Upload DisamBert

DisamBert.py

@@ -0,0 +1,164 @@
+from collections.abc import Generator, Iterable
+from dataclasses import dataclass
+from enum import StrEnum
+
+import pandas as pd
+import torch
+import torch.nn as nn
+from transformers import AutoConfig, AutoModel, AutoTokenizer, PreTrainedModel, PreTrainedConfig
+
+BATCH_SIZE = 64
+
+
+class ModelURI(StrEnum):
+    BASE = "answerdotai/ModernBERT-base"
+    LARGE = "answerdotai/ModernBERT-large"
+
+
+@dataclass(slots=True, frozen=True)
+class LexicalExample:
+    concept: str
+    definition: str
+
+
+@dataclass(slots=True, frozen=True)
+class PaddedBatch:
+    input_ids: torch.Tensor
+    attention_mask: torch.Tensor
+
+
+class DisamBert(PreTrainedModel):
+    def __init__(self, config:PreTrainedConfig):
+        super().__init__(config)
+        self.BaseModel = AutoModel.from_pretrained(config.name_or_path).to("cuda")
+        self.tokenizer = AutoTokenizer.from_pretrained(config.name_or_path)
+        self.classifier_head = nn.UninitializedParameter(device="cuda")
+        self.__entities = None
+        
+    @classmethod
+    def from_base(cls, base_id: ModelURI):
+        return cls(AutoConfig.from_pretrained(base_id))
+        
+    
+    def init_classifier(self, entities: Generator[LexicalExample]) -> None:
+        entity_ids = []
+        vectors = []
+        batch = []
+        n = 0
+        for entity in entities:
+            entity_ids.append(entity.concept)
+            batch.append(entity.definition)
+
+            n += 1
+            if n == BATCH_SIZE:
+                tokens = self.tokenizer(batch, padding=True, return_tensors="pt")
+                encoding = self.BaseModel(
+                    tokens["input_ids"].to("cuda"), tokens["attention_mask"].to("cuda")
+                )
+                vectors.append(encoding.last_hidden_state.detach()[:, 0])
+                n = 0
+                batch = []
+        if n > 0:
+            tokens = self.tokenizer(batch, padding=True, return_tensors="pt")
+            encoding = self.BaseModel(
+                tokens["input_ids"].to("cuda"), tokens["attention_mask"].to("cuda")
+            )
+            vectors.append(encoding.last_hidden_state.detach()[:, 0])
+
+        self.__entities = pd.Series(entity_ids)
+        self.config.entities = entity_ids
+        self.classifier_head = nn.Parameter(torch.cat(vectors, dim=0))
+
+    @property
+    def entities(self) -> pd.Series:
+        if self.__entities is None and hasattr(self.config, "entities"):
+            self.__entities = pd.Series(self.config.entities)
+        return self.__entities
+
+    def forward(self, sentences: Iterable[str], indices: Iterable[list[int]]) -> torch.Tensor:
+        assert not nn.parameter.is_lazy(self.classifier_head), (
+            "Run init_classifier to initialise weights"
+        )
+        all_indices = []
+        all_tokens = []
+
+        for sentence, span_indices in zip(sentences, indices, strict=True):
+            indices = []
+            tokens = []
+            last_span = len(span_indices) - 2
+            for i, position in enumerate(span_indices[:-1]):
+                span = sentence[position : span_indices[i + 1]]
+                span_tokens = self.tokenizer([span], padding=False)["input_ids"][0]
+                if i > 0:
+                    span_tokens = span_tokens[1:]
+                if i < last_span:
+                    span_tokens = span_tokens[:-1]
+                indices.append(len(span_tokens))
+                tokens.extend(span_tokens)
+            all_indices.append(indices)
+            all_tokens.append(tokens)
+        sentence_lengths = [len(boundaries) for boundaries in all_indices]
+        maxlen = max(sentence_lengths)
+        batch = self.pad(all_tokens)
+        token_vectors = self.BaseModel(batch.input_ids, batch.attention_mask).last_hidden_state
+        span_vectors = torch.cat(
+            [
+                torch.vstack(
+                    [
+                        torch.sum(chunk, dim=0)
+                        for chunk in self.split(token_vectors[i], sentence_indices)
+                    ]
+                )
+                for (i, sentence_indices) in enumerate(all_indices)
+            ]
+        )
+        logits = torch.einsum("ij,kj->ki", span_vectors, self.classifier_head)
+        split_logits = torch.split(logits, sentence_lengths, dim=1)
+        return torch.stack(
+            [
+                self.extend_to_max_length(sentence, length, maxlen)
+                for (sentence, length) in zip(split_logits, sentence_lengths, strict=True)
+            ]
+        )
+
+    def split(self, vectors: torch.Tensor, lengths: list[int]) -> tuple[torch.Tensor, ...]:
+        maxlen = vectors.shape[0]
+        total_length = sum(lengths)
+        is_padded = total_length < maxlen
+        chunks = vectors.split((lengths + [maxlen - total_length]) if is_padded else lengths)
+        return chunks[:-1] if is_padded else chunks
+
+    def pad(self, tokens: list[int]) -> PaddedBatch:
+        lengths = [len(sentence) for sentence in tokens]
+        maxlen = max(lengths)
+        input_ids = torch.tensor(
+            [
+                sentence + [self.config.pad_token_id] * (maxlen - length)
+                for (sentence, length) in zip(tokens, lengths, strict=True)
+            ],
+            device="cuda",
+        )
+        attention_mask = torch.vstack(
+            [
+                torch.cat(
+                    (torch.ones(length, device="cuda"), torch.zeros(maxlen - length, device="cuda"))
+                )
+                for length in lengths
+            ]
+        )
+        return PaddedBatch(input_ids, attention_mask)
+
+    def extend_to_max_length(
+        self, sentence: torch.Tensor, length: int, maxlength: int
+    ) -> torch.Tensor:
+        return (
+            torch.cat(
+                [
+                    sentence,
+                    torch.zeros((self.__entities.shape[0], maxlength - length), device="cuda"),
+                ],
+                dim=1,
+            )
+            if length < maxlength
+            else sentence
+        )

README.md

@@ -0,0 +1,199 @@
+---
+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]

model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:765767f2ce35a2118f15cef212da9c3e5159a114e6e1aa080942d3e256b12c22
+size 957523088