commit from mujeen

README.md

@@ -0,0 +1,77 @@
+---
+language:
+- en
+license: mit
+tags:
+- generated_from_trainer
+datasets:
+- glue
+metrics:
+- accuracy
+model-index:
+- name: roberta-base_mnli_bc
+  results:
+  - task:
+      name: Text Classification
+      type: text-classification
+    dataset:
+      name: GLUE MNLI
+      type: glue
+      args: mnli
+    metrics:
+    - name: Accuracy
+      type: accuracy
+      value: 0.9583768461882739
+---
+
+<!-- This model card has been generated automatically according to the information the Trainer had access to. You
+should probably proofread and complete it, then remove this comment. -->
+
+# roberta-base_mnli_bc
+
+This model is a fine-tuned version of [roberta-base](https://huggingface.co/roberta-base) on the GLUE MNLI dataset.
+It achieves the following results on the evaluation set:
+- Loss: 0.2125
+- Accuracy: 0.9584
+
+## Model description
+
+More information needed
+
+## Intended uses & limitations
+
+More information needed
+
+## Training and evaluation data
+
+More information needed
+
+## Training procedure
+
+### Training hyperparameters
+
+The following hyperparameters were used during training:
+- learning_rate: 2e-05
+- train_batch_size: 16
+- eval_batch_size: 8
+- seed: 42
+- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
+- lr_scheduler_type: linear
+- num_epochs: 3.0
+- mixed_precision_training: Native AMP
+
+### Training results
+
+| Training Loss | Epoch | Step  | Validation Loss | Accuracy |
+|:-------------:|:-----:|:-----:|:---------------:|:--------:|
+| 0.2015        | 1.0   | 16363 | 0.1820          | 0.9470   |
+| 0.1463        | 2.0   | 32726 | 0.1909          | 0.9559   |
+| 0.0768        | 3.0   | 49089 | 0.2117          | 0.9585   |
+
+
+### Framework versions
+
+- Transformers 4.13.0
+- Pytorch 1.10.1+cu111
+- Datasets 1.17.0
+- Tokenizers 0.10.3

all_results.json

@@ -0,0 +1,14 @@
+{
+    "epoch": 3.0,
+    "eval_accuracy": 0.9583768461882739,
+    "eval_loss": 0.2125200629234314,
+    "eval_runtime": 18.5055,
+    "eval_samples": 6703,
+    "eval_samples_per_second": 362.217,
+    "eval_steps_per_second": 45.284,
+    "train_loss": 0.1589348805035952,
+    "train_runtime": 5162.793,
+    "train_samples": 261802,
+    "train_samples_per_second": 152.128,
+    "train_steps_per_second": 9.508
+}

config.json

@@ -0,0 +1,39 @@
+{
+  "_name_or_path": "roberta-base",
+  "architectures": [
+    "RobertaForSequenceClassification"
+  ],
+  "attention_probs_dropout_prob": 0.1,
+  "bos_token_id": 0,
+  "classifier_dropout": null,
+  "eos_token_id": 2,
+  "finetuning_task": "mnli",
+  "hidden_act": "gelu",
+  "hidden_dropout_prob": 0.1,
+  "hidden_size": 768,
+  "id2label": {
+    "0": "entailment",
+    "1": "neutral",
+    "2": "contradiction"
+  },
+  "initializer_range": 0.02,
+  "intermediate_size": 3072,
+  "label2id": {
+    "contradiction": 2,
+    "entailment": 0,
+    "neutral": 1
+  },
+  "layer_norm_eps": 1e-05,
+  "max_position_embeddings": 514,
+  "model_type": "roberta",
+  "num_attention_heads": 12,
+  "num_hidden_layers": 12,
+  "pad_token_id": 1,
+  "position_embedding_type": "absolute",
+  "problem_type": "single_label_classification",
+  "torch_dtype": "float32",
+  "transformers_version": "4.13.0",
+  "type_vocab_size": 1,
+  "use_cache": true,
+  "vocab_size": 50265
+}

eval_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 3.0,
+    "eval_accuracy": 0.9583768461882739,
+    "eval_loss": 0.2125200629234314,
+    "eval_runtime": 18.5055,
+    "eval_samples": 6703,
+    "eval_samples_per_second": 362.217,
+    "eval_steps_per_second": 45.284
+}

pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ababe75115c17ee4e75582d689eff69370198676a84d139a0bce7a34ed7830d6
+size 498674093

special_tokens_map.json

@@ -0,0 +1 @@
+{"bos_token": "<s>", "eos_token": "</s>", "unk_token": "<unk>", "sep_token": "</s>", "pad_token": "<pad>", "cls_token": "<s>", "mask_token": {"content": "<mask>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": false}}

tokenizer_config.json

@@ -0,0 +1 @@
+{"unk_token": "<unk>", "bos_token": "<s>", "eos_token": "</s>", "add_prefix_space": false, "errors": "replace", "sep_token": "</s>", "cls_token": "<s>", "pad_token": "<pad>", "mask_token": "<mask>", "model_max_length": 512, "special_tokens_map_file": null, "name_or_path": "roberta-base", "tokenizer_class": "RobertaTokenizer"}

train_results.json

@@ -0,0 +1,8 @@
+{
+    "epoch": 3.0,
+    "train_loss": 0.1589348805035952,
+    "train_runtime": 5162.793,
+    "train_samples": 261802,
+    "train_samples_per_second": 152.128,
+    "train_steps_per_second": 9.508
+}

trainer_state.json

@@ -0,0 +1,640 @@
+{
+  "best_metric": 0.9584578601315002,
+  "best_model_checkpoint": "/hdd1/mujeen/retrieval_prf/output/roberta-base_mnli_bc/checkpoint-49089",
+  "epoch": 3.0,
+  "global_step": 49089,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.9797918067184097e-05,
+      "loss": 0.3858,
+      "step": 500
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 1.959420644136161e-05,
+      "loss": 0.284,
+      "step": 1000
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.9390494815539124e-05,
+      "loss": 0.2536,
+      "step": 1500
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.918719061296828e-05,
+      "loss": 0.2583,
+      "step": 2000
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.89834789871458e-05,
+      "loss": 0.2522,
+      "step": 2500
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.8779767361323312e-05,
+      "loss": 0.2526,
+      "step": 3000
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.8576055735500826e-05,
+      "loss": 0.2457,
+      "step": 3500
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.8372344109678343e-05,
+      "loss": 0.2335,
+      "step": 4000
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.8168632483855853e-05,
+      "loss": 0.224,
+      "step": 4500
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.796492085803337e-05,
+      "loss": 0.2234,
+      "step": 5000
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.7761209232210884e-05,
+      "loss": 0.2335,
+      "step": 5500
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.7557905029640045e-05,
+      "loss": 0.2341,
+      "step": 6000
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.7354193403817555e-05,
+      "loss": 0.2235,
+      "step": 6500
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.7150481777995072e-05,
+      "loss": 0.231,
+      "step": 7000
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.6946770152172586e-05,
+      "loss": 0.2146,
+      "step": 7500
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.67430585263501e-05,
+      "loss": 0.219,
+      "step": 8000
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 1.653975432377926e-05,
+      "loss": 0.2263,
+      "step": 8500
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 1.6336042697956774e-05,
+      "loss": 0.2197,
+      "step": 9000
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 1.6132331072134288e-05,
+      "loss": 0.2072,
+      "step": 9500
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 1.59286194463118e-05,
+      "loss": 0.2196,
+      "step": 10000
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 1.5724907820489318e-05,
+      "loss": 0.2039,
+      "step": 10500
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 1.5521603617918476e-05,
+      "loss": 0.201,
+      "step": 11000
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 1.531789199209599e-05,
+      "loss": 0.2081,
+      "step": 11500
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 1.5114180366273505e-05,
+      "loss": 0.2072,
+      "step": 12000
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 1.4910468740451018e-05,
+      "loss": 0.1949,
+      "step": 12500
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 1.4706757114628534e-05,
+      "loss": 0.2119,
+      "step": 13000
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 1.4503452912057691e-05,
+      "loss": 0.209,
+      "step": 13500
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.4299741286235206e-05,
+      "loss": 0.1992,
+      "step": 14000
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 1.409602966041272e-05,
+      "loss": 0.1953,
+      "step": 14500
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 1.3892318034590235e-05,
+      "loss": 0.2025,
+      "step": 15000
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.368860640876775e-05,
+      "loss": 0.198,
+      "step": 15500
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 1.3485302206196908e-05,
+      "loss": 0.2015,
+      "step": 16000
+    },
+    {
+      "epoch": 1.0,
+      "eval_accuracy": 0.9469515839808726,
+      "eval_loss": 0.18203075230121613,
+      "eval_runtime": 18.9219,
+      "eval_samples_per_second": 353.664,
+      "eval_steps_per_second": 44.234,
+      "step": 16363
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 1.3281590580374423e-05,
+      "loss": 0.1932,
+      "step": 16500
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 1.3077878954551937e-05,
+      "loss": 0.1541,
+      "step": 17000
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 1.2874574751981098e-05,
+      "loss": 0.1556,
+      "step": 17500
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 1.267086312615861e-05,
+      "loss": 0.1607,
+      "step": 18000
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 1.2467151500336125e-05,
+      "loss": 0.1618,
+      "step": 18500
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 1.2263439874513639e-05,
+      "loss": 0.1593,
+      "step": 19000
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 1.2059728248691154e-05,
+      "loss": 0.1511,
+      "step": 19500
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 1.185601662286867e-05,
+      "loss": 0.165,
+      "step": 20000
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 1.1652304997046182e-05,
+      "loss": 0.157,
+      "step": 20500
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 1.1448593371223697e-05,
+      "loss": 0.1385,
+      "step": 21000
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 1.1245289168652856e-05,
+      "loss": 0.1646,
+      "step": 21500
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 1.1041577542830371e-05,
+      "loss": 0.163,
+      "step": 22000
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 1.0837865917007883e-05,
+      "loss": 0.156,
+      "step": 22500
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 1.0634154291185399e-05,
+      "loss": 0.1556,
+      "step": 23000
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 1.043085008861456e-05,
+      "loss": 0.1626,
+      "step": 23500
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 1.0227545886043717e-05,
+      "loss": 0.1605,
+      "step": 24000
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 1.0023834260221232e-05,
+      "loss": 0.1651,
+      "step": 24500
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 9.820122634398746e-06,
+      "loss": 0.1438,
+      "step": 25000
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 9.616411008576261e-06,
+      "loss": 0.1501,
+      "step": 25500
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 9.412699382753775e-06,
+      "loss": 0.1562,
+      "step": 26000
+    },
+    {
+      "epoch": 1.62,
+      "learning_rate": 9.209395180182934e-06,
+      "loss": 0.158,
+      "step": 26500
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 9.005683554360448e-06,
+      "loss": 0.149,
+      "step": 27000
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.801971928537963e-06,
+      "loss": 0.1483,
+      "step": 27500
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.598260302715477e-06,
+      "loss": 0.163,
+      "step": 28000
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 8.39454867689299e-06,
+      "loss": 0.1507,
+      "step": 28500
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 8.19124447432215e-06,
+      "loss": 0.148,
+      "step": 29000
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.987532848499665e-06,
+      "loss": 0.1571,
+      "step": 29500
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.78382122267718e-06,
+      "loss": 0.1467,
+      "step": 30000
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 7.580109596854694e-06,
+      "loss": 0.1417,
+      "step": 30500
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 7.376397971032207e-06,
+      "loss": 0.1526,
+      "step": 31000
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 7.172686345209722e-06,
+      "loss": 0.1327,
+      "step": 31500
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 6.969382142638881e-06,
+      "loss": 0.151,
+      "step": 32000
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 6.7656705168163946e-06,
+      "loss": 0.1463,
+      "step": 32500
+    },
+    {
+      "epoch": 2.0,
+      "eval_accuracy": 0.9559175134488942,
+      "eval_loss": 0.19088995456695557,
+      "eval_runtime": 18.8218,
+      "eval_samples_per_second": 355.545,
+      "eval_steps_per_second": 44.47,
+      "step": 32726
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 6.561958890993909e-06,
+      "loss": 0.1198,
+      "step": 33000
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 6.358247265171424e-06,
+      "loss": 0.0851,
+      "step": 33500
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 6.155350485852228e-06,
+      "loss": 0.1081,
+      "step": 34000
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 5.9516388600297426e-06,
+      "loss": 0.1071,
+      "step": 34500
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 5.747927234207257e-06,
+      "loss": 0.1046,
+      "step": 35000
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 5.544215608384771e-06,
+      "loss": 0.0965,
+      "step": 35500
+    },
+    {
+      "epoch": 2.2,
+      "learning_rate": 5.340503982562285e-06,
+      "loss": 0.0906,
+      "step": 36000
+    },
+    {
+      "epoch": 2.23,
+      "learning_rate": 5.137199779991444e-06,
+      "loss": 0.0927,
+      "step": 36500
+    },
+    {
+      "epoch": 2.26,
+      "learning_rate": 4.933488154168959e-06,
+      "loss": 0.0947,
+      "step": 37000
+    },
+    {
+      "epoch": 2.29,
+      "learning_rate": 4.729776528346473e-06,
+      "loss": 0.1007,
+      "step": 37500
+    },
+    {
+      "epoch": 2.32,
+      "learning_rate": 4.526064902523987e-06,
+      "loss": 0.0981,
+      "step": 38000
+    },
+    {
+      "epoch": 2.35,
+      "learning_rate": 4.322353276701502e-06,
+      "loss": 0.105,
+      "step": 38500
+    },
+    {
+      "epoch": 2.38,
+      "learning_rate": 4.118641650879016e-06,
+      "loss": 0.1025,
+      "step": 39000
+    },
+    {
+      "epoch": 2.41,
+      "learning_rate": 3.915337448308175e-06,
+      "loss": 0.089,
+      "step": 39500
+    },
+    {
+      "epoch": 2.44,
+      "learning_rate": 3.7116258224856896e-06,
+      "loss": 0.1049,
+      "step": 40000
+    },
+    {
+      "epoch": 2.48,
+      "learning_rate": 3.507914196663204e-06,
+      "loss": 0.0982,
+      "step": 40500
+    },
+    {
+      "epoch": 2.51,
+      "learning_rate": 3.304202570840718e-06,
+      "loss": 0.1007,
+      "step": 41000
+    },
+    {
+      "epoch": 2.54,
+      "learning_rate": 3.1004909450182323e-06,
+      "loss": 0.098,
+      "step": 41500
+    },
+    {
+      "epoch": 2.57,
+      "learning_rate": 2.8967793191957468e-06,
+      "loss": 0.0889,
+      "step": 42000
+    },
+    {
+      "epoch": 2.6,
+      "learning_rate": 2.693067693373261e-06,
+      "loss": 0.0975,
+      "step": 42500
+    },
+    {
+      "epoch": 2.63,
+      "learning_rate": 2.4893560675507754e-06,
+      "loss": 0.1031,
+      "step": 43000
+    },
+    {
+      "epoch": 2.66,
+      "learning_rate": 2.28564444172829e-06,
+      "loss": 0.0882,
+      "step": 43500
+    },
+    {
+      "epoch": 2.69,
+      "learning_rate": 2.082747662409094e-06,
+      "loss": 0.0917,
+      "step": 44000
+    },
+    {
+      "epoch": 2.72,
+      "learning_rate": 1.879036036586608e-06,
+      "loss": 0.0935,
+      "step": 44500
+    },
+    {
+      "epoch": 2.75,
+      "learning_rate": 1.6753244107641225e-06,
+      "loss": 0.099,
+      "step": 45000
+    },
+    {
+      "epoch": 2.78,
+      "learning_rate": 1.4716127849416368e-06,
+      "loss": 0.0938,
+      "step": 45500
+    },
+    {
+      "epoch": 2.81,
+      "learning_rate": 1.2679011591191509e-06,
+      "loss": 0.0977,
+      "step": 46000
+    },
+    {
+      "epoch": 2.84,
+      "learning_rate": 1.0641895332966654e-06,
+      "loss": 0.0785,
+      "step": 46500
+    },
+    {
+      "epoch": 2.87,
+      "learning_rate": 8.604779074741797e-07,
+      "loss": 0.106,
+      "step": 47000
+    },
+    {
+      "epoch": 2.9,
+      "learning_rate": 6.571737049033389e-07,
+      "loss": 0.088,
+      "step": 47500
+    },
+    {
+      "epoch": 2.93,
+      "learning_rate": 4.5346207908085317e-07,
+      "loss": 0.087,
+      "step": 48000
+    },
+    {
+      "epoch": 2.96,
+      "learning_rate": 2.4975045325836747e-07,
+      "loss": 0.1096,
+      "step": 48500
+    },
+    {
+      "epoch": 2.99,
+      "learning_rate": 4.6038827435881766e-08,
+      "loss": 0.0768,
+      "step": 49000
+    },
+    {
+      "epoch": 3.0,
+      "eval_accuracy": 0.9584578601315002,
+      "eval_loss": 0.21173284947872162,
+      "eval_runtime": 18.7035,
+      "eval_samples_per_second": 357.794,
+      "eval_steps_per_second": 44.751,
+      "step": 49089
+    },
+    {
+      "epoch": 3.0,
+      "step": 49089,
+      "total_flos": 3.330874138659072e+16,
+      "train_loss": 0.1589348805035952,
+      "train_runtime": 5162.793,
+      "train_samples_per_second": 152.128,
+      "train_steps_per_second": 9.508
+    }
+  ],
+  "max_steps": 49089,
+  "num_train_epochs": 3,
+  "total_flos": 3.330874138659072e+16,
+  "trial_name": null,
+  "trial_params": null
+}

training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9bd94d35ba146b6ad3fb214b5d4aa59a0a9c8d5da7739cbcf3d07bea21c48e55
+size 2991