First model version

README.md

@@ -0,0 +1,21 @@
+---
+language:
+- de
+datasets:
+- gnad10
+tags:
+- text-classification
+- german-news-classification
+metrics:
+- accuracy
+- precision
+- recall
+- f1
+---
+
+# German BERT for News Classification
+
+This a bert-base-german-cased model finetuned for text classification on german news articles
+
+## Training data
+Used the training set from the 10KGNAD dataset (gnad10 on HuggingFace Datasets).

config.json

@@ -0,0 +1,48 @@
+{
+  "_name_or_path": "bert-base-german-cased",
+  "architectures": [
+    "BertForSequenceClassification"
+  ],
+  "attention_probs_dropout_prob": 0.1,
+  "classifier_dropout": null,
+  "hidden_act": "gelu",
+  "hidden_dropout_prob": 0.1,
+  "hidden_size": 768,
+  "id2label": {
+    "0": "Web",
+    "1": "Panorama",
+    "2": "International",
+    "3": "Wirtschaft",
+    "4": "Sport",
+    "5": "Inland",
+    "6": "Etat",
+    "7": "Wissenschaft",
+    "8": "Kultur"
+  },
+  "initializer_range": 0.02,
+  "intermediate_size": 3072,
+  "label2id": {
+    "Web": 0,
+    "Panorama": 1,
+    "International": 2,
+    "Wirtschaft": 3,
+    "Sport": 4,
+    "Inland": 5,
+    "Etat": 6,
+    "Wissenschaft": 7,
+    "Kultur": 8
+  },
+  "layer_norm_eps": 1e-12,
+  "max_position_embeddings": 512,
+  "model_type": "bert",
+  "num_attention_heads": 12,
+  "num_hidden_layers": 12,
+  "pad_token_id": 0,
+  "position_embedding_type": "absolute",
+  "problem_type": "single_label_classification",
+  "torch_dtype": "float32",
+  "transformers_version": "4.12.3",
+  "type_vocab_size": 2,
+  "use_cache": true,
+  "vocab_size": 30000
+}

optimizer.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6f11f628d9742108335fcd75f9b1f9bff7600caacf59dffe0447ae9348f124c5
+size 872822557

pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e52c7e41653b5f7390302f36f791b27014e49f1e944920e3aa1d650c25aebed1
+size 436437165

rng_state.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ef628c018f6f77bb58239b737a3ec5b81bf8b50e882733a3be54a37563dea230
+size 14503

scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:389f9571242e0c2895041392c8e2b216f997fd545ea3b92b7998348587d6d9ce
+size 623

special_tokens_map.json

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

tokenizer_config.json

@@ -0,0 +1 @@
+{"do_lower_case": false, "unk_token": "[UNK]", "sep_token": "[SEP]", "pad_token": "[PAD]", "cls_token": "[CLS]", "mask_token": "[MASK]", "tokenize_chinese_chars": true, "strip_accents": null, "model_max_length": 512, "special_tokens_map_file": null, "name_or_path": "bert-base-german-cased", "tokenizer_class": "BertTokenizer"}

trainer_state.json

@@ -0,0 +1,254 @@
+{
+  "best_metric": 0.40839770436286926,
+  "best_model_checkpoint": "/content/drive/MyDrive/colab/checkpoint-1700",
+  "epoch": 1.6346153846153846,
+  "global_step": 1700,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.7596153846153844e-05,
+      "loss": 1.033,
+      "step": 100
+    },
+    {
+      "epoch": 0.1,
+      "eval_loss": 0.6874200105667114,
+      "eval_runtime": 57.2656,
+      "eval_samples_per_second": 16.153,
+      "eval_steps_per_second": 2.026,
+      "step": 100
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.519230769230769e-05,
+      "loss": 0.6703,
+      "step": 200
+    },
+    {
+      "epoch": 0.19,
+      "eval_loss": 0.5938639044761658,
+      "eval_runtime": 57.2626,
+      "eval_samples_per_second": 16.154,
+      "eval_steps_per_second": 2.026,
+      "step": 200
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 4.278846153846154e-05,
+      "loss": 0.6579,
+      "step": 300
+    },
+    {
+      "epoch": 0.29,
+      "eval_loss": 0.49521076679229736,
+      "eval_runtime": 57.2701,
+      "eval_samples_per_second": 16.152,
+      "eval_steps_per_second": 2.025,
+      "step": 300
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.038461538461539e-05,
+      "loss": 0.5068,
+      "step": 400
+    },
+    {
+      "epoch": 0.38,
+      "eval_loss": 0.5416558384895325,
+      "eval_runtime": 57.2833,
+      "eval_samples_per_second": 16.148,
+      "eval_steps_per_second": 2.025,
+      "step": 400
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 3.798076923076923e-05,
+      "loss": 0.5439,
+      "step": 500
+    },
+    {
+      "epoch": 0.48,
+      "eval_loss": 0.5001481175422668,
+      "eval_runtime": 57.2802,
+      "eval_samples_per_second": 16.149,
+      "eval_steps_per_second": 2.025,
+      "step": 500
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 3.557692307692308e-05,
+      "loss": 0.5429,
+      "step": 600
+    },
+    {
+      "epoch": 0.58,
+      "eval_loss": 0.47204354405403137,
+      "eval_runtime": 57.2593,
+      "eval_samples_per_second": 16.155,
+      "eval_steps_per_second": 2.026,
+      "step": 600
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 3.3173076923076926e-05,
+      "loss": 0.4817,
+      "step": 700
+    },
+    {
+      "epoch": 0.67,
+      "eval_loss": 0.5970540642738342,
+      "eval_runtime": 57.2593,
+      "eval_samples_per_second": 16.155,
+      "eval_steps_per_second": 2.026,
+      "step": 700
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 3.0769230769230774e-05,
+      "loss": 0.5023,
+      "step": 800
+    },
+    {
+      "epoch": 0.77,
+      "eval_loss": 0.44290876388549805,
+      "eval_runtime": 57.2393,
+      "eval_samples_per_second": 16.16,
+      "eval_steps_per_second": 2.027,
+      "step": 800
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 2.8365384615384616e-05,
+      "loss": 0.4756,
+      "step": 900
+    },
+    {
+      "epoch": 0.87,
+      "eval_loss": 0.49230891466140747,
+      "eval_runtime": 57.0994,
+      "eval_samples_per_second": 16.2,
+      "eval_steps_per_second": 2.032,
+      "step": 900
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 2.5961538461538464e-05,
+      "loss": 0.4772,
+      "step": 1000
+    },
+    {
+      "epoch": 0.96,
+      "eval_loss": 0.4519612491130829,
+      "eval_runtime": 57.1095,
+      "eval_samples_per_second": 16.197,
+      "eval_steps_per_second": 2.031,
+      "step": 1000
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 2.355769230769231e-05,
+      "loss": 0.336,
+      "step": 1100
+    },
+    {
+      "epoch": 1.06,
+      "eval_loss": 0.41091373562812805,
+      "eval_runtime": 57.1093,
+      "eval_samples_per_second": 16.197,
+      "eval_steps_per_second": 2.031,
+      "step": 1100
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 2.1153846153846154e-05,
+      "loss": 0.2225,
+      "step": 1200
+    },
+    {
+      "epoch": 1.15,
+      "eval_loss": 0.4540548324584961,
+      "eval_runtime": 57.116,
+      "eval_samples_per_second": 16.195,
+      "eval_steps_per_second": 2.031,
+      "step": 1200
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 1.8750000000000002e-05,
+      "loss": 0.2296,
+      "step": 1300
+    },
+    {
+      "epoch": 1.25,
+      "eval_loss": 0.4974042475223541,
+      "eval_runtime": 57.1073,
+      "eval_samples_per_second": 16.198,
+      "eval_steps_per_second": 2.031,
+      "step": 1300
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 1.6346153846153847e-05,
+      "loss": 0.2511,
+      "step": 1400
+    },
+    {
+      "epoch": 1.35,
+      "eval_loss": 0.48081809282302856,
+      "eval_runtime": 57.1059,
+      "eval_samples_per_second": 16.198,
+      "eval_steps_per_second": 2.031,
+      "step": 1400
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 1.3942307692307693e-05,
+      "loss": 0.3349,
+      "step": 1500
+    },
+    {
+      "epoch": 1.44,
+      "eval_loss": 0.4536380171775818,
+      "eval_runtime": 57.096,
+      "eval_samples_per_second": 16.201,
+      "eval_steps_per_second": 2.032,
+      "step": 1500
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 1.153846153846154e-05,
+      "loss": 0.2529,
+      "step": 1600
+    },
+    {
+      "epoch": 1.54,
+      "eval_loss": 0.4738934636116028,
+      "eval_runtime": 57.1337,
+      "eval_samples_per_second": 16.19,
+      "eval_steps_per_second": 2.03,
+      "step": 1600
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 9.134615384615384e-06,
+      "loss": 0.2064,
+      "step": 1700
+    },
+    {
+      "epoch": 1.63,
+      "eval_loss": 0.40839770436286926,
+      "eval_runtime": 57.1173,
+      "eval_samples_per_second": 16.195,
+      "eval_steps_per_second": 2.031,
+      "step": 1700
+    }
+  ],
+  "max_steps": 2080,
+  "num_train_epochs": 2,
+  "total_flos": 3578535250329600.0,
+  "trial_name": null,
+  "trial_params": null
+}

training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c409fc8b117df90c48570ef0a0610bedf911537d8327c83194ae77411c61dfce
+size 2863