,
+  "model_type": "distilbert",
+  "n_heads": 12,
+  "n_layers": 6,
+  "pad_token_id": 0,
+  "problem_type": "single_label_classification",
+  "qa_dropout": 0.1,
+  "seq_classif_dropout": 0.2,
+  "sinusoidal_pos_embds": false,
+  "tie_weights_": true,
+  "torch_dtype": "float32",
+  "transformers_version": "4.38.2",
+  "vocab_size": 30522
+}

run-5/checkpoint-402/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4e64ad1b6293423a6d6dd58d02ce33bfe560912bf9c605d28cbe5d46e5446019
+size 267832560

run-5/checkpoint-402/optimizer.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:932984d202b3f58895fe5fae5d6c4d75997c82a5428cb5d973a9036c458ec99f
+size 535727290

run-5/checkpoint-402/rng_state.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:52f4a1a7897c982c39ce1fe8d2f78dbe1f6a1dcf9fbfdc0f8e0c3a185e685970
+size 14244

run-5/checkpoint-402/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7ed34ef472c75ddd6829c8ca5de8caec646cfa8342da0c315fa7fe4e1d2f113a
+size 1064

run-5/checkpoint-402/special_tokens_map.json

@@ -0,0 +1,7 @@
+{
+  "cls_token": "[CLS]",
+  "mask_token": "[MASK]",
+  "pad_token": "[PAD]",
+  "sep_token": "[SEP]",
+  "unk_token": "[UNK]"
+}

run-5/checkpoint-402/tokenizer_config.json

@@ -0,0 +1,55 @@
+{
+  "added_tokens_decoder": {
+    "0": {
+      "content": "[PAD]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "100": {
+      "content": "[UNK]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "101": {
+      "content": "[CLS]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "102": {
+      "content": "[SEP]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "103": {
+      "content": "[MASK]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "clean_up_tokenization_spaces": true,
+  "cls_token": "[CLS]",
+  "do_lower_case": true,
+  "mask_token": "[MASK]",
+  "model_max_length": 512,
+  "pad_token": "[PAD]",
+  "sep_token": "[SEP]",
+  "strip_accents": null,
+  "tokenize_chinese_chars": true,
+  "tokenizer_class": "DistilBertTokenizer",
+  "unk_token": "[UNK]"
+}

run-5/checkpoint-402/trainer_state.json

@@ -0,0 +1,53 @@
+{
+  "best_metric": 0.4445067035273154,
+  "best_model_checkpoint": "distilbert-base-uncased-finetuned-cola/run-5/checkpoint-402",
+  "epoch": 3.0,
+  "eval_steps": 500,
+  "global_step": 402,
+  "is_hyper_param_search": true,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 1.0,
+      "eval_loss": 0.5264953970909119,
+      "eval_matthews_correlation": 0.3429695650358,
+      "eval_runtime": 0.34,
+      "eval_samples_per_second": 3068.067,
+      "eval_steps_per_second": 194.144,
+      "step": 134
+    },
+    {
+      "epoch": 2.0,
+      "eval_loss": 0.5003471374511719,
+      "eval_matthews_correlation": 0.4231731541032522,
+      "eval_runtime": 0.3582,
+      "eval_samples_per_second": 2911.918,
+      "eval_steps_per_second": 184.263,
+      "step": 268
+    },
+    {
+      "epoch": 3.0,
+      "eval_loss": 0.5117089748382568,
+      "eval_matthews_correlation": 0.4445067035273154,
+      "eval_runtime": 0.3618,
+      "eval_samples_per_second": 2882.875,
+      "eval_steps_per_second": 182.425,
+      "step": 402
+    }
+  ],
+  "logging_steps": 500,
+  "max_steps": 536,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 4,
+  "save_steps": 500,
+  "total_flos": 0,
+  "train_batch_size": 64,
+  "trial_name": null,
+  "trial_params": {
+    "learning_rate": 1.1390296624220138e-05,
+    "num_train_epochs": 4,
+    "per_device_train_batch_size": 64,
+    "seed": 11
+  }
+}

run-5/checkpoint-402/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7b0a20b7e1ced1e066ae0be151da2892022f2d3caa0764af95b56a89f7778c56
+size 4984

run-5/checkpoint-536/config.json

@@ -0,0 +1,25 @@
+{
+  "_name_or_path": "distilbert-base-uncased",
+  "activation": "gelu",
+  "architectures": [
+    "DistilBertForSequenceClassification"
+  ],
+  "attention_dropout": 0.1,
+  "dim": 768,
+  "dropout": 0.1,
+  "hidden_dim": 3072,
+  "initializer_range": 0.02,
+  "max_position_embeddings": 512,
+  "model_type": "distilbert",
+  "n_heads": 12,
+  "n_layers": 6,
+  "pad_token_id": 0,
+  "problem_type": "single_label_classification",
+  "qa_dropout": 0.1,
+  "seq_classif_dropout": 0.2,
+  "sinusoidal_pos_embds": false,
+  "tie_weights_": true,
+  "torch_dtype": "float32",
+  "transformers_version": "4.38.2",
+  "vocab_size": 30522
+}

run-5/checkpoint-536/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cd6574c9e2af7328df83ed77db62b7779e36c1ae19f4d6a83ac271f25de70bf7
+size 267832560

run-5/checkpoint-536/optimizer.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2b683d725e5b7a5881d0340e1d6879dc804ded6dd9ef30018f2932de31bb8bc3
+size 535727290

run-5/checkpoint-536/rng_state.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:968d4727b85fd0c1a08075f3fb59026e6a950c1b2b8a511a1b1b9ece6c190283
+size 14244

run-5/checkpoint-536/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d155bd932713a4258fa298ae3bdcee81c4a392156a676f81cf065673d4e55234
+size 1064

run-5/checkpoint-536/special_tokens_map.json

@@ -0,0 +1,7 @@
+{
+  "cls_token": "[CLS]",
+  "mask_token": "[MASK]",
+  "pad_token": "[PAD]",
+  "sep_token": "[SEP]",
+  "unk_token": "[UNK]"
+}

run-5/checkpoint-536/tokenizer_config.json

@@ -0,0 +1,55 @@
+{
+  "added_tokens_decoder": {
+    "0": {
+      "content": "[PAD]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "100": {
+      "content": "[UNK]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "101": {
+      "content": "[CLS]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "102": {
+      "content": "[SEP]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "103": {
+      "content": "[MASK]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "clean_up_tokenization_spaces": true,
+  "cls_token": "[CLS]",
+  "do_lower_case": true,
+  "mask_token": "[MASK]",
+  "model_max_length": 512,
+  "pad_token": "[PAD]",
+  "sep_token": "[SEP]",
+  "strip_accents": null,
+  "tokenize_chinese_chars": true,
+  "tokenizer_class": "DistilBertTokenizer",
+  "unk_token": "[UNK]"
+}

run-5/checkpoint-536/trainer_state.json

@@ -0,0 +1,69 @@
+{
+  "best_metric": 0.47896393100906826,
+  "best_model_checkpoint": "distilbert-base-uncased-finetuned-cola/run-5/checkpoint-536",
+  "epoch": 4.0,
+  "eval_steps": 500,
+  "global_step": 536,
+  "is_hyper_param_search": true,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 1.0,
+      "eval_loss": 0.5264953970909119,
+      "eval_matthews_correlation": 0.3429695650358,
+      "eval_runtime": 0.34,
+      "eval_samples_per_second": 3068.067,
+      "eval_steps_per_second": 194.144,
+      "step": 134
+    },
+    {
+      "epoch": 2.0,
+      "eval_loss": 0.5003471374511719,
+      "eval_matthews_correlation": 0.4231731541032522,
+      "eval_runtime": 0.3582,
+      "eval_samples_per_second": 2911.918,
+      "eval_steps_per_second": 184.263,
+      "step": 268
+    },
+    {
+      "epoch": 3.0,
+      "eval_loss": 0.5117089748382568,
+      "eval_matthews_correlation": 0.4445067035273154,
+      "eval_runtime": 0.3618,
+      "eval_samples_per_second": 2882.875,
+      "eval_steps_per_second": 182.425,
+      "step": 402
+    },
+    {
+      "epoch": 3.73,
+      "grad_norm": 3.712400197982788,
+      "learning_rate": 7.65019922522248e-07,
+      "loss": 0.4486,
+      "step": 500
+    },
+    {
+      "epoch": 4.0,
+      "eval_loss": 0.5077244639396667,
+      "eval_matthews_correlation": 0.47896393100906826,
+      "eval_runtime": 0.3802,
+      "eval_samples_per_second": 2743.294,
+      "eval_steps_per_second": 173.593,
+      "step": 536
+    }
+  ],
+  "logging_steps": 500,
+  "max_steps": 536,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 4,
+  "save_steps": 500,
+  "total_flos": 209572221336984.0,
+  "train_batch_size": 64,
+  "trial_name": null,
+  "trial_params": {
+    "learning_rate": 1.1390296624220138e-05,
+    "num_train_epochs": 4,
+    "per_device_train_batch_size": 64,
+    "seed": 11
+  }
+}

run-5/checkpoint-536/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7b0a20b7e1ced1e066ae0be151da2892022f2d3caa0764af95b56a89f7778c56
+size 4984

run-8/checkpoint-535/config.json

@@ -0,0 +1,25 @@
+{
+  "_name_or_path": "distilbert-base-uncased",
+  "activation": "gelu",
+  "architectures": [
+    "DistilBertForSequenceClassification"
+  ],
+  "attention_dropout": 0.1,
+  "dim": 768,
+  "dropout": 0.1,
+  "hidden_dim": 3072,
+  "initializer_range": 0.02,
+  "max_position_embeddings": 512,
+  "model_type": "distilbert",
+  "n_heads": 12,
+  "n_layers": 6,
+  "pad_token_id": 0,
+  "problem_type": "single_label_classification",
+  "qa_dropout": 0.1,
+  "seq_classif_dropout": 0.2,
+  "sinusoidal_pos_embds": false,
+  "tie_weights_": true,
+  "torch_dtype": "float32",
+  "transformers_version": "4.38.2",
+  "vocab_size": 30522
+}

run-8/checkpoint-535/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:602c8c455abf7fad327e2d64c5bedf2ddfbae8013f9edc3c3826cd3c67ff6ed0
+size 267832560

run-8/checkpoint-535/optimizer.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:32e14950f4e0d02186e2a579cfa19af62b9873b0f7d05832d89c104787dc9153
+size 535727290

run-8/checkpoint-535/rng_state.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:629119240156f5bdd665de8b1edbd1247dc5d4322a8f37d92132ef93df16ea3a
+size 14244

run-8/checkpoint-535/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b92135e15e9c7a005c9e9b0c81f2a1edf72b5e5d56aa2f8d079558deecaacc2b
+size 1064

run-8/checkpoint-535/special_tokens_map.json

@@ -0,0 +1,7 @@
+{
+  "cls_token": "[CLS]",
+  "mask_token": "[MASK]",
+  "pad_token": "[PAD]",
+  "sep_token": "[SEP]",
+  "unk_token": "[UNK]"
+}

run-8/checkpoint-535/tokenizer_config.json

@@ -0,0 +1,55 @@
+{
+  "added_tokens_decoder": {
+    "0": {
+      "content": "[PAD]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "100": {
+      "content": "[UNK]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "101": {
+      "content": "[CLS]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "102": {
+      "content": "[SEP]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "103": {
+      "content": "[MASK]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "clean_up_tokenization_spaces": true,
+  "cls_token": "[CLS]",
+  "do_lower_case": true,
+  "mask_token": "[MASK]",
+  "model_max_length": 512,
+  "pad_token": "[PAD]",
+  "sep_token": "[SEP]",
+  "strip_accents": null,
+  "tokenize_chinese_chars": true,
+  "tokenizer_class": "DistilBertTokenizer",
+  "unk_token": "[UNK]"
+}

run-8/checkpoint-535/trainer_state.json

@@ -0,0 +1,42 @@
+{
+  "best_metric": 0.4340396426791047,
+  "best_model_checkpoint": "distilbert-base-uncased-finetuned-cola/run-8/checkpoint-535",
+  "epoch": 1.0,
+  "eval_steps": 500,
+  "global_step": 535,
+  "is_hyper_param_search": true,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.93,
+      "grad_norm": 8.78057861328125,
+      "learning_rate": 3.911277119513222e-06,
+      "loss": 0.529,
+      "step": 500
+    },
+    {
+      "epoch": 1.0,
+      "eval_loss": 0.47478926181793213,
+      "eval_matthews_correlation": 0.4340396426791047,
+      "eval_runtime": 0.3409,
+      "eval_samples_per_second": 3059.972,
+      "eval_steps_per_second": 193.632,
+      "step": 535
+    }
+  ],
+  "logging_steps": 500,
+  "max_steps": 535,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 1,
+  "save_steps": 500,
+  "total_flos": 45196996755108.0,
+  "train_batch_size": 16,
+  "trial_name": null,
+  "trial_params": {
+    "learning_rate": 5.978666454113069e-05,
+    "num_train_epochs": 1,
+    "per_device_train_batch_size": 16,
+    "seed": 15
+  }
+}

run-8/checkpoint-535/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:41134123e6324d2297988ee9a159f309008b9ed0e2c8676bc67bbdad4182508d
+size 4984

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:7b0a20b7e1ced1e066ae0be151da2892022f2d3caa0764af95b56a89f7778c56
+oid sha256:41134123e6324d2297988ee9a159f309008b9ed0e2c8676bc67bbdad4182508d
 size 4984