Upload model

README.md

@@ -0,0 +1,201 @@
+---
+library_name: transformers
+tags: []
+---
+
+# Model Card for Model ID
+
+<!-- Provide a quick summary of what the model is/does. -->
+
+
+
+## Model Details
+
+### Model Description
+
+<!-- Provide a longer summary of what this model is. -->
+
+This is the model card of a 🤗 transformers model that has been pushed on the Hub. This model card has been automatically generated.
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+<!-- Provide the basic links for the model. -->
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
+
+### Direct Use
+
+<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+<!-- This section is meant to convey both technical and sociotechnical limitations. -->
+
+[More Information Needed]
+
+### Recommendations
+
+<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
+
+[More Information Needed]
+
+### Training Procedure 
+
+<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
+
+#### Speeds, Sizes, Times [optional]
+
+<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
+
+[More Information Needed]
+
+## Evaluation
+
+<!-- This section describes the evaluation protocols and provides the results. -->
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+<!-- This should link to a Dataset Card if possible. -->
+
+[More Information Needed]
+
+#### Factors
+
+<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
+
+[More Information Needed]
+
+#### Metrics
+
+<!-- These are the evaluation metrics being used, ideally with a description of why. -->
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+<!-- Relevant interpretability work for the model goes here -->
+
+[More Information Needed]
+
+## Environmental Impact
+
+<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+
+

config.json

@@ -0,0 +1,200 @@
+{
+  "_name_or_path": "liaad/srl-en_roberta-large_hf",
+  "architectures": [
+    "SRLModel"
+  ],
+  "auto_map": {
+    "AutoConfig": "config.SRLModelConfig",
+    "AutoModel": "model.SRLModel"
+  },
+  "bert_model_name": "FacebookAI/roberta-large",
+  "embedding_dropout": 0.1,
+  "id2label": {
+    "0": "O",
+    "1": "<unk>",
+    "10": "I-C-A0",
+    "11": "I-C-A2",
+    "12": "B-A2",
+    "13": "B-R-A1",
+    "14": "I-C-AM-COM",
+    "15": "B-A0",
+    "16": "B-A4",
+    "17": "I-A3",
+    "18": "B-C-AM-COM",
+    "19": "B-C-AM-DIS",
+    "2": "I-C-AM-LOC",
+    "20": "B-C-AM-PRP",
+    "21": "I-AM-ADV",
+    "22": "I-AM-PRP",
+    "23": "I-AM-DIS",
+    "24": "B-C-A3",
+    "25": "B-AM-PRD",
+    "26": "B-C-A1",
+    "27": "I-C-AM-ADV",
+    "28": "I-C-AM-MNR",
+    "29": "I-R-A3",
+    "3": "I-AM-COM",
+    "30": "B-C-AM-NEG",
+    "31": "I-C-A3",
+    "32": "B-C-A2",
+    "33": "I-C-AM-NEG",
+    "34": "B-A3",
+    "35": "B-AM-LOC",
+    "36": "B-C-AM-TMP",
+    "37": "I-AM-DIR",
+    "38": "B-R-A2",
+    "39": "B-A5",
+    "4": "B-R-A0",
+    "40": "B-C-A0",
+    "41": "I-C-AM-PRP",
+    "42": "B-C-AM-PRD",
+    "43": "I-A1",
+    "44": "B-C-AM-EXT",
+    "45": "B-AM-CAU",
+    "46": "B-AM-MOD",
+    "47": "B-AM-COM",
+    "48": "B-AM-NEG",
+    "49": "B-AM-REC",
+    "5": "B-R-A4",
+    "50": "B-AM-ADV",
+    "51": "I-R-A1",
+    "52": "B-AM-DIR",
+    "53": "I-C-AM-TMP",
+    "54": "B-C-AM-MNR",
+    "55": "I-AM-GOL",
+    "56": "I-R-A2",
+    "57": "I-R-A0",
+    "58": "B-C-AM-DIR",
+    "59": "I-C-A4",
+    "6": "I-R-A4",
+    "60": "B-C-AM-LOC",
+    "61": "I-A2",
+    "62": "I-AM-MOD",
+    "63": "B-C-AM-DSP",
+    "64": "B-AM-TMP",
+    "65": "I-AM-REC",
+    "66": "I-C-AM-EXT",
+    "67": "I-C-AM-DIR",
+    "68": "I-AM-CAU",
+    "69": "I-C-AM-DIS",
+    "7": "B-AM-DIS",
+    "70": "I-C-AM-ADJ",
+    "71": "I-AM-LOC",
+    "72": "I-AM-NEG",
+    "73": "B-A1",
+    "74": "B-C-AM-ADV",
+    "75": "B-AM-GOL",
+    "76": "I-C-AM-DSP",
+    "77": "I-AM-PRD",
+    "78": "B-C-AM-ADJ",
+    "79": "B-C-A4",
+    "8": "B-V",
+    "80": "B-AM-EXT",
+    "81": "I-AM-TMP",
+    "82": "I-AM-EXT",
+    "83": "I-A0",
+    "84": "B-AM-MNR",
+    "85": "I-C-A1",
+    "86": "I-A4",
+    "87": "I-AM-MNR",
+    "88": "B-AM-PRP",
+    "89": "B-R-A3",
+    "9": "I-A5"
+  },
+  "label2id": {
+    "<unk>": 1,
+    "B-A0": 15,
+    "B-A1": 73,
+    "B-A2": 12,
+    "B-A3": 34,
+    "B-A4": 16,
+    "B-A5": 39,
+    "B-AM-ADV": 50,
+    "B-AM-CAU": 45,
+    "B-AM-COM": 47,
+    "B-AM-DIR": 52,
+    "B-AM-DIS": 7,
+    "B-AM-EXT": 80,
+    "B-AM-GOL": 75,
+    "B-AM-LOC": 35,
+    "B-AM-MNR": 84,
+    "B-AM-MOD": 46,
+    "B-AM-NEG": 48,
+    "B-AM-PRD": 25,
+    "B-AM-PRP": 88,
+    "B-AM-REC": 49,
+    "B-AM-TMP": 64,
+    "B-C-A0": 40,
+    "B-C-A1": 26,
+    "B-C-A2": 32,
+    "B-C-A3": 24,
+    "B-C-A4": 79,
+    "B-C-AM-ADJ": 78,
+    "B-C-AM-ADV": 74,
+    "B-C-AM-COM": 18,
+    "B-C-AM-DIR": 58,
+    "B-C-AM-DIS": 19,
+    "B-C-AM-DSP": 63,
+    "B-C-AM-EXT": 44,
+    "B-C-AM-LOC": 60,
+    "B-C-AM-MNR": 54,
+    "B-C-AM-NEG": 30,
+    "B-C-AM-PRD": 42,
+    "B-C-AM-PRP": 20,
+    "B-C-AM-TMP": 36,
+    "B-R-A0": 4,
+    "B-R-A1": 13,
+    "B-R-A2": 38,
+    "B-R-A3": 89,
+    "B-R-A4": 5,
+    "B-V": 8,
+    "I-A0": 83,
+    "I-A1": 43,
+    "I-A2": 61,
+    "I-A3": 17,
+    "I-A4": 86,
+    "I-A5": 9,
+    "I-AM-ADV": 21,
+    "I-AM-CAU": 68,
+    "I-AM-COM": 3,
+    "I-AM-DIR": 37,
+    "I-AM-DIS": 23,
+    "I-AM-EXT": 82,
+    "I-AM-GOL": 55,
+    "I-AM-LOC": 71,
+    "I-AM-MNR": 87,
+    "I-AM-MOD": 62,
+    "I-AM-NEG": 72,
+    "I-AM-PRD": 77,
+    "I-AM-PRP": 22,
+    "I-AM-REC": 65,
+    "I-AM-TMP": 81,
+    "I-C-A0": 10,
+    "I-C-A1": 85,
+    "I-C-A2": 11,
+    "I-C-A3": 31,
+    "I-C-A4": 59,
+    "I-C-AM-ADJ": 70,
+    "I-C-AM-ADV": 27,
+    "I-C-AM-COM": 14,
+    "I-C-AM-DIR": 67,
+    "I-C-AM-DIS": 69,
+    "I-C-AM-DSP": 76,
+    "I-C-AM-EXT": 66,
+    "I-C-AM-LOC": 2,
+    "I-C-AM-MNR": 28,
+    "I-C-AM-NEG": 33,
+    "I-C-AM-PRP": 41,
+    "I-C-AM-TMP": 53,
+    "I-R-A0": 57,
+    "I-R-A1": 51,
+    "I-R-A2": 56,
+    "I-R-A3": 29,
+    "I-R-A4": 6,
+    "O": 0
+  },
+  "model_type": "srl",
+  "num_labels": 90,
+  "torch_dtype": "float32",
+  "transformers_version": "4.39.3"
+}

config.py

@@ -0,0 +1,29 @@
+from transformers import PretrainedConfig
+
+class SRLModelConfig(PretrainedConfig):
+    model_type = "srl"
+    
+    def __init__(
+        self,
+        num_labels=0,
+        bert_model_name="bert-base-uncased",
+        embedding_dropout=0.0,
+        label2id = {},
+        id2label = {},
+        **kwargs,
+    ):
+        super().__init__(**kwargs)
+        self.num_labels = num_labels
+        self.bert_model_name = bert_model_name
+        self.embedding_dropout = embedding_dropout
+        self.label2id = label2id
+        self.id2label = id2label
+
+    def to_dict(self):
+        config_dict = super().to_dict()
+
+        config_dict["num_labels"] = self.num_labels
+        # config_dict["bert_model_name"] = self.bert_model_name
+        # config_dict["embedding_dropout"] = self.embedding_dropout
+
+        return config_dict

model.py

@@ -0,0 +1,96 @@
+import torch.nn as nn
+import torch.nn.functional as F
+from transformers import AutoModel, AutoTokenizer, PreTrainedModel
+from .config import SRLModelConfig
+
+
+class SRLModel(PreTrainedModel):
+    config_class = SRLModelConfig
+
+    def __init__(self, config):
+        super().__init__(config)
+
+        print(config.num_labels, config.bert_model_name, config.embedding_dropout)
+
+        # Load pre-trained transformer-based model and tokenizer
+        self.tokenizer = AutoTokenizer.from_pretrained(config.bert_model_name)
+        self.transformer = AutoModel.from_pretrained(
+            config.bert_model_name,
+            num_labels=config.num_labels,
+            output_hidden_states=True,
+        )
+        self.transformer.config.id2label = config.id2label
+        self.transformer.config.label2id = config.label2id
+
+        # The roberta models do not have token_type_embeddings
+        # (the type_vocab_size is 1)
+        # but we use this to pass the verb's position
+        # so we need to change the model and initialize the embeddings randomly
+        if "xlm" in config.bert_model_name or "roberta" in config.bert_model_name:
+            self.transformer.config.type_vocab_size = 2
+            # Create a new Embeddings layer, with 2 possible segments IDs instead of 1
+            self.transformer.embeddings.token_type_embeddings = nn.Embedding(
+                2, self.transformer.config.hidden_size
+            )
+            # Initialize it
+            self.transformer.embeddings.token_type_embeddings.weight.data.normal_(
+                mean=0.0, std=self.transformer.config.initializer_range
+            )
+
+        # Linear layer for tag projection
+        self.tag_projection_layer = nn.Linear(
+            self.transformer.config.hidden_size, config.num_labels
+        )
+
+        # Dropout layer for embeddings
+        self.embedding_dropout = nn.Dropout(p=config.embedding_dropout)
+
+        # Number of labels
+        self.num_labels = config.num_labels
+
+    def forward(self, input_ids, attention_mask, token_type_ids, labels=None):
+
+        # print("FORWARD")
+        # print(labels)
+
+        # Forward pass through the transformer model
+        # Returns BaseModelOutputWithPoolingAndCrossAttentions
+        outputs = self.transformer(
+            input_ids=input_ids,
+            attention_mask=attention_mask,
+            token_type_ids=token_type_ids,
+        )
+
+        # Extract the [CLS] token representation
+        # cls_output = outputs.pooler_output
+
+        bert_embedding = outputs.last_hidden_state
+
+        # Apply dropout to the embeddings
+        embedded_text_input = self.embedding_dropout(bert_embedding)
+
+        # Project to tag space
+        logits = self.tag_projection_layer(embedded_text_input)
+
+        reshaped_log_probs = logits.view(-1, self.num_labels)
+        class_probabilities = F.softmax(reshaped_log_probs, dim=-1).view(
+            logits.size(0), logits.size(1), -1
+        )
+
+        output_dict = {"logits": logits, "class_probabilities": class_probabilities}
+
+        output_dict["attention_mask"] = attention_mask
+        output_dict["input_ids"] = input_ids
+        # output_dict["start_offsets"] = start_offsets
+
+        if labels is not None:
+            # print("Input", logits.view(-1, self.num_labels).size())
+            # print("Target", labels.view(-1).size())
+            # print("C", self.num_labels)
+            # print("AllenNLP function", logits.size(-1))
+            # Could consider passing ignore_index as 0 (pad index) for minor optimization
+            loss = nn.CrossEntropyLoss(ignore_index=-100)(
+                logits.view(-1, self.num_labels), labels.view(-1)
+            )
+            output_dict["loss"] = loss
+        return output_dict

model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a882024c5267e9d07627c0b53281a2e8117619330fa629bd10f789fb61bbd635
+size 1421861880