Added sub-tasks for phi, phi3 and persimmon

src/mlx_transformers/models/__init__.py

@@ -10,9 +10,19 @@
 from .llama import LlamaForCausalLM, LlamaModel
 from .m2m_100 import M2M100ForConditionalGeneration
 from .openelm import OpenELMForCausalLM, OpenELMModel
-from .phi import PhiForCausalLM, PhiModel
-from .phi3 import Phi3ForCausalLM, Phi3Model
-from .persimmon import PersimmonForCausalLM
+from .phi import (
+    PhiForCausalLM,
+    PhiModel,
+    PhiForSequenceClassification,
+    PhiForTokenClassification,
+)
+from .phi3 import (
+    Phi3ForCausalLM,
+    Phi3Model,
+    Phi3ForSequenceClassification,
+    Phi3ForTokenClassification,
+)
+from .persimmon import PersimmonForCausalLM, PersimmonForSequenceClassification
 from .fuyu import FuyuForCausalLM
 from .roberta import (
     RobertaForQuestionAnswering,

src/mlx_transformers/models/persimmon.py

@@ -1,6 +1,6 @@
 import math
 import logging
-from typing import Optional, Tuple, Dict
+from typing import Optional, Tuple, Dict, List, Union
 
 import mlx.core as mx
 import mlx.nn as nn
@@ -9,7 +9,11 @@
 
 from .base import MlxPretrainedMixin
 from .cache import Cache, DynamicCache
-from .modelling_outputs import BaseModelOutputWithPast, CausalLMOutputWithPast
+from .modelling_outputs import (
+    BaseModelOutputWithPast,
+    CausalLMOutputWithPast,
+    SequenceClassifierOutputWithPast,
+)
 from .utils import ACT2FN
 
 logger = logging.getLogger(__name__)
@@ -764,3 +768,120 @@ def sample(logits):
             next_token = sample(next_token_logits)
 
             yield next_token
+
+
+class PersimmonForSequenceClassification(nn.Module, MlxPretrainedMixin):
+    def __init__(self, config):
+        super().__init__()
+        self.config = config
+        self.num_labels = config.num_labels
+        self.model = PersimmonModel(config)
+        self.score = nn.Linear(config.hidden_size, self.num_labels, bias=False)
+
+    def get_input_embeddings(self):
+        return self.model.embed_tokens
+
+    def set_input_embeddings(self, value):
+        self.model.embed_tokens = value
+
+    def __call__(
+        self,
+        input_ids: mx.array = None,
+        attention_mask: Optional[mx.array] = None,
+        position_ids: Optional[mx.array] = None,
+        past_key_values: Optional[Union[Cache, List[mx.array]]] = None,
+        inputs_embeds: Optional[mx.array] = None,
+        labels: Optional[mx.array] = None,
+        use_cache: Optional[bool] = None,
+        output_attentions: Optional[bool] = None,
+        output_hidden_states: Optional[bool] = None,
+        return_dict: Optional[bool] = None,
+    ) -> Union[Tuple, SequenceClassifierOutputWithPast]:
+        r"""
+        labels (`mx.array` of shape `(batch_size,)`, *optional*):
+            Labels for computing the sequence classification/regression loss. Indices should be in `[0, ...,
+            config.num_labels - 1]`. If `config.num_labels == 1` a regression loss is computed (Mean-Square loss), If
+            `config.num_labels > 1` a classification loss is computed (Cross-Entropy).
+        """
+        return_dict = (
+            return_dict if return_dict is not None else self.config.use_return_dict
+        )
+
+        model_outputs = self.model(
+            input_ids,
+            attention_mask=attention_mask,
+            position_ids=position_ids,
+            past_key_values=past_key_values,
+            inputs_embeds=inputs_embeds,
+            use_cache=use_cache,
+            output_attentions=output_attentions,
+            output_hidden_states=output_hidden_states,
+            return_dict=return_dict,
+        )
+        hidden_states = model_outputs.last_hidden_state
+        logits = self.score(hidden_states)
+
+        if input_ids is not None:
+            batch_size = input_ids.shape[0]
+        else:
+            batch_size = inputs_embeds.shape[0]
+
+        if self.config.pad_token_id is None and batch_size != 1:
+            raise ValueError(
+                "Cannot handle batch sizes > 1 if no padding token is defined."
+            )
+        if self.config.pad_token_id is None:
+            sequence_lengths = -1
+        else:
+            if input_ids is not None:
+                # if no pad token found, use modulo instead of reverse indexing for ONNX compatibility
+                sequence_lengths = (
+                    mx.eq(input_ids, self.config.pad_token_id).int().argmax(-1) - 1
+                )
+                sequence_lengths = sequence_lengths % input_ids.shape[-1]
+                sequence_lengths = sequence_lengths.to(logits.device)
+            else:
+                sequence_lengths = -1
+
+        pooled_logits = logits[
+            mx.arange(batch_size, device=logits.device), sequence_lengths
+        ]
+
+        loss = None
+        if labels is not None:
+            labels = labels.to(logits.device)
+            if self.config.problem_type is None:
+                if self.num_labels == 1:
+                    self.config.problem_type = "regression"
+                elif self.num_labels > 1 and (
+                    labels.dtype == mx.long or labels.dtype == mx.int
+                ):
+                    self.config.problem_type = "single_label_classification"
+                else:
+                    self.config.problem_type = "multi_label_classification"
+
+            if self.config.problem_type == "regression":
+                loss_fct = nn.losses.mse_loss
+                if self.num_labels == 1:
+                    loss = loss_fct(pooled_logits.squeeze(), labels.squeeze())
+                else:
+                    loss = loss_fct(pooled_logits, labels)
+            elif self.config.problem_type == "single_label_classification":
+                loss_fct = nn.losses.cross_entropy
+                loss = loss_fct(
+                    pooled_logits.view(-1, self.num_labels), labels.view(-1)
+                )
+            elif self.config.problem_type == "multi_label_classification":
+                loss_fct = nn.losses.binary_cross_entropy
+                loss = loss_fct(pooled_logits, labels)
+        if not return_dict:
+            output = (pooled_logits,) + model_outputs[1:]
+            return ((loss,) + output) if loss is not None else output
+
+        return SequenceClassifierOutputWithPast(
+            loss=loss,
+            logits=pooled_logits,
+            past_key_values=model_outputs.past_key_values,
+            hidden_states=model_outputs.hidden_states,
+            attentions=model_outputs.attentions,
+        )

src/mlx_transformers/models/phi.py

@@ -1,6 +1,6 @@
 import math
 import logging
-from typing import Optional, Tuple, Dict
+from typing import Optional, Tuple, Dict, Union, List
 
 import mlx.core as mx
 import mlx.nn as nn
@@ -9,7 +9,12 @@
 
 from .base import MlxPretrainedMixin
 from .cache import Cache, DynamicCache
-from .modelling_outputs import BaseModelOutputWithPast, CausalLMOutputWithPast
+from .modelling_outputs import (
+    BaseModelOutputWithPast,
+    CausalLMOutputWithPast,
+    SequenceClassifierOutputWithPast,
+    TokenClassifierOutput,
+)
 from .utils import ACT2FN
 
 logger = logging.getLogger(__name__)
@@ -896,3 +901,200 @@ def sample(logits):
             next_token = sample(next_token_logits)
 
             yield next_token
+
+
+class PhiForSequenceClassification(nn.Module, MlxPretrainedMixin):
+    def __init__(self, config):
+        super().__init__()
+        self.config = config
+        self.num_labels = config.num_labels
+        self.model = PhiModel(config)
+        self.score = nn.Linear(config.hidden_size, self.num_labels, bias=False)
+
+    def get_input_embeddings(self):
+        return self.model.embed_tokens
+
+    def set_input_embeddings(self, value):
+        self.model.embed_tokens = value
+
+    def __call__(
+        self,
+        input_ids: mx.array = None,
+        attention_mask: Optional[mx.array] = None,
+        position_ids: Optional[mx.array] = None,
+        past_key_values: Optional[Union[Cache, List[mx.array]]] = None,
+        inputs_embeds: Optional[mx.array] = None,
+        labels: Optional[mx.array] = None,
+        use_cache: Optional[bool] = None,
+        output_attentions: Optional[bool] = None,
+        output_hidden_states: Optional[bool] = None,
+        return_dict: Optional[bool] = None,
+    ) -> Union[Tuple, SequenceClassifierOutputWithPast]:
+        r"""
+        labels (`mx.array` of shape `(batch_size,)`, *optional*):
+            Labels for computing the sequence classification/regression loss. Indices should be in `[0, ...,
+            config.num_labels - 1]`. If `config.num_labels == 1` a regression loss is computed (Mean-Square loss), If
+            `config.num_labels > 1` a classification loss is computed (Cross-Entropy).
+        """
+        return_dict = (
+            return_dict if return_dict is not None else self.config.use_return_dict
+        )
+
+        model_outputs = self.model(
+            input_ids,
+            attention_mask=attention_mask,
+            position_ids=position_ids,
+            past_key_values=past_key_values,
+            inputs_embeds=inputs_embeds,
+            use_cache=use_cache,
+            output_attentions=output_attentions,
+            output_hidden_states=output_hidden_states,
+            return_dict=return_dict,
+        )
+        hidden_states = model_outputs.last_hidden_state
+        logits = self.score(hidden_states)
+
+        if input_ids is not None:
+            batch_size = input_ids.shape[0]
+        else:
+            batch_size = inputs_embeds.shape[0]
+
+        if self.config.pad_token_id is None and batch_size != 1:
+            raise ValueError(
+                "Cannot handle batch sizes > 1 if no padding token is defined."
+            )
+        if self.config.pad_token_id is None:
+            sequence_lengths = -1
+        else:
+            if input_ids is not None:
+                # if no pad token found, use modulo instead of reverse indexing for ONNX compatibility
+                sequence_lengths = (
+                    mx.eq(input_ids, self.config.pad_token_id).int().argmax(-1) - 1
+                )
+                sequence_lengths = sequence_lengths % input_ids.shape[-1]
+                sequence_lengths = sequence_lengths.to(logits.device)
+            else:
+                sequence_lengths = -1
+
+        pooled_logits = logits[
+            mx.arange(batch_size, device=logits.device), sequence_lengths
+        ]
+
+        loss = None
+        if labels is not None:
+            labels = labels.to(logits.device)
+            if self.config.problem_type is None:
+                if self.num_labels == 1:
+                    self.config.problem_type = "regression"
+                elif self.num_labels > 1 and (
+                    labels.dtype == mx.long or labels.dtype == mx.int
+                ):
+                    self.config.problem_type = "single_label_classification"
+                else:
+                    self.config.problem_type = "multi_label_classification"
+
+            if self.config.problem_type == "regression":
+                loss_fct = nn.losses.mse_loss
+                if self.num_labels == 1:
+                    loss = loss_fct(pooled_logits.squeeze(), labels.squeeze())
+                else:
+                    loss = loss_fct(pooled_logits, labels)
+            elif self.config.problem_type == "single_label_classification":
+                loss_fct = nn.losses.cross_entropy
+                loss = loss_fct(
+                    pooled_logits.view(-1, self.num_labels), labels.view(-1)
+                )
+            elif self.config.problem_type == "multi_label_classification":
+                loss_fct = nn.losses.binary_cross_entropy
+                loss = loss_fct(pooled_logits, labels)
+        if not return_dict:
+            output = (pooled_logits,) + model_outputs[1:]
+            return ((loss,) + output) if loss is not None else output
+
+        return SequenceClassifierOutputWithPast(
+            loss=loss,
+            logits=pooled_logits,
+            past_key_values=model_outputs.past_key_values,
+            hidden_states=model_outputs.hidden_states,
+            attentions=model_outputs.attentions,
+        )
+
+
+class PhiForTokenClassification(nn.Module, MlxPretrainedMixin):
+    def __init__(self, config: PhiConfig):
+        super().__init__()
+        self.num_labels = config.num_labels
+        self.config = config
+
+        self.model = PhiModel(config)
+        if (
+            hasattr(config, "classifier_dropout")
+            and config.classifier_dropout is not None
+        ):
+            classifier_dropout = config.classifier_dropout
+        elif hasattr(config, "hidden_dropout") and config.hidden_dropout is not None:
+            classifier_dropout = config.hidden_dropout
+        else:
+            classifier_dropout = 0.1
+        self.dropout = nn.Dropout(classifier_dropout)
+        self.classifier = nn.Linear(config.hidden_size, config.num_labels)
+
+    def __call__(
+        self,
+        input_ids: Optional[mx.array] = None,
+        past_key_values: Optional[Tuple[Tuple[mx.array, mx.array], ...]] = None,
+        attention_mask: Optional[mx.array] = None,
+        inputs_embeds: Optional[mx.array] = None,
+        labels: Optional[mx.array] = None,
+        use_cache: Optional[bool] = None,
+        output_attentions: Optional[bool] = None,
+        output_hidden_states: Optional[bool] = None,
+        return_dict: Optional[bool] = None,
+        **deprecated_arguments,
+    ) -> Union[Tuple[mx.array], TokenClassifierOutput]:
+        r"""
+        labels (`mx.array` of shape `(batch_size,)`, *optional*):
+            Labels for computing the sequence classification/regression loss. Indices should be in `[0, ...,
+            config.num_labels - 1]`. If `config.num_labels == 1` a regression loss is computed (Mean-Square loss), If
+            `config.num_labels > 1` a classification loss is computed (Cross-Entropy).
+        """
+        return_dict = (
+            return_dict if return_dict is not None else self.config.use_return_dict
+        )
+
+        model_outputs = self.model(
+            input_ids,
+            past_key_values=past_key_values,
+            attention_mask=attention_mask,
+            inputs_embeds=inputs_embeds,
+            use_cache=use_cache,
+            output_attentions=output_attentions,
+            output_hidden_states=output_hidden_states,
+            return_dict=return_dict,
+        )
+
+        hidden_states = model_outputs.last_hidden_state
+        hidden_states = self.dropout(hidden_states)
+        logits = self.classifier(hidden_states)
+
+        loss = None
+        if labels is not None:
+            # move labels to correct device to enable model parallelism
+            labels = labels.to(logits.device)
+            batch_size, seq_length = labels.shape
+            loss_fct = nn.losses.cross_entropy
+            loss = loss_fct(
+                logits.view(batch_size * seq_length, self.num_labels),
+                labels.view(batch_size * seq_length),
+            )
+
+        if not return_dict:
+            output = (logits,) + model_outputs[2:]
+            return ((loss,) + output) if loss is not None else output
+
+        return TokenClassifierOutput(
+            loss=loss,
+            logits=logits,
+            hidden_states=model_outputs.hidden_states,
+            attentions=model_outputs.attentions,
+        )