Upload folder using huggingface_hub

.gitignore

@@ -0,0 +1,2 @@
+checkpoint-*/
+.ipynb_checkpoints

README.md

@@ -0,0 +1,145 @@
+---
+library_name: transformers
+language:
+- en
+license: mit
+base_model: FacebookAI/roberta-base
+tags:
+- generated_from_trainer
+datasets:
+- swag
+metrics:
+- accuracy
+model-index:
+- name: swag_base
+  results:
+  - task:
+      name: Multiple Choice
+      type: multiple-choice
+    dataset:
+      name: SWAG
+      type: swag
+      args: regular
+    metrics:
+    - name: Accuracy
+      type: accuracy
+      value: 0.7521243691444397
+---
+
+# swag_base
+
+This model is a fine-tuned version of [FacebookAI/roberta-base](https://huggingface.co/FacebookAI/roberta-base) on the SWAG (Situations With Adversarial Generations) dataset.
+
+## Model description
+
+The model is designed to perform multiple-choice reasoning about real-world situations. Given a context and four possible continuations, it predicts the most plausible ending based on common sense understanding.
+
+Key Features:
+- Base model: RoBERTa-base
+- Task: Multiple Choice Prediction
+- Training dataset: SWAG
+- Performance: 75.21% accuracy on evaluation set
+
+## Training Procedure
+
+### Training hyperparameters
+- Learning rate: 5e-05
+- Batch size: 16
+- Number of epochs: 3
+- Optimizer: AdamW
+- Learning rate scheduler: Linear
+- Training samples: 73,546
+- Training time: 17m 53s
+
+### Training Results
+- Training loss: 0.73
+- Evaluation loss: 0.7362
+- Evaluation accuracy: 0.7521
+- Training samples/second: 205.623
+- Training steps/second: 12.852
+
+## Usage Example
+
+Here's how to use the model:
+
+```python
+from transformers import AutoTokenizer, AutoModelForMultipleChoice
+import torch
+
+# Load model and tokenizer
+model_path = "real-jiakai/roberta-base-uncased-finetuned-swag"
+tokenizer = AutoTokenizer.from_pretrained(model_path)
+model = AutoModelForMultipleChoice.from_pretrained(model_path)
+
+# Example scenarios
+test_examples = [
+    {
+        'context': "Stephen Curry dribbles the ball at the three-point line",
+        'endings': [
+            "He quickly releases a perfect shot that swishes through the net",  # Most plausible
+            "He suddenly starts dancing ballet on the court",
+            "He transforms the basketball into a pizza",
+            "He flies to the moon with the basketball"
+        ]
+    },
+    {
+        'context': "Elon Musk walks into a SpaceX facility and looks at a rocket",
+        'endings': [
+            "He discusses technical details with the engineering team",  # Most plausible
+            "He turns the rocket into a giant chocolate bar",
+            "He starts playing basketball with the rocket",
+            "He teaches the rocket to speak French"
+        ]
+    }
+]
+
+def predict_swag(context, endings, model, tokenizer):
+    encoding = tokenizer(
+        [context] * 4,
+        endings,
+        truncation=True,
+        max_length=128,
+        padding="max_length",
+        return_tensors="pt"
+    )
+    
+    input_ids = encoding['input_ids'].unsqueeze(0)
+    attention_mask = encoding['attention_mask'].unsqueeze(0)
+    
+    outputs = model(input_ids=input_ids, attention_mask=attention_mask)
+    logits = outputs.logits
+    
+    predicted_idx = torch.argmax(logits).item()
+    
+    return {
+        'context': context,
+        'predicted_ending': endings[predicted_idx],
+        'probabilities': torch.softmax(logits, dim=1)[0].tolist()
+    }
+```
+
+## Limitations and Biases
+
+The model's performance is limited by its training data and may not generalize well to all domains
+Performance might vary depending on the complexity and domain of the input scenarios
+The model may exhibit biases present in the training data
+
+## Framework versions
+
+Transformers 4.47.0.dev0
+PyTorch 2.5.1+cu124
+Datasets 3.1.0
+Tokenizers 0.20.3
+
+## Citation
+
+If you use this model, please cite:
+
+```
+@inproceedings{zellers2018swagaf,
+    title={SWAG: A Large-Scale Adversarial Dataset for Grounded Commonsense Inference},
+    author={Zellers, Rowan and Bisk, Yonatan and Schwartz, Roy and Choi, Yejin},
+    booktitle = "Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing (EMNLP)",
+    year={2018}
+}
+```

all_results.json

@@ -0,0 +1,15 @@
+{
+    "epoch": 3.0,
+    "eval_accuracy": 0.7521243691444397,
+    "eval_loss": 0.7361685037612915,
+    "eval_runtime": 28.3884,
+    "eval_samples": 20006,
+    "eval_samples_per_second": 704.725,
+    "eval_steps_per_second": 44.067,
+    "total_flos": 2.508085260375571e+16,
+    "train_loss": 0.7300247143699852,
+    "train_runtime": 1073.0221,
+    "train_samples": 73546,
+    "train_samples_per_second": 205.623,
+    "train_steps_per_second": 12.852
+}

config.json

@@ -0,0 +1,27 @@
+{
+  "_name_or_path": "FacebookAI/roberta-base",
+  "architectures": [
+    "RobertaForMultipleChoice"
+  ],
+  "attention_probs_dropout_prob": 0.1,
+  "bos_token_id": 0,
+  "classifier_dropout": null,
+  "eos_token_id": 2,
+  "hidden_act": "gelu",
+  "hidden_dropout_prob": 0.1,
+  "hidden_size": 768,
+  "initializer_range": 0.02,
+  "intermediate_size": 3072,
+  "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",
+  "torch_dtype": "float32",
+  "transformers_version": "4.47.0.dev0",
+  "type_vocab_size": 1,
+  "use_cache": true,
+  "vocab_size": 50265
+}

eval_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 3.0,
+    "eval_accuracy": 0.7521243691444397,
+    "eval_loss": 0.7361685037612915,
+    "eval_runtime": 28.3884,
+    "eval_samples": 20006,
+    "eval_samples_per_second": 704.725,
+    "eval_steps_per_second": 44.067
+}

model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b639335e7e79363a1eb5ae188010faec74ac19c1f3fc70c6b41da563f6ac74a1
+size 498609724

special_tokens_map.json

@@ -0,0 +1,15 @@
+{
+  "bos_token": "<s>",
+  "cls_token": "<s>",
+  "eos_token": "</s>",
+  "mask_token": {
+    "content": "<mask>",
+    "lstrip": true,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": "<pad>",
+  "sep_token": "</s>",
+  "unk_token": "<unk>"
+}

tokenizer_config.json

@@ -0,0 +1,58 @@
+{
+  "add_prefix_space": false,
+  "added_tokens_decoder": {
+    "0": {
+      "content": "<s>",
+      "lstrip": false,
+      "normalized": true,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "1": {
+      "content": "<pad>",
+      "lstrip": false,
+      "normalized": true,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "2": {
+      "content": "</s>",
+      "lstrip": false,
+      "normalized": true,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "3": {
+      "content": "<unk>",
+      "lstrip": false,
+      "normalized": true,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "50264": {
+      "content": "<mask>",
+      "lstrip": true,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "bos_token": "<s>",
+  "clean_up_tokenization_spaces": false,
+  "cls_token": "<s>",
+  "eos_token": "</s>",
+  "errors": "replace",
+  "extra_special_tokens": {},
+  "mask_token": "<mask>",
+  "model_max_length": 512,
+  "pad_token": "<pad>",
+  "sep_token": "</s>",
+  "tokenizer_class": "RobertaTokenizer",
+  "trim_offsets": true,
+  "unk_token": "<unk>"
+}

train_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 3.0,
+    "total_flos": 2.508085260375571e+16,
+    "train_loss": 0.7300247143699852,
+    "train_runtime": 1073.0221,
+    "train_samples": 73546,
+    "train_samples_per_second": 205.623,
+    "train_steps_per_second": 12.852
+}

trainer_state.json

@@ -0,0 +1,231 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 3.0,
+  "eval_steps": 500,
+  "global_step": 13791,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.10876658690450294,
+      "grad_norm": 7.74190092086792,
+      "learning_rate": 4.818722355159162e-05,
+      "loss": 1.3121,
+      "step": 500
+    },
+    {
+      "epoch": 0.2175331738090059,
+      "grad_norm": 21.71584129333496,
+      "learning_rate": 4.637444710318324e-05,
+      "loss": 1.1551,
+      "step": 1000
+    },
+    {
+      "epoch": 0.3262997607135088,
+      "grad_norm": 14.698501586914062,
+      "learning_rate": 4.4561670654774854e-05,
+      "loss": 1.0899,
+      "step": 1500
+    },
+    {
+      "epoch": 0.4350663476180118,
+      "grad_norm": 12.033799171447754,
+      "learning_rate": 4.2748894206366476e-05,
+      "loss": 1.054,
+      "step": 2000
+    },
+    {
+      "epoch": 0.5438329345225147,
+      "grad_norm": 8.601971626281738,
+      "learning_rate": 4.093611775795809e-05,
+      "loss": 1.015,
+      "step": 2500
+    },
+    {
+      "epoch": 0.6525995214270176,
+      "grad_norm": 12.497455596923828,
+      "learning_rate": 3.912334130954971e-05,
+      "loss": 0.9868,
+      "step": 3000
+    },
+    {
+      "epoch": 0.7613661083315205,
+      "grad_norm": 17.868608474731445,
+      "learning_rate": 3.731056486114133e-05,
+      "loss": 0.9785,
+      "step": 3500
+    },
+    {
+      "epoch": 0.8701326952360235,
+      "grad_norm": 9.991225242614746,
+      "learning_rate": 3.549778841273294e-05,
+      "loss": 0.9544,
+      "step": 4000
+    },
+    {
+      "epoch": 0.9788992821405265,
+      "grad_norm": 30.770978927612305,
+      "learning_rate": 3.368501196432456e-05,
+      "loss": 0.9245,
+      "step": 4500
+    },
+    {
+      "epoch": 1.0876658690450294,
+      "grad_norm": 23.796464920043945,
+      "learning_rate": 3.187223551591618e-05,
+      "loss": 0.7713,
+      "step": 5000
+    },
+    {
+      "epoch": 1.1964324559495323,
+      "grad_norm": 24.789031982421875,
+      "learning_rate": 3.0059459067507794e-05,
+      "loss": 0.7294,
+      "step": 5500
+    },
+    {
+      "epoch": 1.3051990428540352,
+      "grad_norm": 18.538000106811523,
+      "learning_rate": 2.8246682619099413e-05,
+      "loss": 0.7132,
+      "step": 6000
+    },
+    {
+      "epoch": 1.4139656297585381,
+      "grad_norm": 15.591426849365234,
+      "learning_rate": 2.643390617069103e-05,
+      "loss": 0.7076,
+      "step": 6500
+    },
+    {
+      "epoch": 1.5227322166630413,
+      "grad_norm": 22.807159423828125,
+      "learning_rate": 2.462112972228265e-05,
+      "loss": 0.7153,
+      "step": 7000
+    },
+    {
+      "epoch": 1.6314988035675442,
+      "grad_norm": 11.409658432006836,
+      "learning_rate": 2.2808353273874268e-05,
+      "loss": 0.6954,
+      "step": 7500
+    },
+    {
+      "epoch": 1.740265390472047,
+      "grad_norm": 20.154701232910156,
+      "learning_rate": 2.0995576825465886e-05,
+      "loss": 0.707,
+      "step": 8000
+    },
+    {
+      "epoch": 1.84903197737655,
+      "grad_norm": 17.02855110168457,
+      "learning_rate": 1.91828003770575e-05,
+      "loss": 0.6715,
+      "step": 8500
+    },
+    {
+      "epoch": 1.957798564281053,
+      "grad_norm": 19.347368240356445,
+      "learning_rate": 1.737002392864912e-05,
+      "loss": 0.6609,
+      "step": 9000
+    },
+    {
+      "epoch": 2.066565151185556,
+      "grad_norm": 18.68759536743164,
+      "learning_rate": 1.5557247480240738e-05,
+      "loss": 0.5354,
+      "step": 9500
+    },
+    {
+      "epoch": 2.1753317380900588,
+      "grad_norm": 15.24686050415039,
+      "learning_rate": 1.3744471031832355e-05,
+      "loss": 0.4594,
+      "step": 10000
+    },
+    {
+      "epoch": 2.2840983249945617,
+      "grad_norm": 32.36734390258789,
+      "learning_rate": 1.1931694583423973e-05,
+      "loss": 0.453,
+      "step": 10500
+    },
+    {
+      "epoch": 2.3928649118990646,
+      "grad_norm": 19.55815315246582,
+      "learning_rate": 1.011891813501559e-05,
+      "loss": 0.4465,
+      "step": 11000
+    },
+    {
+      "epoch": 2.5016314988035675,
+      "grad_norm": 16.438173294067383,
+      "learning_rate": 8.306141686607208e-06,
+      "loss": 0.4226,
+      "step": 11500
+    },
+    {
+      "epoch": 2.6103980857080704,
+      "grad_norm": 21.047183990478516,
+      "learning_rate": 6.493365238198826e-06,
+      "loss": 0.4324,
+      "step": 12000
+    },
+    {
+      "epoch": 2.7191646726125733,
+      "grad_norm": 29.30052947998047,
+      "learning_rate": 4.680588789790443e-06,
+      "loss": 0.4306,
+      "step": 12500
+    },
+    {
+      "epoch": 2.8279312595170762,
+      "grad_norm": 7.3705010414123535,
+      "learning_rate": 2.8678123413820606e-06,
+      "loss": 0.4371,
+      "step": 13000
+    },
+    {
+      "epoch": 2.936697846421579,
+      "grad_norm": 24.24944305419922,
+      "learning_rate": 1.0550358929736786e-06,
+      "loss": 0.4305,
+      "step": 13500
+    },
+    {
+      "epoch": 3.0,
+      "step": 13791,
+      "total_flos": 2.508085260375571e+16,
+      "train_loss": 0.7300247143699852,
+      "train_runtime": 1073.0221,
+      "train_samples_per_second": 205.623,
+      "train_steps_per_second": 12.852
+    }
+  ],
+  "logging_steps": 500,
+  "max_steps": 13791,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 3,
+  "save_steps": 500,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": true
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 2.508085260375571e+16,
+  "train_batch_size": 16,
+  "trial_name": null,
+  "trial_params": null
+}

training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:83eb56997ccc18225206b4b840525474985d7de9254a3d04681d437858b94051
+size 5304